PDA

View Full Version : John Erardi Gets It -- "Who's Counting"



Pages : 1 [2]

Sea Ray
05-14-2008, 04:26 PM
I can tell you if it were an academic book, it'd be ripped to shreds in peer review.

What do you mean by "it"? You'd rip something to shreads that doesn't exist?

jojo
05-14-2008, 04:28 PM
I don't care how Adam Dunn or any other Reds players create runs; I just care that they create runs.

And really that sums it up-don't blame Dunn if the Reds can't leverage his skill set.

Is he an elite player? No. Offensively he's been pretty close but his defense means he's basically a 3 win player overall when having what has been a typical offensive season for him.

I find it difficult to label a roughly three win over replacement player as a major problem on a roster.....

Cyclone792
05-14-2008, 04:32 PM
How do they contradict themselves? In a vacuum sure they contradict themselves. But when you look at each individual game and each individual situation they are not contradictary.

You're applying micro to macro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

Avoiding outs correlates to actual run production at 90+ percent. Acquiring bases also correlates to actual run production at 90+ percent. If your goal is maximize run production, then your goal is to accomplish what most correlates to run production ... which is avoiding outs and acquiring bases.

kaldaniels
05-14-2008, 04:35 PM
You're applying micro to macro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

Avoiding outs correlates to actual run production at 90+ percent. Acquiring bases also correlates to actual run production at 90+ percent. If your goal is maximize run production, then your goal is to accomplish what most correlates to run production ... which is avoiding outs and acquiring bases.

Which would correspond to OBP and SLG %...am I correct in that assumption.

Sea Ray
05-14-2008, 04:39 PM
And really that sums it up-don't blame Dunn if the Reds can't leverage his skill set.

Is he an elite player? No. Offensively he's been pretty close but his defense means he's basically a 3 win player overall when having what has been a typical offensive season for him.

I find it difficult to label a roughly three win over replacement player as a major problem on a roster.....

I'm with you there.

Anybody else disagree with Jojo? Is Dunn an elite player?

dabvu2498
05-14-2008, 04:41 PM
I find it difficult to label a roughly three win over replacement player as a major problem on a roster.....

What is one who is going into his 29 year old season worth per year?

That is the 13-18 million dollar question.

Cyclone792
05-14-2008, 04:41 PM
Which would correspond to OBP and SLG %...am I correct in that assumption.

Precisely. OBP is the rate of avoiding outs and SLG is the rate of base acquisition.

bucksfan2
05-14-2008, 04:42 PM
You're applying micro to macro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

Avoiding outs correlates to actual run production at 90+ percent. Acquiring bases also correlates to actual run production at 90+ percent. If your goal is maximize run production, then your goal is to accomplish what most correlates to run production ... which is avoiding outs and acquiring bases.

But baseball is played in the micro correct? Its not Strat-O-Matic out there. My point behind this discussion is that you can take all the macro tools and a make a gereralizatoin about something but what happens on a given pitch or atbat (the micro) can defy all logic.

To combat your agument I will say this: You're applying macro to micro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

vaticanplum
05-14-2008, 04:43 PM
What you're failing to understand is those two points above contradict each other. In order to maximize the amount of runs scored in a given game, it means making as few outs as possible.

Bingo. If you don't make outs, you will eventually score runs. It's a direct causal relationship.

I get the impression that some people don't think watching four consecutive guys walk is real baseball. But guess what, four consecutive guys walk and you score a run. Every time.

MWM
05-14-2008, 04:43 PM
But baseball is played in the micro correct? Its not Strat-O-Matic out there. My point behind this discussion is that you can take all the macro tools and a make a gereralizatoin about something but what happens on a given pitch or atbat (the micro) can defy all logic.

To combat your agument I will say this: You're applying macro to micro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.


I have no idea what you just said.

kaldaniels
05-14-2008, 04:44 PM
Precisely. OBP is the rate of avoiding outs and SLG is the rate of base acquisition.

Got it. Thanks.

jojo
05-14-2008, 04:45 PM
What is one who is going into his 29 year old season worth per year?

That is the 13-18 million dollar question.

Well being a roughly 3 win over replacement guy over the last 6 years and still being one over the next five seasons are two different issues.....

Cyclone792
05-14-2008, 04:45 PM
But baseball is played in the micro correct? Its not Strat-O-Matic out there. My point behind this discussion is that you can take all the macro tools and a make a gereralizatoin about something but what happens on a given pitch or atbat (the micro) can defy all logic.

To combat your agument I will say this: You're applying macro to micro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

Uh, what?

I don't intend to be mean here, but the more you post on this subject the more it's apparent that you are not understanding the premise. You're either willing to learn and understand the foundation of run production ... or you're not.

kaldaniels
05-14-2008, 04:45 PM
Bingo. If you don't make outs, you will eventually score runs. It's a direct causal relationship.

I get the impression that some people don't think watching four consecutive guys walk is real baseball. But guess what, four consecutive guys walk and you score a run. Every time.

Ryan Freel would have have an argument with that. :D

pahster
05-14-2008, 04:48 PM
What do you mean by "it"? You'd rip something to shreads that doesn't exist?

Your entire argument throughout this thread.

Highlifeman21
05-14-2008, 04:48 PM
I'm with you there.

Anybody else disagree with Jojo? Is Dunn an elite player?

I don't know that anyone's made the argument he's an elite player.

Compared to other Reds, he's an elite player, but he's in the 2nd tier of elite, league-wide. He's not in the same category as ARod or Pujols or Manny or Ortiz, but he's in the category directly below them.

And that's not a bad thing.

jojo
05-14-2008, 04:50 PM
I'm with you there.

Anybody else disagree with Jojo? Is Dunn an elite player?

I'm not sure it's important to focus on whether Dunn has been an "elite" player or simply an above average one. The important point is that he's been a valuable player for the Reds and that shouldn't get lost in labels because the appropriate baseline isn't AROD.

Sea Ray
05-14-2008, 04:53 PM
Your entire argument throughout this thread.

Nobody's posts on here would stand up to academic review. That's not what we do around here

Sea Ray
05-14-2008, 04:54 PM
I don't know that anyone's made the argument he's an elite player.

Compared to other Reds, he's an elite player, but he's in the 2nd tier of elite, league-wide. He's not in the same category as ARod or Pujols or Manny or Ortiz, but he's in the category directly below them.

And that's not a bad thing.

HOF has been mentioned a few times...

bucksfan2
05-14-2008, 04:57 PM
Uh, what?

I don't intend to be mean here, but the more you post on this subject the more it's apparent that you are not understanding the premise. You're either willing to learn and understand the foundation of run production ... or you're not.

The object of the game is to score more runs than your opponent? Agree.

Even the best hitters have a success rate of about 40% and a failure rate of about 60%. Agree.

In a perfect world every batter up would get a hit and you wouldn't have to deal with outs. My opinions on the game are to make the best usage of the 27 outs. You can try and avoid outs all you want but they are going to happen. If it takes an out to score a run then so be it.

Here is an example. Tie game bottom of the ninth, runner on 3rd, 1 out. What is the batter's responsibility? IMO his responsibility is to score that runner by any means possible. A walk doesn't score that runner, nor does a lazy fly ball or a grounder to an infielder, nor does a strikeout. If the batter gets on base but doesn't score the run there is a very likely situation in which the next batter could ground into a double play eliminating the scoring opportunity.

I understand the notion of not making outs but that is easier said than done. Thats like telling a golfer to drive the ball 280 off the tee every time. I believe that there is such a thing as a productive out. I know there are others out there who don't so we disagree, fair enough.

jojo
05-14-2008, 05:00 PM
HOF has been mentioned a few times...

You mean like here:


If Dunn could actually catch, he'd be in HOF discussions.

I think that's hardly calling him a HOFer. Basically it's suggesting his bat is within a range that would get him into the debate all other things ignored.

BCubb2003
05-14-2008, 05:01 PM
These debates almost always come down to stats vs. anecdotes. The 90% correlation is a strong argument but it doesn't seem to overcome the memories of your big RBI guy walking so a weaker batter can strike out. Maybe OBP should be adjusted for LOB.

pahster
05-14-2008, 05:05 PM
Nobody's posts on here would stand up to academic review. That's not what we do around here

I agree with you in the sense that there has been no new knowledge unearthed here, but I disagree in that many of the arguments put forth in this thread are based on rigorous analyses. We know that it's SLG that drives RBI, not AVG (as if it matters; RBI is a silly metric for individuals) and we know that OBP trumps AVG when it comes to scoring runs. This makes your BA w/RISP argument silly when it comes to Dunn specifically; Dunn's OBP improves with runners on. That's a good thing because it will lead to more runs being scored over the course of a season.

Yachtzee
05-14-2008, 05:06 PM
But baseball is played in the micro correct? Its not Strat-O-Matic out there. My point behind this discussion is that you can take all the macro tools and a make a gereralizatoin about something but what happens on a given pitch or atbat (the micro) can defy all logic.

To combat your agument I will say this: You're applying macro to micro, and that's a huge mistake. When you do that you'll draw some extremely misinformed and flawed conclusions.

I don't know. Earl Weaver seemed to be quite successful as a manager and much of his thinking supports the macro view. If you avoid outs and acquire bases, runs will score. In micro situations, you may think "oh that productive out was successful because it scored one run." But also happened is that it decreased the likelihood of scoring multiple runs. Any team playing in GABP should be playing for multiple runs. I'd say the only time the micro view is warranted is in late inning ties and when the opposing pitcher is throwing a no-no against you and you need to get on the board.

I think many baseball managers like to operate under the notion that things like situational hitting and bunting and sacrifice flies and hit and runs and such are important because it gives them something to do during the game. A small ball manager probably gets the feeling that he is more a part of the team's wins when he's out there micromanaging things. But managers have found success by filing out the lineup card and sitting back and letting the players do what they do best.

Highlifeman21
05-14-2008, 05:07 PM
The object of the game is to score more runs than your opponent? Agree.

Even the best hitters have a success rate of about 40% and a failure rate of about 60%. Agree.

In a perfect world every batter up would get a hit and you wouldn't have to deal with outs. My opinions on the game are to make the best usage of the 27 outs. You can try and avoid outs all you want but they are going to happen. If it takes an out to score a run then so be it.

Here is an example. Tie game bottom of the ninth, runner on 3rd, 1 out. What is the batter's responsibility? IMO his responsibility is to score that runner by any means possible. A walk doesn't score that runner, nor does a lazy fly ball or a grounder to an infielder, nor does a strikeout. If the batter gets on base but doesn't score the run there is a very likely situation in which the next batter could ground into a double play eliminating the scoring opportunity.

I understand the notion of not making outs but that is easier said than done. Thats like telling a golfer to drive the ball 280 off the tee every time. I believe that there is such a thing as a productive out. I know there are others out there who don't so we disagree, fair enough.

Not sure how this analogy helps your argument. What kind of golfer are we talking about?

And as for productive outs, we might as well start talking about win efficient runs, or whatever that wacky topic was a year or two ago.

Yachtzee
05-14-2008, 05:15 PM
I understand the notion of not making outs but that is easier said than done. Thats like telling a golfer to drive the ball 280 off the tee every time. I believe that there is such a thing as a productive out. I know there are others out there who don't so we disagree, fair enough.

It's easier to take what the pitcher gives you and avoid an out than it is to hit an intentional sack fly or slap the ball to the right side on a hit and run on command. If I'm trying to avoid an out, I look for pitches I can drive and lay off the ones I can't handle. If it's borderline then I swing and hope I foul it off or get lucky with a hit. If I get a pitch I can drive, it's likely to be extra bases and if I get 4 pitches out of the zone, I avoid the out and get a base for my trouble, giving the next guy the opportunity to plate at least two runs, me and himself (more if there were any on ahead of me). If I'm trying to hit an intentional sac fly or slap the ball to the right side on a hit and run, I'm doing all kinds of bad things I don't normally do to make contact with a pitch that I might not be able to handle, just because someone's calling for a productive out. I might get a productive out, but I'm much more likely to make a non-productive out than I would be if I were just trying to avoid outs all together.

RedsManRick
05-14-2008, 05:15 PM
My whole thing with Dunn is that he get's treated as though he is a borderline HOF'er. I just don't see how anyone with as big a flaw as he has at making contact is treated as such. To me those who feel that way are being way too bias to one philosophy. That philosophy is IMO flawed if it says that a guy who cannot make better consistent contact than that is a HOF'er if his defense were marginally better.

Perhaps you are misunderstanding the "philosophy". I'll try to state it as clearly as possible. Two statements.

1.) Outcome measurement (run production in this case) should be done objectively where possible, through careful analysis about what actions and events lead to the desired outcome and in what proportion.

2.) The players who's actions produce the most runs, are the best offensive players.

That's it. This isn't about "liking" certain stats, it's not about championing walks, or preferring a certain style of play. It's not about HR totals or leading the team in RBI. It's about Adam Dunn doing things at the plate which lead to a ton of run production.

It's not that I "like" walks. I'll take a hit over a walk every time. It's not that I "like" a low batting average. I wish Adam Dunn hit .270 too. Heck, I wish he hit .340. It's that I don't want to be biased by the measures I "like". I want to judge a player based on his actual production, not my personal aesthetic preferences.


Just so I'm understood I don't have a major problem with Adam Dunn, but obviously with this philosophy. I do have a small problem with Dunn at times because I wonder if he is really serious about improving because I feel he can still improve on some aspects but he never really does. My vision for him would be about a consistent .270 Hitter w/o giving up any power by "sewing up" that hole in his swing all the way (or at least most of the way) and moving to 1st base. If he removed the obvious major issues I would be happy to keep him and plant him squarely in the #3 or 4 holes for the next 5-6 yrs, otherwise I will not feel bad if he moves on.

I'd love for Dunn to hit .270. I'd love for Jeff Keppinger to slug .500. I'd love for Joey Votto to play a gold glove CF. So what? Every player has certain things they do well and certain things they do poorly. Why does it make sense to bemoan Dunn's lack of batting average (something borne out of his inherent skill set of being a big guy with mediocre hand/eye coordination) but not Ryan Freel's lack of slugging? Just as Dunn could cut down his swing and try to put the ball in play more often, Freel could muscle up and swing big. But nobody asks Freel to do that or calls him lazy for not doing it. They recognize that he's using his skill set to maxmize his production. For some reason, Dunn doesn't get that benefit of the doubt.

I just don't get this "feeling" that Adam Dunn must not care because he happens to hit for a low average. They've turned their personal bias regarding the style of play they prefer in to a moral failing of Dunn's. I just don't understand it.

I don't love Dunn as a player. He's not very fun to watch play. His defense is poor and his HR-driven SLG% leads to a somewhat inconsistent distribution of production. That said, I champion his cause not because I think he's a model baseball player, but because he's under appreciated. He gets called lazy and ridiculed constantly because of other people's bias against the way he plays the game. They have certain notions about what "should" matter, and he offends those notions. He's a very valuable player and I hope he ends up in a place that can better appreciate the tremendous value he brings. The Reds deserve to have Norris Hopper instead.

vaticanplum
05-14-2008, 05:16 PM
I think many baseball managers like to operate under the notion that things like situational hitting and bunting and sacrifice flies and hit and runs and such are important because it gives them something to do during the game. A small ball manager probably gets the feeling that he is more a part of the team's wins when he's out there micromanaging things. But managers have found success by filing out the lineup card and sitting back and letting the players do what they do best.

That is quite astute. You should manage.

There's a saying in theater that 90 percent of directing is casting. It's usually true.

Raisor
05-14-2008, 05:20 PM
I'd love for Dunn to hit .270. I'd love for Jeff Keppinger to slug .500. I'd love for Joey Votto to play a gold glove CF. So what? Every player has certain things they do well and certain things they do poorly. Why does it make sense to bemoan Dunn's lack of batting average (something borne out of his inherent skill set of being a big guy with mediocre hand/eye coordination) but not Ryan Freel's lack of slugging? Just as Dunn could cut down his swing and try to put the ball in play more often, Freel could muscle up and swing big. But nobody asks Freel to do that or calls him lazy for not doing it. They recognize that he's using his skill set to maxmize his production. For some reason, Dunn doesn't get that benefit of the doubt.

.



I'm having some serious man-love for you right now.

:thumbup:

Highlifeman21
05-14-2008, 05:21 PM
I'm having some serious man-love for you right now.

:thumbup:

Get a room, you two.

Cyclone792
05-14-2008, 05:26 PM
The object of the game is to score more runs than your opponent? Agree.

Even the best hitters have a success rate of about 40% and a failure rate of about 60%. Agree.

In a perfect world every batter up would get a hit and you wouldn't have to deal with outs. My opinions on the game are to make the best usage of the 27 outs. You can try and avoid outs all you want but they are going to happen. If it takes an out to score a run then so be it.

Here is an example. Tie game bottom of the ninth, runner on 3rd, 1 out. What is the batter's responsibility? IMO his responsibility is to score that runner by any means possible. A walk doesn't score that runner, nor does a lazy fly ball or a grounder to an infielder, nor does a strikeout. If the batter gets on base but doesn't score the run there is a very likely situation in which the next batter could ground into a double play eliminating the scoring opportunity.

You just cited an example that probably occurs in less than 0.001 percent of all plate appearances to try to defend your stance.

If ever there was a case of misinformation and flawed analysis, that right there is it.

Highlifeman21
05-14-2008, 05:28 PM
The object of the game is to score more runs than your opponent? Agree.

Even the best hitters have a success rate of about 40% and a failure rate of about 60%. Agree.

In a perfect world every batter up would get a hit and you wouldn't have to deal with outs. My opinions on the game are to make the best usage of the 27 outs. You can try and avoid outs all you want but they are going to happen. If it takes an out to score a run then so be it.

Here is an example. Tie game bottom of the ninth, runner on 3rd, 1 out. What is the batter's responsibility? IMO his responsibility is to score that runner by any means possible. A walk doesn't score that runner, nor does a lazy fly ball or a grounder to an infielder, nor does a strikeout. If the batter gets on base but doesn't score the run there is a very likely situation in which the next batter could ground into a double play eliminating the scoring opportunity.
I understand the notion of not making outs but that is easier said than done. Thats like telling a golfer to drive the ball 280 off the tee every time. I believe that there is such a thing as a productive out. I know there are others out there who don't so we disagree, fair enough.

So in your example, the next batter is Brandon Phillips?

Mario-Rijo
05-14-2008, 10:13 PM
Perhaps you are misunderstanding the "philosophy". I'll try to state it as clearly as possible. Two statements.

1.) Outcome measurement (run production in this case) should be done objectively where possible, through careful analysis about what actions and events lead to the desired outcome and in what proportion.

2.) The players who's actions produce the most runs, are the best offensive players.

That's it. This isn't about "liking" certain stats, it's not about championing walks, or preferring a certain style of play. It's not about HR totals or leading the team in RBI. It's about Adam Dunn doing things at the plate which lead to a ton of run production.

It's not that I "like" walks. I'll take a hit over a walk every time. It's not that I "like" a low batting average. I wish Adam Dunn hit .270 too. Heck, I wish he hit .340. It's that I don't want to be biased by the measures I "like". I want to judge a player based on his actual production, not my personal aesthetic preferences.

I do not misunderstand it at all, in fact I have it down cold. RC is a statistic that measures how many runs a player creates through his actions. The actions that has been decided on which correlate to higher runs created is OBP% and SLG% which combine for OPS%. OBP% tells us how often a player avoids an out, and SLG% a measurement of power.



I'd love for Dunn to hit .270. I'd love for Jeff Keppinger to slug .500. I'd love for Joey Votto to play a gold glove CF. So what? Every player has certain things they do well and certain things they do poorly. Why does it make sense to bemoan Dunn's lack of batting average (something borne out of his inherent skill set of being a big guy with mediocre hand/eye coordination) but not Ryan Freel's lack of slugging? Just as Dunn could cut down his swing and try to put the ball in play more often, Freel could muscle up and swing big. But nobody asks Freel to do that or calls him lazy for not doing it. They recognize that he's using his skill set to maxmize his production. For some reason, Dunn doesn't get that benefit of the doubt.

I just don't get this "feeling" that Adam Dunn must not care because he happens to hit for a low average. They've turned their personal bias regarding the style of play they prefer in to a moral failing of Dunn's. I just don't understand it.

I don't love Dunn as a player. He's not very fun to watch play. His defense is poor and his HR-driven SLG% leads to a somewhat inconsistent distribution of production. That said, I champion his cause not because I think he's a model baseball player, but because he's under appreciated. He gets called lazy and ridiculed constantly because of other people's bias against the way he plays the game. They have certain notions about what "should" matter, and he offends those notions. He's a very valuable player and I hope he ends up in a place that can better appreciate the tremendous value he brings. The Reds deserve to have Norris Hopper instead.

And this is where this argument always ends up and I hate it. We simply cannot get to the bottom of it w/o someone making fun of someone else's view or someone making the accusation that he is not appreciated. I know he pretty much is what he is and I don't fault anyone for lack of a specific talent it's not something they can control. But I do fault a guy for a lack of skill at times if that player can improve it and doesn't. His inability to make consistent contact has some to do with talent but also with a lack of skill. What's wrong with asking a guy with his considerable talents to increase certain skills to maximize the player he can be.

And oh BTW we do bemoan others for their skillset every single player on this team has been torn down. I know for instance that BP get's ripped because he cannot lay off certain pitches and thus has been labeled an out-machine at times. I know Ryan Freel get's ripped for his pathetic baserunning and poor angles to balls hit to him. EE get's hammered because of his inability to consistently throw the ball to 1st base w/o issue. These are all skills and not necc. talents and therefore everyone of them are viable complaints about them. The big difference between Dunn and the rest? We have been watching it for way too long (at least 2 yrs longer than any of the others) with little improvement in any way.

You might say he's not capable of improving his contact w/o sacrificing power or plate discipline, I say I think he can. Now I am no pro scout so I could be wrong he may not be able to maximize his contact rate but until I hear someone who is qualified tell me he cannot improve the above w/o sacrificing the above I will feel this way. You can call that bias if you want to but too me it's baseball, good players maximize their skillsets and mediocre or bad players don't.

RedsManRick
05-14-2008, 11:25 PM
I do not misunderstand it at all, in fact I have it down cold. RC is a statistic that measures how many runs a player creates through his actions. The actions that has been decided on which correlate to higher runs created is OBP% and SLG% which combine for OPS%. OBP% tells us how often a player avoids an out, and SLG% a measurement of power.

Decided? Do you think there was a big stat wanker convention and they voted on the stats they liked the most? There was no "deciding". There was running analysis which showed that certain actions actually DO lead to runs scoring. This isn't a matter of opinion.



And this is where this argument always ends up and I hate it. We simply cannot get to the bottom of it w/o someone making fun of someone else's view or someone making the accusation that he is not appreciated. I know he pretty much is what he is and I don't fault anyone for lack of a specific talent it's not something they can control. But I do fault a guy for a lack of skill at times if that player can improve it and doesn't. His inability to make consistent contact has some to do with talent but also with a lack of skill. What's wrong with asking a guy with his considerable talents to increase certain skills to maximize the player he can be.

Nothing at all. What makes you think that Dunn is choosing not to improve his contact rate? Why is his inability to make more contact a choice? I've yet to see a reasonable argument to that end.

There's absolutely nothing wrong with expecting players to continue to work on increasing their abilities. What's wrong is expecting Dunn to increase his contact rate, but not expecting Freel to increase his power, not complaining that Phillips working to improve his plate discipline and to stop hitting in to so many double plays, to have Belisle working to reduce his hit rate? Why isn't Joey Votto working to run faster?

Dunn's weakness get labeled as personal failings -- as laziness. Other people merely have limited abilities. That right there IS the complaint I have. Dunn's weakness get converted to character flaws. It's simply inaccurate and not fair to Dunn the person.


You might say he's not capable of improving his contact w/o sacrificing power or plate discipline, I say I think he can. Now I am no pro scout so I could be wrong he may not be able to maximize his contact rate but until I hear someone who is qualified tell me he cannot improve the above w/o sacrificing the above I will feel this way. You can call that bias if you want to but too me it's baseball, good players maximize their skillsets and mediocre or bad players don't.

Ok. Good for you. I think that Ryan Freel can hit 30 HR. Because he only hits 5, that means he's lazy and choosing to not hit for power. Wow, that was fun.

Who prey tell do you determined as qualified? Adam Dunn can't make contact as much as most players. What evidence do you have that Dunn, or any other player, despite immense training and instruction is willfully choosing not to develop a significant portion of his skill set. If Dunn could make more contact, he would. It's in his best interest to do so. That you assume he chooses not to, but refuse to apply that logic to other players based on no evidence whatsoever simply strikes me as laughable. I guess we'll have to agree to disagree here.

Mario-Rijo
05-14-2008, 11:55 PM
Decided? Do you think there was a big stat wanker convention and they voted on the stats they liked the most? There was no "deciding". There was running analysis which showed that certain actions actually DO lead to runs scoring. This isn't a matter of opinion.



Nothing at all. What makes you think that Dunn is choosing not to improve his contact rate? Why is his inability to make more contact a choice? I've yet to see a reasonable argument to that end.

There's absolutely nothing wrong with expecting players to continue to work on increasing their abilities. What's wrong is expecting Dunn to increase his contact rate, but not expecting Freel to increase his power, not complaining that Phillips working to improve his plate discipline and to stop hitting in to so many double plays, to have Belisle working to reduce his hit rate? Why isn't Joey Votto working to run faster?

Dunn's weakness get labeled as personal failings -- as laziness. Other people merely have limited abilities. That right there IS the complaint I have. Dunn's weakness get converted to character flaws. It's simply inaccurate and not fair to Dunn the person.

Really, since when have you seen me state that Dunn is lazy. I merely have wondered if he is content and if he is then I would say he is lazy or afraid. I believe if you can get better and don't you have a character flaw, is that to say Dunn has one I don't know but I wonder.

Who hasn't complained about BP's plate discipline???? I guarantee you if I have to watch BP swing at that outside pitch off the plate that he cannot hit for 3,4,5 more years like I have watched Dunn I will have the same problem with him. I already wonder about BP's shortcomings I just feel he still has time to improve. And let me add I echo those exact sentiments with Edwin.

Freel to increase his power huh, I see what you are trying to do. Is Ryan Freel someone you think we should be building around? Guys like Freel have proven that they are part time/complimentary players and not building blocks, Dunn is and therefore I expect more out of him. That said I expect Freel to be as good as he can as well. But more power is a talent not a skill and you can't ask him to swing any harder than he already does that would likely be impossible w/o torquing his body.

Votto run faster, again a talent not a skill. And I think he runs extremely well for his position. Belisle, you may have a point there, he does need to improve. And none of these guys have been above criticism here, yet you seem to feel differently.

Mario-Rijo
05-15-2008, 12:00 AM
Decided? Do you think there was a big stat wanker convention and they voted on the stats they liked the most? There was no "deciding". There was running analysis which showed that certain actions actually DO lead to runs scoring. This isn't a matter of opinion.



Nothing at all. What makes you think that Dunn is choosing not to improve his contact rate? Why is his inability to make more contact a choice? I've yet to see a reasonable argument to that end.

There's absolutely nothing wrong with expecting players to continue to work on increasing their abilities. What's wrong is expecting Dunn to increase his contact rate, but not expecting Freel to increase his power, not complaining that Phillips working to improve his plate discipline and to stop hitting in to so many double plays, to have Belisle working to reduce his hit rate? Why isn't Joey Votto working to run faster?

Dunn's weakness get labeled as personal failings -- as laziness. Other people merely have limited abilities. That right there IS the complaint I have. Dunn's weakness get converted to character flaws. It's simply inaccurate and not fair to Dunn the person.



Ok. Good for you. I think that Ryan Freel can hit 30 HR. Because he only hits 5, that means he's lazy and choosing to not hit for power. Wow, that was fun.

Who prey tell do you determined as qualified? Adam Dunn can't make contact as much as most players. What evidence do you have that Dunn, or any other player, despite immense training and instruction is willfully choosing not to develop a significant portion of his skill set. If Dunn could make more contact, he would. It's in his best interest to do so. That you assume he chooses not to, but refuse to apply that logic to other players based on no evidence whatsoever simply strikes me as laughable. I guess we'll have to agree to disagree here.

Why all the venom????

BuckeyeRedleg
05-15-2008, 12:03 AM
I think he's frustrated he has to continually defend basic logic.

I'm glad he's doing it, because it saves me a bunch of heartache and he conveys the message much better.

Thanks, Rick. :thumbup:

RedsManRick
05-15-2008, 12:05 AM
MR, if skill is the realization of talent as you assert, what talent has Dunn displayed that suggests to you he should be capable of increasing his contact skill? I don't mean any "venom", I simply don't understand the origin of this assertion.

Mario-Rijo
05-15-2008, 12:13 AM
I think he's frustrated he has to continually defend basic logic.

I'm glad he's doing it, because it saves me a bunch of heartache and he conveys the message much better.

Thanks, Rick. :thumbup:

Wow that's pretty rude. You might as well be calling me an idiot. That's cool.

Mario-Rijo
05-15-2008, 12:22 AM
MR, if skill is the realization of talent as you assert, what talent has Dunn displayed that suggests to you he should be capable of increasing his contact skill? I don't mean any "venom", I simply don't understand the origin of this assertion.

No talent but technique. He improved last season when he got his hands down some which is what I have been calling for, for quite some time. He had regressed 2 yrs in a row before this change which suggests that the change worked. I feel if he dropped his hands more he would improve more. I see him as a LH Mark McGwire with where his hands should be. And all big leaguers hate this but why not choke up a tad in certain situations to improve his bat control.

I have already been told by a pro scout that he has bad bat control, this is why he isn't a great contact hitter. The scout gave me no indication that he couldn't improve this, although I don't expect alot that's why I expect a .270 hitter.

SteelSD
05-15-2008, 12:36 AM
You might say he's not capable of improving his contact w/o sacrificing power or plate discipline, I say I think he can. Now I am no pro scout so I could be wrong he may not be able to maximize his contact rate but until I hear someone who is qualified tell me he cannot improve the above w/o sacrificing the above I will feel this way.

I think there's something to be learned from the above passage so I want to address it. M-R, while I very much appreciate that you've made a legitimate effort to better understand the metrics often used in these discussions (seriously), you've produced a logic error.

If you can't adopt an opinion that Dunn can't improve his contact rate without first hearing from someone "qualified", then how can you adopt a position that he can without hearing from a similarly "qualified" individual?

In short, you're stating that the only people who could change your mind are those who weren't involved in the production of your original position.


I have already been told by a pro scout that he has bad bat control, this is why he isn't a great contact hitter. The scout gave me no indication that he couldn't improve this, although I don't expect alot that's why I expect a .270 hitter.

Aha. So now you've spoken with a "pro scout" about Dunn's contact rate and he gave you "no indication" that Dunn could improve his contact rate. How many "pro scouts" have you talked to who've "indicated" that Dunn should be able to improve his contact rate?

Mario-Rijo
05-15-2008, 12:44 AM
I think there's something to be learned from the above passage so I want to address it. M-R, while I very much appreciate that you've made a legitimate effort to better understand the metrics often used in these discussions (seriously), you've produced a logic error.

If you can't adopt an opinion that Dunn can't improve his contact rate without first hearing from someone "qualified", then how can you adopt a position that he can without hearing from a similarly "qualified" individual?

In short, you're stating that the only people who could change your mind are those who weren't involved in the production of your original position.

See my last post above.

Mario-Rijo
05-15-2008, 12:49 AM
Aha. So now you've spoken with a "pro scout" about Dunn's contact rate and he gave you "no indication" that Dunn could improve his contact rate. How many "pro scouts" have you talked to who've "indicated" that Dunn should be able to improve his contact rate?

No he gave me no indication that he couldn't improve it. Are you trying to act as though I am making it up because your words read like I am. I have only spoke to 1 scout and at the time I didn't get to ask him further about what he could do to improve it.

But I seen him improve his contact rate just last season and the statistics back that up.

SteelSD
05-15-2008, 01:20 AM
No he gave me no indication that he couldn't improve it. Are you trying to act as though I am making it up because your words read like I am. I have only spoke to 1 scout and at the time I didn't get to ask him further about what he could do to improve it.

Honestly, my intent wasn't to imply that you're making anything up regarding your interaction with the "pro scout".

And "no indication" that Dunn "couldn't" improve his contact rate is exactly the same thing as "no indication" that Dunn "could" improve his contact rate. You received absolutely no information about the matter from the "pro scout", but you're now using your interaction as support for your position. That's a basic error in logic. The fact that you didn't get to ask him more questions is irrelevant.


But I seen him improve his contact rate just last season and the statistics back that up.

No, you actually didn't see that. Dunn's Contact Rate on pitches from 2007 was actually lower than it was in 2005 (a season in which you think he "regressed"). In 2007, Dunn's Contact Rate was 72.00%. In 2005, it was 72.85%. For Strikes, Dunn produced similar results across all three seasons, with 2005 being his best year of the three. 81.69% strike zone contact for 2005, 79.13% in 2006, and 81.56% in 2007.

While it's unfortunate that you're using conclusion-based analysis, it is what it is. You have a conclusion and the only information you've been able to present to back it up is non-information from a "pro scout" you met for a couple of minutes. Nothing objective about that.

Mario-Rijo
05-15-2008, 01:35 AM
Honestly, my intent wasn't to imply that you're making anything up regarding your interaction with the "pro scout".

And "no indication" that Dunn "couldn't" improve his contact rate is exactly the same thing as "no indication" that Dunn "could" improve his contact rate. You received absolutely no information about the matter from the "pro scout", but you're now using your interaction as support for your position. That's a basic error in logic. The fact that you didn't get to ask him more questions is irrelevant.



No, you actually didn't see that. Dunn's Contact Rate on pitches from 2007 was actually lower than it was in 2005 (a season in which you think he "regressed"). In 2007, Dunn's Contact Rate was 72.00%. In 2005, it was 72.85%. For Strikes, Dunn produced similar results across all three seasons, with 2005 being his best year of the three. 81.69% strike zone contact for 2005, 79.13% in 2006, and 81.56% in 2007.

While it's unfortunate that you're using conclusion-based analysis, it is what it is. You have a conclusion and the only information you've been able to present to back it up is non-information from a "pro scout" you met for a couple of minutes. Nothing objective about that.

Actually he (scout) said that "He has horrible bat control. Therefore he lacks the ability to make adjustments". At the time I didn't get the rest of the story and yes I know that it's somewhat irrelevant. But since you have proven to me that his contact rate wasn't actually higher (just the end result, BA) I will be sure to get ahold of said scout and finish the discussion for the sake of all.

Ron Madden
05-15-2008, 05:13 AM
Let me first point out that this is not a shot at anyone in this thread.

We've had this same discussion for years. It has been proven that OBP and SLG are the main ingredients to Runs Scored.

It's illogical or stubborn to argue that BA and productive outs are as important to Runs Scored as OBP and SLG.


Adam Dunns OBP and SLG numbers are very good. Adam Dunn is not the problem. The problem is some Fans have unrealistic expectations of Dunn. That's not Adams fault.

BuckeyeRedleg
05-15-2008, 09:56 AM
Wow that's pretty rude. You might as well be calling me an idiot. That's cool.

My apologies for coming off rude, Mario. I did not intend for it to come across that way. Poor choice of words on my part.

It's just for years now, I have seen Cyclone, Rick, etc. explain themselves in a very detailed manner and a few posters on here just disregard it and have nothing to back up their opinions. Sometimes they even take a shot at them for being "statheads" or never watching the game. Rarely in all this time have I seen Cyclone, Rick, or anyone explaining this side get snarky. They continue to put their thoughts out there and continue to get dismissed by a select few that have little to back up their disagreement. I envy their patience and passion for explaining their thoughts.

You on the the other hand have been very thoughtful with your posts and polite and I wouldn't want to come across that way to you. I have enjoyed your contributions for many years on this board. I was just explaining my take as to why Rick might have been getting somewhat annoyed - because he keeps explaining the same thing over and over again.

Again, I really have enjoyed reading your posts for all these years and wouldn't want any hard feelings.

Sea Ray
05-15-2008, 11:07 AM
For me the frustrating thing with Dunn is that he has not improved his game since he broke in many years ago. When he first came up the thinking was he had little baseball experience , was a great athlete and would improve as he learns the game of baseball. Well he has improved precious little since his 1st or 2nd year in the league. He really shouldn't peak as a 23 yr old but some do.

I don't let it bother me so much any more. I look at it as he's established what he is and it's time for us to accept it and that is a power hitting #2 hitter who strikes out a ton but does draw lots of walks and hits horribly w/ RISP.

RedsManRick
05-15-2008, 11:29 AM
For me the frustrating thing with Dunn is that he has not improved his game since he broke in many years ago. When he first came up the thinking was he had little baseball experience , was a great athlete and would improve as he learns the game of baseball. Well he has improved precious little since his 1st or 2nd year in the league. He really shouldn't peak as a 23 yr old but some do.

I don't let it bother me so much any more. I look at it as he's established what he is and it's time for us to accept it and that is a power hitting #2 hitter who strikes out a ton but does draw lots of walks and hits horribly w/ RISP.

This is an interesting perspective Sea Ray, and quite fair I would say. The expectation placed upon Dunn by many people was based on an assumption that he would follow the standard progression.

This is really the "old skills" issue from the fan perspective. Certain skills improve over time, some regress. Unfortunately, most of Dunn's value comes from skills that usually mature over time. He happened to come out of the minor league womb fully formed. The weaknesses in his game were never likely to improve, and he arguably had reached the limit of where existing strengths could grow.

Because of that, many fans have always judged Dunn on what he "could be", failing to appreciate what he is. When a player is able to grow in to that expectation, it can be a great experience as a fan. But when the player doesn't, he's often never appreciated fully and the fan can feel robbed.

Raisor
05-15-2008, 11:34 AM
hits horribly w/ RISP.

.410 .457 .868


http://strangevehicles.greyfalcon.us/Pictures/Ritter3.png

BRM
05-15-2008, 11:36 AM
.410 .457 .868


You left out the BA.

Raisor
05-15-2008, 11:37 AM
You left out the BA.

That's because I'm Evil.

BRM
05-15-2008, 11:38 AM
That's because I'm Evil.

You also left out the K's. KRISPY...

Sea Ray
05-15-2008, 11:39 AM
.410 .457 .868




You want to put some reference on those numbers or would you rather just spend your time searching the internet for fancy graphics?

BRM
05-15-2008, 11:43 AM
You want to put some reference on those numbers or would you rather just spend your time searching the internet for fancy graphics?

That's Adam Dunn's career OBP/SLG/OPS with RISP.

Sea Ray
05-15-2008, 11:47 AM
That's Adam Dunn's career OBP/SLG/OPS with RISP.

For a year? a career? vs LHPs with a snake tatoo on Sundays?

BRM
05-15-2008, 11:48 AM
For a year? a career? vs LHPs with a snake tatoo on Sundays?

I said those are his career numbers with RISP.

jojo
05-15-2008, 11:52 AM
Here's the Reds record with Dunn: 473-537.

Here's his splits:
Reds win: .281/.425/.623 OPS: 1.048
Reds loss: .217/.336/.420 OPS: .756

To me it couldn't be clearer.....

It's Dunn's fault the Reds have lost 537 times he's been in the line up. :cool:

Factor in his poor performance as judged by the ultimate metric- (PAs-((k/productive outs)+BBwRISP*(1/BAwRISP))- and it's easy to see what the problem is...

Ultimately he's failed given his lack of elite status. I'm guessing it's work ethic related since Bonds, Griffey, Mays, Mantle, Aaron, Ruth, Pujols, AROD etc didn't have any problem making their games elite.

MWM
05-15-2008, 12:25 PM
I will say that I've also been disappointed that Adam hasn't progressed like I thought he would. His lack of progression is a fair criticism, especially on the defensive side of the ball.

Sea Ray
05-15-2008, 12:30 PM
I said those are his career numbers with RISP.

Raisor thanks you for clearing up his ambiguous post.

As we all know those numbers are walk driven and walks aren't nearly as valuable as hits where RISP are involved. To lump them in with base hits misses the forest from the trees.

Sea Ray
05-15-2008, 12:32 PM
Here's the Reds record with Dunn: 473-537.

Here's his splits:
Reds win: .281/.425/.623 OPS: 1.048
Reds loss: .217/.336/.420 OPS: .756

To me it couldn't be clearer.....

It's Dunn's fault the Reds have lost 537 times he's been in the line up. :cool:



Shows me he's had more bad days than good

Raisor
05-15-2008, 12:33 PM
As we all know those numbers are walk driven and walks aren't nearly as valuable as hits where RISP are involved. To lump them in with base hits misses the forest from the trees.

Actually, most of us know that OPS w/RISP is way more accurate.

Even earlier, you admited that OBP w/RISP was more accurate then BA.


Simple question requiring a simple answer:

Which do you think more closely correlates to actual runs scored, BA w/RISP or OBP w/RISP?

It's a simple question that you can answer in merely two or three letters, nothing more.


To actual runs scored: OBP.




Three steps forward, four steps back.

BRM
05-15-2008, 12:44 PM
It still can't figure out why the fact Dunn gets pitched around with runners on is a mark against him.

pahster
05-15-2008, 12:47 PM
It still can't figure out why the fact Dunn gets pitched around with runners on is a mark against him.

Because he's not Vlad, Bonds, Pujols, A-Rod, Ortiz, Manny Ramirez, or any other elite level player. He sucks because he's in the tier right below them. :p:

westofyou
05-15-2008, 12:52 PM
Mike Schmidt, another guy who was bad and good.... shudder


in Wins 1261 .310 .428 .644 1.072
in Losses 1140 .221 .324 .397 .721

Sea Ray
05-15-2008, 01:00 PM
Actually, most of us know that OPS w/RISP is way more accurate.

Even earlier, you admited that OBP w/RISP was more accurate then BA.



That's doesn't mean that's the only metric you consider. If it is top heavy in either hits or walks further information is needed. In Dunn's case it's top heavy in walks. For example an OBP of .450 in a guy who draws 100walks a year is not equal to an OBP of .450 in a guy who walks 10 times

Raisor
05-15-2008, 01:12 PM
For example an OBP of .450 in a guy who draws 100walks a year is not equal to an OBP of .450 in a guy who walks 10 times

Except it is, or at least it's so close as to be undistinguishable.

I posted it once in this thread. Go back and look at Sean Casey and Adam Dunn's 2004 seasons.

Here, I'll post their numbers again.

2004 Reds

Sean Casey 324/381/534 24-99
Adam Dunn 266/388/569 46-102

Dunn 124.9 RC in 681 PA (18 RC per 100 TPA)
Casey 110.6 RC in 633 PA (17 RC per 100 TPA)

RISP

Casey 304/404/529
Dunn 239/438/514

Casey 30 RC in 171 PA (17 RC per 100 TPA)
Dunn 35.5 RC in 192 PA (18 RC per 100 TPA)

BRM
05-15-2008, 01:15 PM
Casey had more non-HR RBI's though.

Mario-Rijo
05-15-2008, 03:43 PM
My apologies for coming off rude, Mario. I did not intend for it to come across that way. Poor choice of words on my part.

It's just for years now, I have seen Cyclone, Rick, etc. explain themselves in a very detailed manner and a few posters on here just disregard it and have nothing to back up their opinions. Sometimes they even take a shot at them for being "statheads" or never watching the game. Rarely in all this time have I seen Cyclone, Rick, or anyone explaining this side get snarky. They continue to put their thoughts out there and continue to get dismissed by a select few that have little to back up their disagreement. I envy their patience and passion for explaining their thoughts.

You on the the other hand have been very thoughtful with your posts and polite and I wouldn't want to come across that way to you. I have enjoyed your contributions for many years on this board. I was just explaining my take as to why Rick might have been getting somewhat annoyed - because he keeps explaining the same thing over and over again.

Again, I really have enjoyed reading your posts for all these years and wouldn't want any hard feelings.

Thanks for the response, I'm sorry for misunderstanding your post. Don't I feel like a creep now. Anyhow I don't disagree with your thoughts behind the post. I get that and appreciate what those people you mentioned and more do for the board, and I respect their opinion always. But when I see something like I have with Dunn it's going to take a little more than usual to change my mind.

I also want to apologize to others I may have offended with the "qualified" comment. I meant no disrespect to anyone, but and again no disrespect we all question things around here like mechanics and technique but how many of us really are in a position to truely answer something like that. I know i'm not although I sometimes like to think I can I will always go with a professionals word on it. It's no different than your furnace banging around and your buddy stops by and says "hey I know what's wrong with that, X,Y,Z". That's wonderful and maybe he's right but just in case I'm gonna call a pro and get it right for sure.

Sea Ray
05-15-2008, 03:59 PM
Except it is, or at least it's so close as to be undistinguishable.

I posted it once in this thread. Go back and look at Sean Casey and Adam Dunn's 2004 seasons.

Here, I'll post their numbers again.

2004 Reds

Sean Casey 324/381/534 24-99
Adam Dunn 266/388/569 46-102

Dunn 124.9 RC in 681 PA (18 RC per 100 TPA)
Casey 110.6 RC in 633 PA (17 RC per 100 TPA)

RISP

Casey 304/404/529
Dunn 239/438/514

Casey 30 RC in 171 PA (17 RC per 100 TPA)
Dunn 35.5 RC in 192 PA (18 RC per 100 TPA)

Let's look at Jeff Keppinger and Adam Dunn with RISP.

Career numbers RISP Avg/OBP/slugging/OPS:

Dunn

.220/.410/.457/868 338 RBIs in 723 games

Keppinger

.333/.389/.462/.851 53 RBI in 99 games

Here Kepp has a lower OBP but he knocked in more runs than Dunn did. How can that be? Must be that hits drive in more runs than walks. Add on to that the fact that Dunn had more opportunities per game than Kepp (1.63 PA/game compared to 1.52 for Kepp) and he still knocked in less people.

This gets back to my original point. Dunn, despite his high OPS is not very good at driving in runs.

Raisor
05-15-2008, 04:07 PM
Let's look at Jeff Keppinger and Adam Dunn with RISP.

Career numbers RISP Avg/OBP/slugging/OPS:

Dunn

.220/.410/.457/868 338 RBIs in 723 games

Keppinger

.333/.389/.462/.851 53 RBI in 99 games

Here Kepp has a lower OBP but he knocked in more runs than Dunn did. How can that be? Must be that hits drive in more runs than walks. Add on to that the fact that Dunn had more opportunities per game than Kepp (1.63 PA/game compared to 1.52 for Kepp) and he still knocked in less people.

This gets back to my original point. Dunn, despite his high OPS is not very good at driving in runs.


If RBI are what belts your buckle, so be it.

If you've decided you've learned everything there is to learn, so be it again.

I guess you know what you know.

(By the way, you're looking at a difference of about 7 RBI per 100 games in your example)

BRM
05-15-2008, 04:10 PM
Take a look at RBI per AB w/RISP. They are very close.

bucksfan2
05-15-2008, 04:15 PM
Let's look at Jeff Keppinger and Adam Dunn with RISP.

Career numbers RISP Avg/OBP/slugging/OPS:

Dunn

.220/.410/.457/868 338 RBIs in 723 games

Keppinger

.333/.389/.462/.851 53 RBI in 99 games

Here Kepp has a lower OBP but he knocked in more runs than Dunn did. How can that be? Must be that hits drive in more runs than walks. Add on to that the fact that Dunn had more opportunities per game than Kepp (1.63 PA/game compared to 1.52 for Kepp) and he still knocked in less people.

This gets back to my original point. Dunn, despite his high OBP is not very good at driving in runs.

SeaRay I agree with this assesment. I would agree that a high OBP may lead to more runs being scored in a typical inning but it isn't as big of a factor as BA is for driving in runs. It has been aruged that RBI is stat that is driven by outside forces, mainly people ahead of you getting on base. I agree that without runners on in advance of you RBI doesn't matter much but to counter that I think driving in runs is more difficult than some give it. Is there a reason that Victor Martinez is such a good hitter with RISP?

I think we all agree that Dunn's high OBP is an important thing to have on a club. The only problem I have with it is that 20% of the time when Dunn gets on base it is via the walk. While that is a good thing it doesn't help the team out as much when Bako and P follow Dunn. When you are type cast as a power hitter you would like to see the BA come up in proportion to the OBP. If Dunn were a 2 hole hitter then I don't think we would have as much of a problem with him.

Raisor
05-15-2008, 04:15 PM
Anyone want to take a shot at proving that RBI are a more accurate way to look at a player then Runs Created?

Mario-Rijo
05-15-2008, 04:36 PM
For anyone who is interested I have spoken to the scout and here's what was said.

My question to him was:

Can Adam Dunn improve his contact rate w/o sacrificing plate discipline or power, can he increase his bat control?

His response was this:

Not only can he do it, he has done it. 2 years ago he he went to a lighter bat and was tearing it up. As soon as his #'s were good he dumped the lighter bats and struggled again. And he didn't go back to what worked for him (lighter bats).

(I recall that period of him tearing it up back in mid-late july of '06. I remember a specific series in Houston where he was on an absolute tear.)

He went on to say:

Just like the last two ST's he goes the other way, stays on the ball longer during his swing etc. But as soon as the season starts he goes right back to swinging for the fences. As long as that is his approach he will not improve. The older he get's the tougher it is going to be for him to reverse field as his bat is already a hair slower, it really noticeable when he tries to pull those outside pitches.

Sea Ray
05-15-2008, 04:41 PM
SeaRay I agree with this assesment. I would agree that a high OBP may lead to more runs being scored in a typical inning but it isn't as big of a factor as BA is for driving in runs. It has been aruged that RBI is stat that is driven by outside forces, mainly people ahead of you getting on base.



In my example there were runners on base (RISP) so that part of the equation was already accounted for.

Sea Ray
05-15-2008, 04:47 PM
Anyone want to take a shot at proving that RBI are a more accurate way to look at a player then Runs Created?

Then we get into runs created vs runs produced. We all agree Dunn can be effectively used to create runs if placed properly in the lineup. My point is Dunn fits best as a #2 hitter because of what he doesn't do well. I don't know what point you're trying to make since you've not shown him to have any offensive weaknesses.

MWM
05-15-2008, 04:50 PM
MR, nothing personal, but I don't buy any of that. It makes no sense whatsoever.

Caveat Emperor
05-15-2008, 04:56 PM
Anyone want to take a shot at proving that RBI are a more accurate way to look at a player then Runs Created?

RBI = On back of baseball card.

RC = Not on back of baseball card.

It couldn't be easier, as far as I'm concerned. :p:

Sea Ray
05-15-2008, 05:18 PM
RBI = On back of baseball card.

RC = Not on back of baseball card.

It couldn't be easier, as far as I'm concerned. :p:

Before we can throw out stats like RC we have to agree on the formula. I believe there are 24 different formulas out there for RC. Folks on this site throw out a number for RC without giving any indication as to how it was figured and who did the figuring

Raisor
05-15-2008, 05:19 PM
I don't know what point you're trying to make since you've not shown him to have any offensive weaknesses.


The point I'm trying to make is that you're developing opinions without using the best tools we have available.

Look at those numbers for Casey and Dunn again. If what you were saying was true, there's no way their numbers would be as close as they were.

Mario-Rijo
05-15-2008, 05:35 PM
MR, nothing personal, but I don't buy any of that. It makes no sense whatsoever.

I gotta be honest, I really don't see how it doesn't make sense. Why do you feel that it doesn't?

Cyclone792
05-15-2008, 05:39 PM
Before we can throw out stats like RC we have to agree on the formula. I believe there are 24 different formulas out there for RC. Folks on this site throw out a number for RC without giving any indication as to how it was figured and who did the figuring

All the information is out there; you're more than welcome to research it yourself.

I've already researched the formulas, studied them, and analyzed them. Some are over two decades old, others merely a few years old. Some are more advanced and more accurate, others are more basic and slightly less accurate.

But one thing is clear, which is that all of them are more accurate than archaic stats such as BA w/RISP and RBI.

The fact that you keep harping on archaic stats tells me that you either haven't studied RC at all, or you attempted to look at it, failed to understand it and threw it away simply because you failed to understand it.

Sea Ray
05-15-2008, 05:41 PM
The point I'm trying to make is that you're developing opinions without using the best tools we have available.

Look at those numbers for Casey and Dunn again. If what you were saying was true, there's no way their numbers would be as close as they were.

Define RC. How do you arrive at it?

Raisor
05-15-2008, 05:43 PM
Define RC. How do you arrive at it?


I've been using the one that's easy to find on ESPN.com

Runs created
[(H + BB + HBP - CS - GIDP) times (Total bases + .26[BB - IBB + HBP] + .52[SH + SF + SB])] divided by (AB + BB + HBP + SH+ SF)

Sea Ray
05-15-2008, 05:43 PM
The fact that you keep harping on archaic stats tells me that you either haven't studied RC at all, or you attempted to look at it, failed to understand it and threw it away simply because you failed to understand it.

Your answer tells me you're harping on a stat you can't even define. You had a chance to define it and didn't. I told you, there are 24 different ways to define RC. Which one do you use?

Call my stats archaic if you want but I like them because they're easily defined. I don't trust stats that include more formulas than the federal tax system.

BuckeyeRedleg
05-15-2008, 05:47 PM
Basic RC =

OBP × SLG × AB

or

OBP × TB

Which correlates to actual run production way better better than BA, RBI, BA RISP, etc.

The more complex the RC formulas are, the higher the correlation to run production.

The point is that the even the most simple form of RC still blows away everything else.

Sea Ray
05-15-2008, 05:47 PM
I've been using the one that's easy to find on ESPN.com

Runs created
[(H + BB + HBP - CS - GIDP) times (Total bases + .26[BB - IBB + HBP] + .52[SH + SF + SB])] divided by (AB + BB + HBP + SH+ SF)

Why use factors .26 and .52?

BuckeyeRedleg
05-15-2008, 05:48 PM
From Wikipedia:

Runs created (RC) is a baseball statistic invented by Bill James to estimate the number of runs a hitter contributes to his team.

James explains in his Bill James Historical Baseball Abstract why runs created is an essential thing to measure:

"With regard to an offensive player, the first key question is how many runs have resulted from what he has done with the bat and on the basepaths. Willie McCovey hit .270 in his career, with 353 doubles, 46 triples, 521 home runs and 1,345 walks -- but his job was not to hit doubles, nor to hit singles, nor to hit triples, nor to draw walks or even hit home runs, but rather to put runs on the scoreboard. How many runs resulted from all of these things?"

http://en.wikipedia.org/wiki/Runs_created

MWM
05-15-2008, 05:50 PM
I gotta be honest, I really don't see how it doesn't make sense. Why do you feel that it doesn't?

It's not believable. You mean to tell me that:

- someone convinced him in to using a lighter bat
- it actually worked with him hitting much better
- then he just randomly switched back
- followed by him going to back to not hitting as well
- and he didn't back to the lighter bat

I'm not questioning you here, but your source. The whole "lighter bat" sounds like the type of thing someone would say when they're trying to sound like they know their stuff, but really don't know anything. I'd just question the credibility of this scout.

Cyclone792
05-15-2008, 05:51 PM
Your answer tells me you're harping on a stat you can't even define. You had a chance to define it and didn't. I told you, there are 24 different ways to define RC. Which one do you use?

I can't define it, huh?

A = H+BB+HBP-CS-GIDP
B = TB+.24*(BB-IBB+HBP)+.62*SB+.5*(SH+SF)-.03*SO
C = AB+BB+HBP+SH+SF

RC = (((2.4*C+A)*(3*C+B))/(9*C))-(.9*C)

Now it's your turn. Go tell me the correlation of that stat to actual runs scored for all MLB teams from 1991-2005.


Call my stats archaic if you want but I like them because they're easily defined. I don't trust stats that include more formulas than the federal tax system.

You better not ever leave your house then if that's the case.

Sea Ray
05-15-2008, 05:53 PM
Basic RC =

OBP × SLG × AB

or

OBP × TB



OK. So Adam Dunn's OBP is .380 X .515 slugging X 3475 ABs = 680 runs created over his career. Is that your formula?

Raisor
05-15-2008, 05:53 PM
By the way, in the NL central this season, RC is 98% accurate.

Yachtzee
05-15-2008, 05:54 PM
I've been using the one that's easy to find on ESPN.com

Runs created
[(H + BB + HBP - CS - GIDP) times (Total bases + .26[BB - IBB + HBP] + .52[SH + SF + SB])] divided by (AB + BB + HBP + SH+ SF)

I find it interesting that this stat hasn't taken off among people who believe in "playing the game the right way." What other stat tries to capture sacrifices (i.e. "productive outs") as a part of run creation? You've got your hits, your walks, your HBP, your bunting and sac flies,. . . they're all in there! It's a stat for everybody!

BRM
05-15-2008, 05:55 PM
By the way, in the NL central this season, RC is 98% accurate.

How accurate is BA w/RISP?

Raisor
05-15-2008, 05:56 PM
How accurate is BA w/RISP?

hush now.

BuckeyeRedleg
05-15-2008, 05:58 PM
For many years, I would not accept some of the "newer" stats. I felt people using them were just using them to sound smart and I had little desire to investigate their meaning. I would respond with my own subjective arguments and that was that. Finally, I got bored of my own stubbornness and I figured maybe I should investigate what these individuals were talking about. It started with my unfamiliarity with OPS. I did some simple research on the internet of OPS and some of the basic sabermetric measures. Everything is out there for free. Anyway, eventually this all led me to Runs Created. Runs Created led me to Bill James and that led me to pick up a copy of his latest Historical Baseball Abstract. The next thing you know I'm reading Moneyball.

What did the most for me was to do a simple correlation test. The most important goal in an offense is to score runs. I wanted to determine on my own whether OPS and RC were the true measures of a hitter's worth compared with BA and the older "bean-counting" stats. So, with the tutorials of Cyclone and rdiersin, I pulled yearly offensive totals from retrosheet.com and ran a correlation of runs scored to BA, OPS, SLG, OBP, and many versions of RC on an Excel spreadsheet. Because this type of thing is not my area of expertise, this test took me hours and hours, but I was adamant about completing the task and seeing for myself. What I found was not only fascinating but helped me tremendously to make sense of the whole BA vs. OPS (RC) argument.

In summary, instead of taking everyone's word for it, I investigated it on my own. My point, is that you make what you want of this whole thing. The information is out there. Whether you agree or disagree, it is up to you to at least do some research before claiming it is all wrong. You owe it to yourself and to the members of this fine board that you disagree with.

I'll be honest though. Some of the concepts I have not been able to research yet, but I trust in the posters that value it, based on my research of other concepts and running the correlation study.

BRM
05-15-2008, 05:59 PM
Raisor's formula doesn't contain strikeouts in it anywhere.

Raisor
05-15-2008, 06:00 PM
Raisor's formula doesn't contain strikeouts in it anywhere.

Hush now.

Sea Ray
05-15-2008, 06:19 PM
I can't define it, huh?

A = H+BB+HBP-CS-GIDP
B = TB+.24*(BB-IBB+HBP)+.62*SB+.5*(SH+SF)-.03*SO
C = AB+BB+HBP+SH+SF

RC = (((2.4*C+A)*(3*C+B))/(9*C))-(.9*C)

Now it's your turn. Go tell me the correlation of that stat to actual runs scored for all MLB teams from 1991-2005.




It's your stat. I don't have to explain any of it. OK, we've got three definitions of RC so far. We've got 21 to go. :D

You're making my point beautifully. Your RC is not equal to Raisor's. It's not that it's useless but at least numbers like RBI and BA are static and don't change from person to person.

I think RC can be a very useful stat when evaluating the overall value of a player but it has its limitations when discussing specific skills of a player

Cyclone792
05-15-2008, 06:21 PM
It's your stat. I don't have to explain any of it. OK, we've got three definitions of RC so far. We've got 21 to go. :D

You're making my point beautifully. Your RC is not equal to Raisor's. It's not that it's useless but at least numbers like RBI and BA are static and don't change from person to person.

I think RC can be a very useful stat when evaluating the overall value of a player but it has its limitations when discussing specific skills of a player

I'm still waiting.


Now it's your turn. Go tell me the correlation of that stat to actual runs scored for all MLB teams from 1991-2005.

Let's see what you've got, big boy. Or do you not know how to do it?

BRM
05-15-2008, 06:23 PM
Don't the different variations of RC all arrive at pretty close to the same numbers anyway? Or am I mistaken?

Cyclone792
05-15-2008, 06:24 PM
Don't the different variations of RC all arrive at pretty close to the same numbers anyway?

We can let Sea Ray tell us.

Because once he runs the correlation for the version I listed above, he can then run the correlations for the version Raisor listed and the correlations for the version that BuckeyeRedleg listed.

Then he'll be able to answer that question for you, right Sea Ray?

Raisor
05-15-2008, 06:29 PM
I'll even help.

For the Reds, this year

The ESPN formula comes to 194.7 RC
The "simple" RC formula (OBP*SLG*AB) comes to 192.8

We're talking less then 2 runs over 41 games.

The Reds vs the ESPN RC
Actual RS-182
ESPN RC-194.7

Difference of about 0.30 Runs/Game.

Mario-Rijo
05-15-2008, 06:32 PM
It's not believable. You mean to tell me that:

- someone convinced him in to using a lighter bat
- it actually worked with him hitting much better
- then he just randomly switched back
- followed by him going to back to not hitting as well
- and he didn't back to the lighter bat

I'm not questioning you here, but your source. The whole "lighter bat" sounds like the type of thing someone would say when they're trying to sound like they know their stuff, but really don't know anything. I'd just question the credibility of this scout.

Fair enough I can understand how that might seem like an unlikely set of events. That said this scout's credibility I don't worry about.

However it doesn't seem that far fetched to me with regards to Dunn. I remember when Dunn was rumored around here to have gone to a lighter bat briefly and it coencided with him going the other way alot more (again that Houston series in late July of '06 and that whole week or 2), hitting everything but just wasn't hitting for as much power. And then he fell off again sometime thereafter. That was Dunn's worst season since he wasn't considered a youngster anymore. Perhaps his struggles and the media and fans complaining about it brought on the idea of trying something new for awhile just to get everyone off his back. As soon as he felt comfortable where he was he went back to what he liked to do.

I for one think Dunn is content with who he is as a player and doesn't really feel like changing anything would bring that much of a difference. I agree that most of the time he can be the version of Adam Dunn we all know (offensively speaking) and I would be ok with him talking a BB or swinging for the fences. But there are times in a game where I would like to see a bit different approach.

An often used example is late in the game (7th-9th+) and with runners on base. And that is one time I agree with. But just to throw something else out there how about this?

Mid-May, top of the 2nd 0 outs down 2-0, runners on 1st and 2nd. Belisle's on the hill for us, we are playing a division foe on the road. Dunn's up and the defense shifts for him to pull the ball. The pitcher is throwing Dunn strikes, so what do you want him to do here? He's likely swinging for the fences since he isn't likely to draw a BB here, but it's a golden opportunity (and therefore better odds) to go the other way.

Sea Ray
05-15-2008, 06:32 PM
I've answered everything up to this point. Anything further would be repetitive. I've made my points. If you have further points to make have it!

Cyclone792
05-15-2008, 06:34 PM
I've answered everything up to this point. Anything further would be repetitive.

I'm still waiting.

If you can't answer what I've asked you above, then my original premise will hold true, which is:


The fact that you keep harping on archaic stats tells me that you either haven't studied RC at all, or you attempted to look at it, failed to understand it and threw it away simply because you failed to understand it.

Because quite honestly, if you've studied RC then you would know it would take all of about five minutes - not even five minutes - to accomplish what I asked. And I know you've spent more than five minutes in this thread in just the last hour.

Raisor
05-15-2008, 06:35 PM
I've answered everything up to this point. Anything further would be repetitive. I've made my points. If you have further points to make have it!

So what problem do you have with the ESPN RC stat? It's easy to use, there's no math involved since you can just look it up.

Caveat Emperor
05-15-2008, 06:37 PM
It's not that it's useless but at least numbers like RBI and BA are static and don't change from person to person.

That's why RBI is referred to as a "counting stat" -- it merely "counts" the occurrences of something (in this case, runs batted in). Counting stats (HR, BB, H, SB, K, AB, etc.) are the equivalent of baseball constants -- they don't change based on the metric used, they just "are".

And, I don't get the argument that RC is useless because there are multiple ways to get to the number -- if the number has validity and correlates significantly with run production, why is the formula an issue?

Furthermore, why is a stat with less validity "better" because it's simpler to use?

Sea Ray
05-15-2008, 06:43 PM
At the risk of being repetitive I'll remind you one more time what I said earlier:


I think RC can be a very useful stat when evaluating the overall value of a player but it has its limitations when discussing specific skills of a player

As to the ESPN formula it is one of many. Who's to say that formula is better or worse than someone else's?

I'm done discussing semantics. If we get back to discussing a specific player and a specific skill we can revisit various methods to use. I've made my points re: Adam Dunn here and I've done it in a way without namecalling like "big boy".

Cyclone792
05-15-2008, 06:45 PM
At the risk of being repetitive I'll remind you one more time what I said earlier:

As to the ESPN formula it is one of many. Who's to say that formula is better or worse than someone else's?

I'm done discussing semantics. If we get back to discussing a specific player and a specific skill we can revisit various methods to use. I've made my points re: Adam Dunn here and I've done it in a way without namecalling like "big boy".

You're trying to persuade people that a statistic in which A) you do not understand, and B) you refuse to even run a simple correlation test on, has limitations?

If you don't understand the statistic in the first place, then how are you reaching the conclusion that it has limitations?

And, BTW, you still haven't run those correlations.

Caveat Emperor
05-15-2008, 06:47 PM
As to the ESPN formula it is one of many. Who's to say that formula is better or worse than someone else's?

Basic statistical analysis can answer that question -- run the regression and determine which statistic correlates most highly with run production.

I'd wager a large sum of money that every single RC formula spits out a number with a higher correlation-coefficient than BA, RBI or BA w/ RISP.

Raisor
05-15-2008, 06:50 PM
At the risk of being repetitive I'll remind you one more time what I said earlier:



As to the ESPN formula it is one of many. Who's to say that formula is better or worse than someone else's?

.

You do it by looking at results.

Up above, I ran the numbers for the NL Central.

The ESPN RC formula is 98% accurate this year in the NL Central.

That should give you an idea how accurate it is.

jojo
05-15-2008, 08:53 PM
Before we can throw out stats like RC we have to agree on the formula. I believe there are 24 different formulas out there for RC. Folks on this site throw out a number for RC without giving any indication as to how it was figured and who did the figuring

That's not true.....

http://www.redszone.com/forums/showpost.php?p=1248453&postcount=354

http://www.redszone.com/forums/showpost.php?p=1268045&postcount=380

BTW, THT incorporates "clutch" into their calculations. I'd think you'd be all over that....

jojo
05-15-2008, 09:18 PM
The fact that you keep harping on archaic stats tells me that you either haven't studied RC at all, or you attempted to look at it, failed to understand it and threw it away simply because you failed to understand it.

Ignoring the metrics for a moment, I also have issues with the very narrow definition of "run producer" being applied to Dunn by some. Defining run production only by RISP splits ignores what Dunn does in roughly 75% of his PA's. For instance here's splits that summarize his whole career:

Bases empty: .250/.355/.522 OPS: .877; RC (simple formula)= 362;
Runners on: .243/.407/.507 OPS: .914; RC (simple formula)= 314 (162 were created with RISP);

That's roughly 68 wins worth of production as a Red. So why frame the debate by focusing on less than 25% of his production??????

ochre
05-15-2008, 09:37 PM
This is always the first card that's played when someone's POV has overwhelmingly been show to lack merit. It's like the sun rising in the east.
we call that 'ad hominem' around these parts, and that ain't no fancy Italian name for corn that's been soaked in lye.

Raisor
05-15-2008, 09:37 PM
2004 Reds

Sean Casey 324/381/534 24-99
Adam Dunn 266/388/569 46-102

Dunn 124.9 RC in 681 PA (18 RC per 100 TPA)
Casey 110.6 RC in 633 PA (17 RC per 100 TPA)

RISP

Casey 304/404/529
Dunn 239/438/514

Casey 30 RC in 171 PA (17 RC per 100 TPA)
Dunn 35.5 RC in 192 PA (18 RC per 100 TPA)

Still waiting for someone to tell me how this was possible.

How could they have had essentially the same offensive season while Casey hit 60 points higher then Dunn in batting average (and 65 points higher w/RISP).

ochre
05-15-2008, 09:39 PM
To actual runs scored: OBP.

Now, here's one for you, if he were more adept at driving in runs would you still advocate hitting him 2nd?

(Talk about crickets)
If he were more adept at driving in runs the aggregate team OBP would likely be higher and some of the constituents of said aggregation might present themselves as adequate candidates for duty as the team's #2 hitter. Having said that, he'd be a fine candidate to hit second, regardless of his RBI totals.

ochre
05-15-2008, 09:46 PM
Because hits drive in runs. Walks rarely do
runs and rbis are functions of the team as a collective organism. Other than solo home runs (which are likely statistically insignificant under most conditions), teams score runs through a cooperative effort of avoiding outs and acquiring bases. A team has, outside of extra innings games, twenty seven outs to acquire as many bases as they can. Individuals that are able to maximize the teams Plate Appearances/game by avoiding outs tend to allow for the team, as an organism, to have the best chance at optimal performance. In the big picture (which is made up of hundreds and thousands of small pictures) the players that make the lowest rate of outs per PA are groovy.

ochre
05-15-2008, 09:46 PM
I'm still waiting.

If you can't answer what I've asked you above, then my original premise will hold true, which is:



Because quite honestly, if you've studied RC then you would know it would take all of about five minutes - not even five minutes - to accomplish what I asked. And I know you've spent more than five minutes in this thread in just the last hour.
that's (five minutes) only true unless one is using excel and trying to break it down into percentiles. Then one has to figure out where MS messed up their percentile formulas. That takes way longer then five minutes I hear... :)

jojo
05-15-2008, 09:52 PM
For many years, I would not accept some of the "newer" stats. I felt people using them were just using them to sound smart and I had little desire to investigate their meaning.

That's a sentiment that crops up pretty often (a lot in PM's actually).

But really when an opinion (argument) is supported by laying out the premises (stats in many cases) explicitly, it actually should help stimulate discussion. When readers can see the reasons behind a poster's opinion, it should be much easier to critique the opinion because the flaws/assumptions/biases/mistakes etc are naked for all to see.

To me it's not about being pompous or trying to look smart or win an argument or bludgeon others with an avalanche of stats. It's about laying my reasoning out there so that others can find my goofs or stimulate me to think about things in a new way by sharing their perspective (which I might not have ever considered on my own). Basically, it's about becoming smarter through other's comments and questions. Hopefully it's a reciprocal process.

That's why this place is so cool. There are plenty of stat guys to mind the P's and Q's and plenty of "eye" guys to raise a flag when the stats aren't jiving with the crack of the bat.

That said, I can sympathize with those who are put to sleep by too much stat-based discussion. Everybody in the ORG is capable of becoming an expert stathead. Not everybody cares to though.

edabbs44
05-15-2008, 09:56 PM
Still waiting for someone to tell me how this was possible.

How could they have had essentially the same offensive season while Casey hit 60 points higher then Dunn in batting average (and 65 points higher w/RISP).

Gee Mr. Wizard, do tell. :)

Raisor
05-15-2008, 09:57 PM
Everybody in the ORG is capable of becoming an expert stathead. .

I drool on the keyboard too much.

jojo
05-15-2008, 09:59 PM
I drool on the keyboard too much.

I've sometimes wondered if you type with your tongue.... :cool:

GAC
05-15-2008, 10:19 PM
http://news.enquirer.com/apps/pbcs.dll/article?AID=/AB/20080511/SPT04/805110428/

Preaching to the choir on here Rick. We all get it.

Email the article to our illustrious manager at ihavenoclue@geewhizmail.com :lol:

pahster
05-15-2008, 10:27 PM
Basic statistical analysis can answer that question -- run the regression and determine which statistic correlates most highly with run production.


Regression doesn't test correlation, it seeks to explain the variance caused by any number of independent variables observed in a single dependent variable.

jojo
05-15-2008, 10:35 PM
Preaching to the choir on here Rick. We all get it.

Email the article to our illustrious manager at ihavenoclue@geewhizmail.com :lol:

Unfortunately Bruce's "on base" average is only 32 points higher than his batting average.... he'll need more time in AAA.... :cool:

BRM
05-16-2008, 12:28 AM
Still waiting for someone to tell me how this was possible.

How could they have had essentially the same offensive season while Casey hit 60 points higher then Dunn in batting average (and 65 points higher w/RISP).

Casey was more productive. He drove in more of his teammates than Dunn did. And with far fewer K's to boot.

OnBaseMachine
05-16-2008, 12:31 AM
Casey was more productive. He drove in more of his teammates than Dunn did. And with far fewer K's to boot.

And Sean Casey was a better hugger.

Raisor
05-16-2008, 12:36 AM
Casey was more productive. He drove in more of his teammates than Dunn did. And with far fewer K's to boot.

Hush

KronoRed
05-16-2008, 12:38 AM
Don't forget all the hopping.

SteelSD
05-16-2008, 01:01 AM
It's not believable. You mean to tell me that:

- someone convinced him in to using a lighter bat
- it actually worked with him hitting much better
- then he just randomly switched back
- followed by him going to back to not hitting as well
- and he didn't back to the lighter bat

I'm not questioning you here, but your source. The whole "lighter bat" sounds like the type of thing someone would say when they're trying to sound like they know their stuff, but really don't know anything. I'd just question the credibility of this scout.

Question the credibility? You're a generous guy, bud. And you're entirely correct on how unlikely that scenario is considering how superstitious baseball players tend to be. Heck, they won't wash their socks during a hot streak for fear of jinxing themselves. Wade Boggs probably ate chicken about 2,400 times during his 18-year career...just during the season.

But even assuming that said unlikely scenario occurred, here's what we're hearing:

1. In 2006 (two years ago), Adam Dunn switched to a lighter bat, which allegedly improved his contact rate versus his prior performance. This transition should have improved both his Zone (Z-Contact%) and Out of Zone (O-Contact%) contact rates due to improved bat control. The combined result should be an overall improvement in Contact rate (Contact%) for 2006 versus his previous "heavy bat" season.

2. Adam Dunn then dropped the lighter bat in favor of his previous heavier bat. This resulted in a drop in his contact rates; which should be reflected in his Z-Contact%, O-Contact%, and Contact% after he switched back.

Well, anyone who's able to access fangraphs.com and enter "Adam Dunn" in the search engine would know better. Here are the numbers:

2005:

Z-Contact%: 81.69%
O-Contact%: 36.41%
Contact%: 72.85%

2006:

Z-Contact%: 79.31%
O-Contact%: 38.11%
Contact%: 70.39%

2007:

Z-Contact%: 81.56%
O-Contact%: 44.48%
Contact%: 72.00%

During the season in which Dunn allegedly used the "lighter" bat to produce more effective contact rates, he did nothing of the sort. His contact rate with strikes (Z-Contact) went down. His overall contact rate (70.39%) went down. His O-Zone contact rate went up, but if we're to attribute that to a "lighter" bat, then why would his O-Zone% go UP in 2007 while using a heavier bat? In fact, why would all three contact rate metrics actually improve versus 2006 while Dunn was using a bat that was, allegedly, hurting his performance prior to 2006? And why didn't that heavier bat in 2005 result in lower contact rates in 2005 versus 2006?

This kind of thing is why any kind of subjective "expert" analysis should be taken with a block of salt. Not only was the scenario presented unlikely, but the subjective analysis simply doesn't mesh with the objective data we have at our fingertips. If I were Mario-Rijo, I'd go back the the "pro scout" and ask him why the data doesn't support the story he's been fed by the "pro scout". And if I were Mario-Rijo, I probably wouldn't ask the question very nicely given the data we're seeing because it means that the "pro scout" pretty much hung M-R out to dry.

*BaseClogger*
05-16-2008, 01:12 AM
To be fair Steel, this is what M-R said:


Not only can he do it, he has done it. 2 years ago he he went to a lighter bat and was tearing it up. As soon as his #'s were good he dumped the lighter bats and struggled again. And he didn't go back to what worked for him (lighter bats).

(I recall that period of him tearing it up back in mid-late july of '06. I remember a specific series in Houston where he was on an absolute tear.)

This implies that he didn't use the lighter bat the entire season, which means you will need to find data for that "hot streak" around the middle of the season...

BRM
05-16-2008, 01:31 AM
Hush

Raisor is no fan of sarcasm.

SteelSD
05-16-2008, 01:32 AM
This implies that he didn't use the lighter bat the entire season, which means you will need to find data for that "hot streak" around the middle of the season...

To be really fair, M-R's "source" would need to provide those contact rates with the alleged "lighter" bat and then need provide some type of correlation between said "lighter" bat usage and contact rates versus the larger sample (both before and after).

But I suspect that M-R's "pro scout" isn't doing anything but using subjective hindsight analysis in an effort to explain overall performance variances. I highly doubt that his "pro scout" has the ability or the interest to get that granular. Nor do I have any confidence that historical data was used.

The problem with the "analysis" M-R provided is that "hot streaks" can happen at any point during the season. For any player. Without objective data, we tend to attribute those steaks to any possible variance in player behavior, equipment, mindset, what have you. But sans data to support such a contention, all we have is a subjective opinion from someone who isn't busy enough to be on M-R's cell phone speed dial, but who also was so busy that he couldn't finish a conversation M-R started with him previously.

It's good that you chimed in, because what you're seeing is nothing but a fallacious argument in the form of an appeal to authority. Yet the "authority" hasn't been named and thus hasn't been properly vetted. In short, the "pro scout" could be the worst scout in all creation who knows little about proper evaluation techniques, knows even less about how to properly support his subjective opinions via simple objective analysis, and is feeding M-R nothing but what he wants to hear.

But he's allegedly a "pro scout", so that's all one needs if they're appealing to authority. Let's not consider that said "pro scout" just might suck on all levels.

bucksfan2
05-16-2008, 09:26 AM
I've been using the one that's easy to find on ESPN.com

Runs created
[(H + BB + HBP - CS - GIDP) times (Total bases + .26[BB - IBB + HBP] + .52[SH + SF + SB])] divided by (AB + BB + HBP + SH+ SF)

Stat guys please explain this to me.

I don't want to rehash this entire discussion but a stat like his has always puzzled me. The problem I have with a stat as detailed as this one is there are two arbitrary numbers in this formula. There are two numbers that have no correlation to what actually happens on the field. I can understand obp, ops, slg, stats like this that what actually happens on the field. But when you throw in .52 and .26 IMO you lose the integrity of the stat.

Chip R
05-16-2008, 10:00 AM
This thread reminds me of something.

Highlifeman21
05-16-2008, 10:15 AM
To be really fair, M-R's "source" would need to provide those contact rates with the alleged "lighter" bat and then need provide some type of correlation between said "lighter" bat usage and contact rates versus the larger sample (both before and after).

But I suspect that M-R's "pro scout" isn't doing anything but using subjective hindsight analysis in an effort to explain overall performance variances. I highly doubt that his "pro scout" has the ability or the interest to get that granular. Nor do I have any confidence that historical data was used.

The problem with the "analysis" M-R provided is that "hot streaks" can happen at any point during the season. For any player. Without objective data, we tend to attribute those steaks to any possible variance in player behavior, equipment, mindset, what have you. But sans data to support such a contention, all we have is a subjective opinion from someone who isn't busy enough to be on M-R's cell phone speed dial, but who also was so busy that he couldn't finish a conversation M-R started with him previously.

It's good that you chimed in, because what you're seeing is nothing but a fallacious argument in the form of an appeal to authority. Yet the "authority" hasn't been named and thus hasn't been properly vetted. In short, the "pro scout" could be the worst scout in all creation who knows little about proper evaluation techniques, knows even less about how to properly support his subjective opinions via simple objective analysis, and is feeding M-R nothing but what he wants to hear.

But he's allegedly a "pro scout", so that's all one needs if they're appealing to authority. Let's not consider that said "pro scout" just might suck on all levels.

At least we weren't given bat model numbers....

I found this with a simple google search....

http://www.baseball-fever.com/showpost.php?p=1001625&postcount=2

The full thread can be found here.... http://www.baseball-fever.com/showthread.php?t=67654



Anyone that wants to harp on Dunn's 2007 numbers can point to the simple fact that he switched to Jeff Conine's model C353... ;)

ochre
05-16-2008, 10:28 AM
Stat guys please explain this to me.

I don't want to rehash this entire discussion but a stat like his has always puzzled me. The problem I have with a stat as detailed as this one is there are two arbitrary numbers in this formula. There are two numbers that have no correlation to what actually happens on the field. I can understand obp, ops, slg, stats like this that what actually happens on the field. But when you throw in .52 and .26 IMO you lose the integrity of the stat.
Those are weights that balance the relative values of those events. From what I understand they have been fairly finely tuned to the point that the correlation between the results of that formula and actual run production is pretty tight.

jojo
05-16-2008, 10:31 AM
Those are weights that balance the relative values of those events. From what I understand they have been fairly finely tuned to the point that the correlation between the results of that formula and actual run production is pretty tight.

Yes, it's James mimicking linear weights.

Yachtzee
05-16-2008, 11:02 AM
This thread reminds me of something.

I'm sorry, I thought it was "abuse." ;)

pahster
05-16-2008, 11:10 AM
Those are weights that balance the relative values of those events. From what I understand they have been fairly finely tuned to the point that the correlation between the results of that formula and actual run production is pretty tight.

I haven't read any of his abstracts. How did he come up with the weighted constants? Are they coefficients?

Raisor
05-16-2008, 11:38 AM
I got a PM last night that I'm going to post. I'll take the name off of it, but it fits with our discussion.


Hi,


You asked a question in this thread, "John Erardi Gets It -- "Who's Counting""

"How could they (Dunn and Casey in 2004) have had essentially the same offensive season while Casey hit 60 points higher then Dunn in batting average (and 65 points..."

You used these numbers.

Sean Casey 324/381/534 24-99
Adam Dunn 266/388/569 46-102

Dunn 124.9 RC in 681 PA (18 RC per 100 TPA)
Casey 110.6 RC in 633 PA (17 RC per 100 TPA)

RISP

Casey 304/404/529
Dunn 239/438/514

Casey 30 RC in 171 PA (17 RC per 100 TPA)
Dunn 35.5 RC in 192 PA (18 RC per 100 TPA)

You are asking how a player who has an OBP driven by BA can have the same RC as a player whose OBP is driven by walks.

Well here is my response.

It is very simple. The Runs Created formula you used treats walks the same as hits. You are playing a fixed game. You are using a stat that equates walks and hits to show that walks and hits produce the same results. Using your formula, players with similar OPS will always have similar RC's.

Also, while Runs Created is very accurate when applied to teams, the version you used, has a flaw when applied to individuals. This has been known for at least five years.
The formula assumes that everyone else on the team has the same OBP and SLG as the individual in question. So, according to the formula you used, a player with a higher OPS will create more runs than a player with a lower OPS, even when that is not the case.
The Runs created formula you used unfairly favors OBP and SLG. It favors sluggers like Dunn, and hurts single hitters like Casey. Therefore, you can not fairly use it to make a point about how which is more productive, OBP or BA.

bucksfan2
05-16-2008, 11:38 AM
Those are weights that balance the relative values of those events. From what I understand they have been fairly finely tuned to the point that the correlation between the results of that formula and actual run production is pretty tight.

I had HS Algebra quite a few years ago but could you say in other words; James has used numbers (linear weights) in order to manipulate a formula in order to get a desired return?

BRM
05-16-2008, 11:41 AM
Does the formula really favor OBP and SLG? Or is it that OBP and SLG just have a higher correlation to run production which in turn makes them "favorable"?

westofyou
05-16-2008, 11:43 AM
Some critics point to the ecological fallacy when discussing the validity of runs created (or any run estimator, for that matter). Just because a formula does a good job of estimating team runs (so the argument goes), that does not mean it can be accurately applied to individual production. However, this may be a misuse of the ecological fallacy; while that fallacy does deal with incorrect assumptions being made about individuals based on group data, the actions it often applies to are purely individual in nature—voting distribution in a community is a common example. Since each person makes their choice independent of everyone else, voting data at the community level is a very different function than voting at the individual level. But run scoring is, by nature, a team process. And individual run creation is essentially the same function as team run scoring, except on a smaller scale, because all of things that make teams score runs (getting on base and driving runners in) are being done by individuals. So, because team runs and player runs created are essentially measuring the same thing, the ecological fallacy is generally seen as not applicable to run estimation formulas. However, the ecological fallacy aside, there are still other significant problems with runs created.

While even the simplest version of Runs Created estimates team runs with reasonable accuracy, the multiplicative (A*B)/C structure of the formula is fundamentally flawed when estimating the runs produced by each individual hitter, particularly in the case of hitters with extremely high on-base and slugging percentages. The reason for this is that it is impossible for a player to get on base and then drive himself in -- players' on-base and slugging averages must interact with those of their teammates. Yet RC's simple OBP*TB form assumes that a player's own slugging is interacting with his own on-base percentage, which artificially inflates RC for players who score well in both categories.

Take an example: in isolation, Ryan Howard's on-base percentage and slugging average each have a real, discrete effect on the Philadelphia Phillies' offense, but when combined they overstate Howard's contribution by treating it as though he is both driving in players with equal on-base ability as himself, and simultaneously being driven in by players with equal slugging ability as himself. This model would be appropriate with regard to a theoretical lineup of nine Ryan Howards, each of whose on-base and slugging abilities would interact in precisely this manner; however, Howard is in a lineup with players of lesser on-base and slugging abilities—his actual contribution to the Phillies in terms of runs is influenced by the fact that some of his on-base skills are being wasted by teammates who lack his slugging ability, and that some of his slugging skills are being wasted by teammates who lack his on-base ability. Therefore, Howard's RC production must be adjusted downward to reflect this reality.

This is generally not a major issue for most players, as their OBPs and SLGs are not high enough to significantly distort their Runs Created; however, superstars who put up impressive OBPs and SLGs will frequently see their RC artificially inflated by this phenomenon. In recent years, James has modified the Runs Created to correct this error, effectively placing a player in a lineup of average players, rather than assume that a player's own slugging is interacting with his own on-base percentage.

Runs created does not take into account the stadiums in which a player hits. Certain stadiums, such as Denver's Coors Field prior to the introduction of the baseball humidor, generally increase offensive production in games played there. Since each run scored in such stadiums is less valuable, the same number of runs created will translate into fewer wins in a stadium like Coors than it would elsewhere.

Runs created also does not take into account the era in which a player played. Due to various factors, some eras of baseball history have had lower or higher average levels of offensive production.

BRM
05-16-2008, 11:50 AM
Thanks for that explanation WOY. This has definitely been an educational thread.

ochre
05-16-2008, 11:54 AM
I had HS Algebra quite a few years ago but could you say in other words; James has used numbers (linear weights) in order to manipulate a formula in order to get a desired return?
You could, but manipulate is a loaded word.

More appropriately, I think, one could say that we know there are disparate values associated with some of these events, but nobody knows what those values exactly are. So, in the "B" component of this compound statistic, you have TB + some other stuff. What this is saying is that this 'some other stuff' isn't quite worth what a TB event is worth, but it does contribute to the team goals. So, a walk type event that is unintentional is worth, roughly, a quarter of a TB event. A Sacrifice, or a SB is worth, roughly, half of a TB event. It's applying a weight factor, which can be seen as a bit arbitrary, but is something that needs to be done, as common sense tells us we can't count those oranges as apples.

For the most part, I don't get too caught up in Runs Created. It's pretty advanced as far a correlation, but OPS isn't too much worse, while being much more readily available (to calculate) from the program, or scoreboard :).

MWM
05-16-2008, 12:06 PM
It is very simple. The Runs Created formula you used treats walks the same as hits. You are playing a fixed game. You are using a stat that equates walks and hits to show that walks and hits produce the same results. Using your formula, players with similar OPS will always have similar RC's.

If this were really a problem, the formulate wouldn't work so well. You'd see more variation and it could be tied back to hits versus walks. But you don't see that (at least I'm not aware of it) and I'm sure it's been tested. And while most would take a hit over a walk, in most cases, the result ins't that much better. When no one is on base, a walk produces the same result as a single, and in some ways better because the pitcher probably had to throw more pitches. If a runner doesn't advance from 1st to third on a single, it's the same result. Now there are plenty of instances with runners on where it produces a better result, but it's probably not enough to be a significant influencer on the RC formula.

edabbs44
05-16-2008, 12:06 PM
A problem I have with a stat like "Runs created" is that it gives me the same feeling that passing yards, TDs or even points scored does in football. It is missing context.

The stat obviously has value. It obviously correlates to runs. That's wonderful.

But if someone goes 4-4 in a game with 2 doubles and 2 singles but the team gets shutout, he created zero runs in that game no matter what any formula tells you.

Context is what is missing from these equations. Watching a QB throw for 60 yds and a TD in the last drive of a blowout game puts points on the board and pads stats, but has zero direct correlation to winning a game. Watching Marcus Allen score 15 one yard TDs looks great on paper, but doesn't do much otherwise.

The same goes for looking at walks in a vacuum. Doubles in a vacuum. Even HRs in a vacuum. Hitting a solo HR off of Mike Lincoln in a 10-2 blowout equates to about 1/10th (or less) of the value of hitting a 2 run game winning single off of Mariano Rivera, even if that HR "created" more runs than the single.

The stat has value, but it is difficult to use it as the measurement to end all measurements of a player.

bucksfan2
05-16-2008, 12:12 PM
A problem I have with a stat like "Runs created" is that it gives me the same feeling that passing yards, TDs or even points scored does in football. It is missing context.

The stat obviously has value. It obviously correlates to runs. That's wonderful.

But if someone goes 4-4 in a game with 2 doubles and 2 singles but the team gets shutout, he created zero runs in that game no matter what any formula tells you.

Context is what is missing from these equations. Watching a QB throw for 60 yds and a TD in the last drive of a blowout game puts points on the board and pads stats, but has zero direct correlation to winning a game. Watching Marcus Allen score 15 one yard TDs looks great on paper, but doesn't do much otherwise.

The same goes for looking at walks in a vacuum. Doubles in a vacuum. Even HRs in a vacuum. Hitting a solo HR off of Mike Lincoln in a 10-2 blowout equates to about 1/10th (or less) of the value of hitting a 2 run game winning single off of Mariano Rivera, even if that HR "created" more runs than the single.

The stat has value, but it is difficult to use it as the measurement to end all measurements of a player.

Well said :thumbup:

Cyclone792
05-16-2008, 12:14 PM
Does the formula really favor OBP and SLG? Or is it that OBP and SLG just have a higher correlation to run production which in turn makes them "favorable"?

The earlier versions did for star players, as pointed out by woy's explanation, though the recent versions have cleaned it up for the most part as described below.


This is generally not a major issue for most players, as their OBPs and SLGs are not high enough to significantly distort their Runs Created; however, superstars who put up impressive OBPs and SLGs will frequently see their RC artificially inflated by this phenomenon. In recent years, James has modified the Runs Created to correct this error, effectively placing a player in a lineup of average players, rather than assume that a player's own slugging is interacting with his own on-base percentage.

If you get right down to it, there's really only two or three players on a career level where RC now may be slightly overstating their offensive contributions, and those players are Babe Ruth, Ted Williams, and Barry Bonds. On a single season level, there may be a couple additional seasons by other players where that might be the case ... such as Mickey Mantle's 1957 season. We're pretty much talking about maybe 0.01 percent of players in baseball history where this is the case though.

What's interesting is even some of the "shortcomings" with raw RC can be adjusted very easily, such as the ones described below.


Runs created does not take into account the stadiums in which a player hits. Certain stadiums, such as Denver's Coors Field prior to the introduction of the baseball humidor, generally increase offensive production in games played there. Since each run scored in such stadiums is less valuable, the same number of runs created will translate into fewer wins in a stadium like Coors than it would elsewhere.

Runs created also does not take into account the era in which a player played. Due to various factors, some eras of baseball history have had lower or higher average levels of offensive production.

It's pretty easy to both park adjust and era adjust RC, especially once you convert it over to RC/27. One of the quick and easy ways to compare players across eras via an accurate method is to divide their RC/27 to their league's RC/27. An example:

Player A: 8.50 RC/27 in a league with 5.50 RC/27 = 1.55 RQ (RC quotient)
Player B: 7.50 RC/27 in a league with 4.75 RC/27 = 1.58 RQ (RC quotient)

In raw RC, Player A produced more, but he also played in a more offensive environment (think 1990s for Player A vs. 1950s for Player B, for example). Once you adjust for league you see that they created runs at a pretty similar level to each other relative to their respective peers.

Then you can take that league adjusted RC/27 and throw a park adjustment on there.

MWM
05-16-2008, 12:16 PM
A problem I have with a stat like "Runs created" is that it gives me the same feeling that passing yards, TDs or even points scored does in football. It is missing context.

I hear arguments like this all the time, but they don't make sense to me. Context is more important in the micro. If context were so important, then why is the formula so accurate in predicting runs scored. When it's all said and done, it is incredibly accurate which trumps all the other contextual arguments. And if you break down the formula, it makes sense. It's basically measure how many bases the team gets per time they come to the plate. And the fact that the team RC is just a compilation of all the singles, double, triples, walks, HRs, HBP, etc..., it makes sense that an individual hitter's overall RC number is an accurate representation of their individual contribution to runs scored.

We can argue all the problems with it, but it works so well, I don't think it becomes difficult to do so. It's not perfect, but it's pretty darn good.

MWM
05-16-2008, 12:19 PM
The stat obviously has value. It obviously correlates to runs. That's wonderful.


You're bringing up a single situation that happens rarely. This always happens. With the formula being so accurate over the course of a season, what it's telling us is all these that are brought up (like your example above) are just natural random noise that is filtered out over the course of an entire season. Anomalies are anomalies. They don't disprove anything.

pahster
05-16-2008, 12:19 PM
You could, but manipulate is a loaded word.

More appropriately, I think, one could say that we know there are disparate values associated with some of these events, but nobody knows what those values exactly are. So, in the "B" component of this compound statistic, you have TB + some other stuff. What this is saying is that this 'some other stuff' isn't quite worth what a TB event is worth, but it does contribute to the team goals. So, a walk type event that is unintentional is worth, roughly, a quarter of a TB event. A Sacrifice, or a SB is worth, roughly, half of a TB event. It's applying a weight factor, which can be seen as a bit arbitrary, but is something that needs to be done, as common sense tells us we can't count those oranges as apples.

For the most part, I don't get too caught up in Runs Created. It's pretty advanced as far a correlation, but OPS isn't too much worse, while being much more readily available (to calculate) from the program, or scoreboard :).

Hm. I'm not sure I'm comfortable with the more advanced RC formulas if this is the case. I wouldn't think it'd be particularly difficult to accurately discern weighted constants through a bit of statistical modeling, though.

One thing I think we should be aware of is that just because X correlates a great deal with Y does not mean that X is a good measurement. I like to look at OPS because it's a nice quick and dirty calculation of a batter's offensive value, but it's a highly problematic tool when it comes to prediction and explanatory power (of runs scored). OPS has a higher correlation with runs scored not because it's a better measure than OBP or SLG, but because it counts hits twice, thus artificially inflating itself. There are other problems inherent in the measure; should OBP and SLG be equally weighted? Probably not. Just something to keep in mind.

Cyclone792
05-16-2008, 12:20 PM
A problem I have with a stat like "Runs created" is that it gives me the same feeling that passing yards, TDs or even points scored does in football. It is missing context.

The stat obviously has value. It obviously correlates to runs. That's wonderful.

But if someone goes 4-4 in a game with 2 doubles and 2 singles but the team gets shutout, he created zero runs in that game no matter what any formula tells you.

Context is what is missing from these equations. Watching a QB throw for 60 yds and a TD in the last drive of a blowout game puts points on the board and pads stats, but has zero direct correlation to winning a game. Watching Marcus Allen score 15 one yard TDs looks great on paper, but doesn't do much otherwise.

The same goes for looking at walks in a vacuum. Doubles in a vacuum. Even HRs in a vacuum. Hitting a solo HR off of Mike Lincoln in a 10-2 blowout equates to about 1/10th (or less) of the value of hitting a 2 run game winning single off of Mariano Rivera, even if that HR "created" more runs than the single.

The stat has value, but it is difficult to use it as the measurement to end all measurements of a player.

If you were correct, then the formula wouldn't correlate to actual runs ~96-97 percent. But it does correlate to actual runs, which means that what you're saying above isn't accurate.

bucksfan2
05-16-2008, 12:26 PM
If you were correct, then the formula wouldn't correlate to actual runs ~96-97 percent. But it does correlate to actual runs, which means that what you're saying above isn't accurate.

Does the formula correlate to the total runs scored in a season? I am curious as to what the RC formula is acurate to 96-97 percent.

pahster
05-16-2008, 12:30 PM
Does the formula correlate to the total runs scored in a season? I am curious as to what the RC formula is acurate to 96-97 percent.

Yes, team runs scored. Depending on the formula, the r = .95-.97. RC correlates to a very high degree with runs scored.

jojo
05-16-2008, 12:34 PM
I had HS Algebra quite a few years ago but could you say in other words; James has used numbers (linear weights) in order to manipulate a formula in order to get a desired return?

It's not untoward but rather more desirable to increase accuracy of a model.....

BTW, linear weights aren't arbitrary...they're dictated by the data.

jojo
05-16-2008, 12:37 PM
I got a PM last night that I'm going to post. I'll take the name off of it, but it fits with our discussion.

I think it's been brought up a couple of times in passing during this thread that there are better run estimators out there.

RC is just more accessible primarily because of tradition.

edabbs44
05-16-2008, 12:37 PM
If you were correct, then the formula wouldn't correlate to actual runs ~96-97 percent. But it does correlate to actual runs, which means that what you're saying above isn't accurate.

But some runs mean more than others.

SteelSD
05-16-2008, 12:38 PM
We can argue all the problems with it, but it works so well, I don't think it becomes difficult to do so. It's not perfect, but it's pretty darn good.

That's really the point and protestations to the contrary are lacking context. For example, any discussions about the accuracy of things like Runs Created are held with things like Batting Average or RBI already sucking dust. When we're talking about the more advanced metrics, folks who don't understand them don't realize that we're talking about correlation differentials versus actual Run scoring measured in fractions of percentage points while already sitting in the high 90% range. No metric is ever going to be perfect, but when comparing advanced metrics, it's "more perfect" versus "less perfect" rather than the "accurate" versus "inaccurate" we see when comparing RC to Batting Average. RC versus Base Runs? Ferrari versus Jaguar. RC versus Batting Average? Ferrari versus Pinto.

woy's wikipedia citation of the "Problems with Runs Created" is a good example of how these discussions tend to break out. As he understands RC, woy either read (or posted) that wiki entry knowing that the "problems" create only intensely small variances in correlation. And as Cyclone noted, they're easily worked through for all but the most extreme outliers in baseball history.

pahster
05-16-2008, 12:40 PM
But some runs mean more than others.

RC is used as a predictor of run scoring. They are all equal for this particular measure. 1 run = 1 run no matter when or how it scores, no matter the outcome of the game. RC doesn't seek to predict wins and losses, only run scoring.

Cyclone792
05-16-2008, 12:41 PM
Does the formula correlate to the total runs scored in a season? I am curious as to what the RC formula is acurate to 96-97 percent.

Sure it does.

The easiest way to test the correlation is to gather a sample of teams from several seasons (at least five seasons). Then take their team total RC and correlate those figures to the actual runs each team scored. The correlation is 96-97 percent, give or take a few tenths.

If you want, just add each player's RC on the Reds. For example, THT RC version has each individual Reds players RC totals adding up to 182 runs created this season. How many runs have the Reds actually scored this season? 182 runs.

Last season, THT RC version had the Reds create 759 runs. The Reds actually scored 783 runs, and that's a difference of only 3 percent. You can take any team in the history of baseball, apply the RC formula to it, and I'd be willing to guess that over 95 percent of all teams RC will come within 3-4 percent of their actual runs total.

If the stat wasn't incredibly accurate, that wouldn't be possible.

bucksfan2
05-16-2008, 12:42 PM
I hear arguments like this all the time, but they don't make sense to me. Context is more important in the micro. If context were so important, then why is the formula so accurate in predicting runs scored. When it's all said and done, it is incredibly accurate which trumps all the other contextual arguments. And if you break down the formula, it makes sense. It's basically measure how many bases the team gets per time they come to the plate. And the fact that the team RC is just a compilation of all the singles, double, triples, walks, HRs, HBP, etc..., it makes sense that an individual hitter's overall RC number is an accurate representation of their individual contribution to runs scored.

We can argue all the problems with it, but it works so well, I don't think it becomes difficult to do so. It's not perfect, but it's pretty darn good.

This is a reason I struggle with these kinds of stats. "Context is more important in the micro". Isn't what happens in the micro (each at bat) more indicitive of the outcome of a game than what happens in the macro (on average). I would imagine a RC stat would be important for an agent trying to get a bigger contract and a GM trying to decide who is worth the money.

You also lose the importance of an occurance. Is every run created equal? Is every hit, BB, IBB, 2B, etc created equal? On paper a 2 run HR is the same whenever it happens but in the context of a given game it can have two completely different impacts.

pahster
05-16-2008, 12:44 PM
This is a reason I struggle with these kinds of stats. "Context is more important in the micro". Isn't what happens in the micro (each at bat) more indicitive of the outcome of a game than what happens in the macro (on average). I would imagine a RC stat would be important for an agent trying to get a bigger contract and a GM trying to decide who is worth the money.

You also lose the importance of an occurance. Is every run created equal? Is every hit, BB, IBB, 2B, etc created equal? On paper a 2 run HR is the same whenever it happens but in the context of a given game it can have two completely different impacts.

RC doesn't seek to measure the "importance" of runs; it seeks to attribute the number of runs a team scores to individual players. It measures run production, not wins. Whether or not a team wins a game is irrelevant.

Roy Tucker
05-16-2008, 12:54 PM
I think it just takes a while for new stats to become assimilated and accepted by the mainstream fan.

OPS/OBP/SLG used to be considered fairly whacked-out stats used only by propeller heads. But now it seems more and more accepted by the mainstream fan as they get used to what it means. Also used by mainstream media.

RC will get there as it gets past the evangelical stage into accepted. Just takes time.

edabbs44
05-16-2008, 12:55 PM
RC is used as a predictor of run scoring. They are all equal for this particular measure. 1 run = 1 run no matter when or how it scores, no matter the outcome of the game. RC doesn't seek to predict wins and losses, only run scoring.

Then RC should only be used as a measurement of "creating runs" and not as a center of discussion in talks of which player is better, MVP voting or HOF voting.

Chip R
05-16-2008, 12:58 PM
Then RC should only be used as a measurement of "creating runs" and not as a center of discussion in talks of which player is better, MVP voting or HOF voting.


Then all stats should be excluded then, right? Batting average, HRs, RBIs, ERA, the whole lot.

pahster
05-16-2008, 01:02 PM
Then RC should only be used as a measurement of "creating runs" and not as a center of discussion in talks of which player is better, MVP voting or HOF voting.

Do you disagree with the notion that better players create more runs than lesser players?

BuckeyeRedleg
05-16-2008, 01:07 PM
A problem I have with a stat like "Runs created" is that it gives me the same feeling that passing yards, TDs or even points scored does in football. It is missing context.

The stat obviously has value. It obviously correlates to runs. That's wonderful.

But if someone goes 4-4 in a game with 2 doubles and 2 singles but the team gets shutout, he created zero runs in that game no matter what any formula tells you.

Context is what is missing from these equations. Watching a QB throw for 60 yds and a TD in the last drive of a blowout game puts points on the board and pads stats, but has zero direct correlation to winning a game. Watching Marcus Allen score 15 one yard TDs looks great on paper, but doesn't do much otherwise.

But why would you want to know this in the micro? You know he went 4-4 with 2 doubles and 2 singles. Of course runs created wouldn't do anything for you in this situation.

What do you need to know other than he had a great game?

RC, as MWM mentioned, is more about the macro, but it can be divided to measure what a player contributes per game (on average).

Raisor
05-16-2008, 01:09 PM
Let's not forget that there can be negative runs created too.

osuceltic
05-16-2008, 01:11 PM
This is a reason I struggle with these kinds of stats. "Context is more important in the micro". Isn't what happens in the micro (each at bat) more indicitive of the outcome of a game than what happens in the macro (on average). I would imagine a RC stat would be important for an agent trying to get a bigger contract and a GM trying to decide who is worth the money.

You also lose the importance of an occurance. Is every run created equal? Is every hit, BB, IBB, 2B, etc created equal? On paper a 2 run HR is the same whenever it happens but in the context of a given game it can have two completely different impacts.

This gets at my issue with these measures. I absolutely acknowledge their value in measuring data over time, but each game is an individual battle that must be played and managed to be won. One run in a given game often is more valuable than the potential for four runs. That one run may knock a starting pitcher out of the game earlier. It may force the opponent to begin playing the infield in in later situations, creating opportunities for hitters. That one run could be incredibly valuable, even if the percentages say you're more likely to get a big inning if you don't play for one run (by sacrificing, making "productive outs," etc.).

This really isn't a criticism of RC as a stat. It's more of an observation on the imperfections built into judging players almost entirely on these seasonal and career statistical measures. There are plenty of guys whose value to winning a given game goes far beyond a stat like RC (Keith Hernandez, Derek Jeter come to mind) and other players whose value to winning a given game doesn't seem to match their RC or similar stats (Dunn, Bobby Abreu, etc.).

Clearly, RC isn't designed to measure my "value to winning a given game" stat. But it and other similar stats often are thrown around on this board and elsewhere as the best way to measure the value of a player. They're a valuable piece in measuring value, but only a piece.


RC doesn't seek to measure the "importance" of runs; it seeks to attribute the number of runs a team scores to individual players. It measures run production, not wins. Whether or not a team wins a game is irrelevant.
Some might say that's the only relevant stat.

SteelSD
05-16-2008, 01:12 PM
But some runs mean more than others.

Only in hindsight. Is the 6th Run scored in a 6-0 game really less valuable than any of the other Runs? What happens to how we view that 6th Run if the opposing team comes back to score 6 in the ninth Inning to tie the game? Is the 7th Run that ends it really more valuable than the 6th, when we know that the 7th couldn't have been scored without each preceeding Run already on the board?

Raisor
05-16-2008, 01:15 PM
Only in hindsight. Is the 6th Run scored in a 6-0 game really less valuable than any of the other Runs? What happens to how we view that 6th Run if the opposing team comes back to score 6 in the ninth Inning to tie the game? Is the 7th Run that ends it really more valuable than the 6th, when we know that the 7th couldn't have been scored without each preceeding Run already on the board?

That game really was a "soft run" advocate's nightmare, wasn't it?

edabbs44
05-16-2008, 01:15 PM
Do you disagree with the notion that better players create more runs than lesser players?

Absolutely agree with you.

edabbs44
05-16-2008, 01:16 PM
Then all stats should be excluded then, right? Batting average, HRs, RBIs, ERA, the whole lot.

They all should be viewed together.

pahster
05-16-2008, 01:16 PM
Some might say that's the only relevant stat.

Some might. But it's silly to criticize a measurement for failing to measure something that it doesn't attempt to.

BRM
05-16-2008, 01:18 PM
Only in hindsight. Is the 6th Run scored in a 6-0 game really less valuable than any of the other Runs? What happens to how we view that 6th Run if the opposing team comes back to score 6 in the ninth Inning to tie the game? Is the 7th Run that ends it really more valuable than the 6th, when we know that the 7th couldn't have been scored without each preceeding Run already on the board?

Here we go with the donut theory again.

BCubb2003
05-16-2008, 01:29 PM
There's always been an effort to adjust stats for various things, like park-adjusted stats for instance. Or the assumption that good stats against September call-ups don't mean as much. So I guess that's why there's an urge to adjust stats for game conditions.

The two schools of thought seem to be ... if you need a run in this situation, you want your most solid runs creator up there ... vs. ... a guy who knows that the third baseman is not moving well so he exploits that weakness and disrupts the defense until he can manufacture a run in this particular case.

I suppose the argument will continue, with one fan saying, "all we needed was a run and the big dude struck out again" and the fan in the next seat saying, "but over the course of a season he has an RC value of ..."

edabbs44
05-16-2008, 01:30 PM
Only in hindsight. Is the 6th Run scored in a 6-0 game really less valuable than any of the other Runs? What happens to how we view that 6th Run if the opposing team comes back to score 6 in the ninth Inning to tie the game? Is the 7th Run that ends it really more valuable than the 6th, when we know that the 7th couldn't have been scored without each preceeding Run already on the board?

Many discussions are in hindsight. MVP voting, for example, happens after the season. So that way you can fully appreciate what the player meant to the team.

MWM
05-16-2008, 01:38 PM
This is a reason I struggle with these kinds of stats. "Context is more important in the micro". Isn't what happens in the micro (each at bat) more indicitive of the outcome of a game than what happens in the macro (on average). I would imagine a RC stat would be important for an agent trying to get a bigger contract and a GM trying to decide who is worth the money.

You also lose the importance of an occurance. Is every run created equal? Is every hit, BB, IBB, 2B, etc created equal? On paper a 2 run HR is the same whenever it happens but in the context of a given game it can have two completely different impacts.

But the purpose and value of these things isn't in a single game. It's in measuring a player's overall value, or in trying to understand where your team weaknesses are so you can address them. It's a decision making tool. You don't build teams for what might happen in an individual game. If a guy creates a lot of runs, you generally want him on your team. If he doesn't, you generally don't unless they bring something else to the table such as great defense.

None of the stats you've listed as what you like to look at work in the micro either. Help us understand how your micro (each at bat) way of looking at things helps you put together a team. It doesn't. It helps looking at an individual game and understanding how that game was won or lost. But over the long term, games are won or lost by scoring more runs than the other team.

That's where Pythagorus comes in to complete the story. People have argued that certain runs mean more, so that socring more runs doesn't always lead to more wins because you either score 10 or 2. Ultimately, if that were the case, then Pythagorus wouldn't be so accurate and you'd see more variation that could be tied to more variable run scoring. But I'm not aware that we've seen any of this.

The accuracy of runs created added to the accuracy of Pythagorus is pretty hard to dispute. If you fill your lineup with people who "create" runs, you're going to score a lot of runs. And with equal pitching and defense, you're going to win more games. All the little things that can happen in an individual game you can't plan for and make personnel decisions based on; and over the course of the year, the randomness equals out so that you get back to runs created approximating runs scored.

I used to scoff at stats. But I've worked with statistics in my life so much outside of baseball, that when I started to see how closely things correlated and how consistent the data was from year to year, I realized any arguments against it were futile.

BuckeyeRedleg
05-16-2008, 01:41 PM
Here was my experience with the correlation study. Some of this is copy and paste from some of my older posts to make it relevant for this discussion.

So how accurate is RC and OPS compared with Batting Average? Why is it better? Well, here is where correlation comes into play and I had never heard of it until just recently, when some posters on the Redszone guided me through the process. It was an awesome learning experience. You can correlate the different measurements (AVG, OBP, SLG, OPS, H, BB, K, HR, etc) to the actual number of runs scored to determine which stat is the most accurate in predicting the most important objective in offense - scoring runs. The reason I undertook the challenge was because I really wanted to understand why SLG, OBP, OPS were the best ways to measure a hitter or teams ability to score runs.

Ok, so what I did was I went to a cool website called retrosheet.com and copy-and-pasted (into an excel doc) team stats for every season from 1955 on to 2005 (50 years of data). I wanted a large sample size to draw from, but it only took me 30 minutes or so to copy-and-paste each season into the spreadsheet. Once I had all team information from each year (Like 1200 rows of data), I could create separate columns off to the side and plug in the various RC formula's. Once these are figured out and it really doesn't take long, you can then correlate. To correlate all you have to do his highlight the two columns you want and then hit a button on your toolbar or type some weird jargon (can't recall right now) in the field and hit "enter". This then gives you a correlation between those two columns. What we want to correlate is runs per season (per team) to BA, OBP, SLG, and RC. You also can figure the standard deviation of error from one column to another, as long as it is a counting stat. However with the correlation feature, you can even correlate counting stats to Runs scored as well. Well, what I found was that RC correlates the best with the actual runs scored, followed, by OPS, SLG, OBP, and way behind was BA (batting average).

RC (basic) .9650
RC (SB) .9707
RC (tech) .9750
XR (extrapolated runs ) .9774

Since the Bill James original (basic) RC formula, he as well as other baseball stat-wankers have altered it to make it even more closely correlated with actual runs scored. XRC, which is "EXTRAPOLATED RUNS CREATED" uses linear weights to determine the most accurate number that correlates with runs scored. The point here isn't that one RC formula does better than the other, because it's obvious that the more complicated the formula the higher the correlation. I think some are missing the forest through the trees with this fact. The point is that even the simplest form of RC correlates to actual run production at .9650 (at least in my study), which is significantly better than any other rate or counting stat.

Now, see what follows with some other rate stats......

OPS .8448
SLG .8191
OBP .7829
AVG .7176

and some counting stats....

HR .7733
BB .6166
K .3611


In fact to give you an idea of how bad batting average is to determine the most efficient run producers, all you have to do is look at the Reds of 2005. 4th in all of baseball in runs scored (820). 19th in team batting average (.261). This is an extreme micro example, but over the macro this proves true. What about strikeouts? Even K's correlate positively to scoring runs. Hmmm. We hear all the time how bad they are, but guess what, there is a correlation to the top run scoring teams and higher strikeout totals. In 2005, the Reds finished #4 in runs scored. In K's? Numero uno. The Big Red Machine teams of 1975 and 1976? 1975-1st in the major's with 840 runs scored. 2nd in OPS and believe it or not 4th in strikeouts. 1976? 1st in runs scored with 857 and #1 in OPS and 2nd in team K's.

registerthis
05-16-2008, 01:46 PM
They all should be viewed together.

But that assumes that all stats are equally valid and equally meaningful. A stat like BA has a small modicum of usefulness, but really is about as pedestrian--and uselsss--as stats get. I don't understand why you'd want to mix in a stat such as BA with a stat such as RC, which correlates so much more closely with the run production value--and thus the inherent superiority--of a particular player.

RichRed
05-16-2008, 01:48 PM
Let's not forget that there can be negative runs created too.

There's no reason to bring Juan Castro into this.

jojo
05-16-2008, 01:49 PM
But some runs mean more than others.

You should be a big fan of win expectancy-based metrics like WPA.....

BuckeyeRedleg
05-16-2008, 01:56 PM
There's no reason to bring Juan Castro into this.

The mere mention of his name results in a one day ban I've heard.

edabbs44
05-16-2008, 02:03 PM
But that assumes that all stats are equally valid and equally meaningful. A stat like BA has a small modicum of usefulness, but really is about as pedestrian--and uselsss--as stats get. I don't understand why you'd want to mix in a stat such as BA with a stat such as RC, which correlates so much more closely with the run production value--and thus the inherent superiority--of a particular player.

This is when the discussion becomes ridiculous.

GAC
05-16-2008, 02:12 PM
Unfortunately Bruce's "on base" average is only 32 points higher than his batting average.... he'll need more time in AAA.... :cool:

And Patterson's OB% is only 10 points higher then Dunn's I.Q. :p:

registerthis
05-16-2008, 02:15 PM
This is when the discussion becomes ridiculous.

Only if you're dead-set on clinging to stats such as BA and RBI as a means by which to adequately judge the value of a player. You can dismiss as "ridiculous" the notion that all stats are not created equal, but that doesn't make it so.

RedsManRick
05-16-2008, 02:50 PM
We often seem to go back and forth between the way we're using a stat. It’s important to make sure we’re all trying to accomplish the same thing. Are you measuring performance or skill? Are you trying to measure historically or predict future value? Retrospective analysis can consider specific context in assigning value – when considering context, some singles are more valuable than some HR.

For example:

Game 1: Jeff Keppinger hits a single with two outs in the bottom of the 9th, driving in 2 runs to win the game, 5-4.

Game 2: Adam Dunn hits a solo HR in the top for the 6th, extended the Reds lead to 8-2.

In a retrospective sense, Keppinger's single was more "productive" than Dunn's HR. It did more to further the team's chances of winning that game. However, the differential "run creation" value of Keppinger's single versus Dunn's HR was due to context. While we can celebrate the event of Keppinger’s single winning the game, we cannot fairly give Keppinger credit for the facts of the situation. He didn’t put those runners on 2nd and 3rd. He didn’t put the Reds down 1. (If they had been down/up 5, would that make Keppinger a less productive player?) The event was great – but the credit for the event should be shared across all of those players who “created” it. And that those specific runs were the “game winning” ones, they weren’t any more or less important in the final outcome than the first 3.

A similar thought process can be applied to Dunn’s HR.

So yes, Keppinger's single, as an event, was more valuable to winning a given game than was Dunn’s HR. But the value of that level of detail ends there. As you try to look at production in the aggregate, you’d need to apply that context specific valuation. That would be incredibly complicated. It has been done, leverage indices for pitchers for example. But that’s for a very specific reason. One which does not apply to starters or to offense.

The Tony Perez example used earlier is instructive. When it comes to offense, there are certain situations which have much more game-specific value than others. However, over time, offensive players tend to experience this situations at roughly the same rate and their performance tends to be (more or less) evenly distributed across those scenarios. Any clutch/choke effect is at the very margins. As players accumulate PA, the context issue comes out in the wash. The exception here is relief pitchers (and pinch hitters), who are used in such a way that the leverage of their usage follows a bias. Closers/Set-Up guys get placed in high leverage situations, and thus their production tends to have a disproportionate effect relative to, say, a long reliever. Their performance has a consistently stronger effect on actual game outcome. Unlike with offense, it does not even out over time.

Because of this, prospective analysis (essentially skill, as opposed to performance) does not need to (and could not) consider context. Given that offensive performers don’t have a significant ability to align their best performances with the highest leverage scenarios and given the full slate of potential game scenarios, the prospective HR is always more valuable than the prospective single. As you look forward, you cannot assume any given context. While you can accurately say that player X has done well in high leverage situations in the past, you cannot accurately infer that he is likely to keep doing so. You can make such an inference about the actual event though (the walk, single, HR, etc.)

The evidence of this is precisely the extremely high correlation between RC and actual observed runs. So yes, that particular Keppinger single is not going to get nearly the credit it deserves in the RC formula, relative to its real life importance. And vice versa for Dunn's HR. But both at the team level, and individually, those events will balance out over time. Each player's performances will be more or less randomly distributed across the universe of situations as they get more experience. Even over the course of a season, that RC total will quickly become a pretty accurate measure. Of course, it won't be perfect. If this wasn't true, RC wouldn't model against actual runs so darn well. But it does.

If you want to nitpick, find cases where it's not perfect, you are free to do so. No formula will ever model reality perfectly. For me, the key is understanding that despite its imperfects, it's better than just about anything else -- including our personal intuition. That's not to say we shouldn't listen to our gut, that we shouldn't believe our eyes. But we should vet them -- heavily. Humans are not very good at all about properly assigning value to events. Our brains just aren't wired to do it well. We are detail oriented thinkers and have trouble with aggregation. In fact, quite often our brain tricks us. How much did you mourn for the ~3000 dead in 9/11 versus the ~230,000 from the 2004 Tsunami? How much did you fret over the fate of Elizabeth Smart compared to the dozens, if not hundreds, which occur in your local community each year? How many of you realize that car transportation is orders of magnitude more dangerous than flying in a plane? Our eyes deceive us all the time.

But if you want a pretty darn good assessment of the value of a player's past value, and even more so, a good predictor of future value, Runs Created does this. Sure, it's hard to link it back to a memory. You can't think of that game where Dunn "created" a really crucial run. It doesn’t give you that warm fuzzy feeling of feeling like you’re modeling reality by calling them “Runs Batted In”. But the reality of the situation is that advanced stats like Runs Created, among others, do a really darn good job at assessing the value of performance.

Stats like RC certainly aren’t beyond criticism; They aren’t perfect. But they deserve to be attacked on their merits, and doing this requires an understanding of how and why those metrics work – and how they’re intended to be interpreted. At the risk of sounding pompous, us “stat guys” deserve the benefit of the doubt. We watch the game. We appreciate the difference between reality and models. But performance analysis isn’t a relativistic science. There are better and worse measures of player performance and they can be proven as such. Threads like this are very helpful both for those interested in learning more about the stat side and for keeping us stat-wankers grounded. I just hope we can continue to keep things civil.

RedsManRick
05-16-2008, 03:04 PM
Many discussions are in hindsight. MVP voting, for example, happens after the season. So that way you can fully appreciate what the player meant to the team.

I find that question fascinating. Do you look at the actual value given the context in which he operated or the theoretical value, controlling for the randomness of context?

If Ryan Howard gets 30 more RBI than Albert Pujols because he gets 80 more opportunities to work with, but drives in runners at a lower rate, who is "most valuable?"

I personally like to remove context, find the guy who did the best given the actions for which he's responsible. But the voting body at large clearly either wants the context, or doesn't understand/care about it's effect.

Kc61
05-16-2008, 03:08 PM
I've been reading through this thread, and I appreciate that over time, modeling all players' performances and their numbers, there are good statistical metrics to attempt to predict performance and to analyze hitters' contributions. But many of them are based on the concept that things "even out" and that timing of particular hitting performances is basically random.

This may be generally true when looking at major league players as a whole. But it does point up the following question:

Aren't there hitters who don't do well in high pressure situations?

Aren't there hitters who tend to press, and to fail, with men on base or in close games?

In hindsight, a home run hit when leading by 5 runs may or may not turn out to be important. But it would seem to me that some individuals may perform better in those (lower pressured) situations than they do in higher pressure situations. And that this should be a relevant consideration in looking at performance.

Sea Ray
05-16-2008, 03:08 PM
This is when the discussion becomes ridiculous.

Oh it got ridiculous at least 100 posts ago. If one still hasn't made his point after 400 posts, you got problems.

Carry on... :cool:

jojo
05-16-2008, 03:11 PM
Oh it got ridiculous at least 100 posts ago. If one still hasn't made his point after 400 posts, you got problems.

Carry on... :cool:

That doesn't bode well for an active participant in the discussion who posted, say, the 444th post of the thread..... :cool:

RedsManRick
05-16-2008, 03:16 PM
They all should be viewed together.

I don't understand this sentiment. What value do you derive from the RBI stat? What is it telling you about that player than you cannot derive more accurately from elsewhere? The same could be said for ERA, BA, etc.

I think as a community, we've let the statistic tail wag the dog. The stat-head movement is really at heart, an attempt to correct this problem. The purpose of stats is to inform evaluation and decision making. But we've become so attached to certain stats, that they've gained value in and of themselves. The "save" stat is a perfect example of this. It's ridiculously convoluted and borderline meaningless. But it's become a management tool. That's just intellectual laziness.

There are better ways to measure the things we want to measure. I think a very interesting discussion would be one centered around the measurement question. Starting from scratch, setting aside all existing "stats" for a second, what are we trying to get at? What are the things we're measure.

Once we've articulated that, we have a common framework from which we can evaluate what stat/metric best answers those questions. We need to get outside of our historical bounds for this conversation to really move forward.

dabvu2498
05-16-2008, 03:16 PM
I've really enjoyed (most of) this discussion. Good work to all of those who have taken the time and effort to explain this stuff.


I find that question fascinating. Do you look at the actual value given the context in which he operated or the theoretical value, controlling for the randomness of context?

If Ryan Howard gets 30 more RBI than Albert Pujols because he gets 80 more opportunities to work with, but drives in runners at a lower rate, who is "most valuable?"

I personally like to remove context, find the guy who did the best given the actions for which he's responsible. But the voting body at large clearly either wants the context, or doesn't understand/care about it's effect.

RMR -- If I'm reading you right, you'd say that someone who has 150 RC for a 70-92 team has the same "valuableness" as someone who has 150 RC for a 92-70 team. Correct?

If so, that's a hard thing for some of us to grasp.

BCubb2003
05-16-2008, 03:21 PM
The MVP has always been a subject of those kinds of debates. It's the "we could have finished last without you" syndrome.

Is the best player on the Yankees more valuable to his team than the best of a few good players on another team or the only good player on a last-place team?

RedsManRick
05-16-2008, 03:27 PM
I've really enjoyed (most of) this discussion. Good work to all of those who have taken the time and effort to explain this stuff.



RMR -- If I'm reading you right, you'd say that someone who has 150 RC for a 70-92 team has the same "valuableness" as someone who has 150 RC for a 92-70 team. Correct?

If so, that's a hard thing for some of us to grasp.

I partially answered this in the post I just made and edited, above yours.

If we're asking the question, which player is better, the answer is that both guys are equally good. That 70-92 team wasn't 70-92 and because the performance it got from its 150 RC guy was worse than the 150 RC performance from the 92-70 team. It's all of the other things going on around those 150 RC that meant the difference. When evaluating those two players, we shouldn't be giving one of them extra credit for having better teammates.

I don't think "value" has been defined well at all. Value is a pretty complex concept that has been bent in a bunch of different ways. A ming vase could be worth millions in New York, and pennies in Africa. Value is a context dependent metric, by definition. It's an ascribed characteristic based on the particulars of the environment.

When we evaluate players, value is rarely what we're actually getting at. We're interested in some absolute measure of how well the player does the things which tend to lead to runs scoring. The HR hit when there are 3 guys on base (player A) is more valuable than the HR hit when the bases are empty (player B). But we cannot extrapolate that point to the player himself. He didn't create the context. So in evaluating that player, we evaluate only the portion of that event for which that player was responsible.

Using RBI to measure the production of the player suggests that player A did something drastically different than player B. That's simply not true. And clutch study after clutch study is showing us that players simply do not have much, if any, control over the "when" of their performances.

Sure, as an event, player A's HR was more valuable. But to credit that differential value due to context to player A, suggesting that his HR makes him a better player than player B, is a mistake. I personally don't want to evaluate specific players based on the environment in which they operate when they don't have any impact on that environment. It simply doesn't tell me anything I find important. Aaron Harang is not a less valuable pitcher (to me) because his teammates haven't given him good run support.

Cyclone792
05-16-2008, 03:30 PM
RMR -- If I'm reading you right, you'd say that someone who has 150 RC for a 70-92 team has the same "valuableness" as someone who has 150 RC for a 92-70 team. Correct?

If so, that's a hard thing for some of us to grasp.

An individual player really cannot control how productive his teammates are. Sure, players can influence the teammate dependent statistics of one another - the stats that really are misleading because of that influence - but they really can't influence real, isolated production measures. Accurately determining a player's value means you must correctly isolate his value from the team and eliminate influences that are teammate dependent.

I'll use an example that's nearly a century old and in which nobody should have any biases.

In 1908, Ty Cobb hit .324/.367/.468 with a 169 OPS+ and helped lead the Tigers to the AL Pennant with a 90-63 record. In 1912, Cobb hit .409/.456/.584 with a 200 OPS+ and the Tigers went 69-84.

Was the 1908 season a better year for Cobb than the 1912 season? No, not at all. It was a better year for his team, but it wasn't a better year for Cobb himself. Cobb's "valuableness" was higher in 1912 than in 1908.

The "valuableness" of the player is largely determined by the player's own isolated production. It's not determined by his teammates. Over a 162-game season, Ernie Banks' teams had an average season of 75-87. Banks was a great player, a Hall of Famer, but it should not be held against him that he often played for poor teams.

Raisor
05-16-2008, 03:37 PM
Oh it got ridiculous at least 100 posts ago. If one still hasn't made his point after 400 posts, you got problems.

Carry on... :cool:

I'm going to assume you still think BAw/RISP >OPS w/RISP

dabvu2498
05-16-2008, 03:38 PM
Thanks for the responses guys.

Let me throw another one at you -- when evaluating player decisions (lineups, contracts, etc.) should we put any weight on team context or, as RMR said above should we strictly "remove context, find the guy who did the best given the actions for which he's responsible?"

Also, jojo mentioned WPA above. I know just enough about WPA to say something stupid about it, so instead, I'll defer to those more well-versed than me. What are your thoughts on it as a useful tool?

edabbs44
05-16-2008, 03:49 PM
I don't understand this sentiment. What value do you derive from the RBI stat? What is it telling you about that player than you cannot derive more accurately from elsewhere? The same could be said for ERA, BA, etc.

I'm not saying that RBI is the be all, end all stat. Far from it. But why would someone discount it? Because one may have had more opportunities than someone else? Cool. Add that to the equation.

Player A had 115 RBI in 250 ABs with RISP.
Player B had 100 RBI in 150 ABs with RISP.

Pretty relevant info. But I think it is easier to add that to the equation rather than discounting a statistic just because there may be other information skewing the data.

ERA is useless? Tell me, when it is all said and done would you rather have a pitcher on your team have an ERA of 3 when other stats say it should have been 5 or a pitcher post an ERA of 5 when other stats say it should be 3?

Give me the guy who actually finds success on the mound over a guy who should have had success.

Now, solely using ERA to project what the pitcher will do next year may be a bit distorted and is a different story. But what's wrong with listening to ERA and then disputing it with other data if that is the case? Don't throw out ERA, supplement it if necessary.

There's a reason why pitchers like Webb and Johan are typically near the top of the ERA leaderboard and why hitters like ARod and Pujols are typically driving in the most runs.

The stats may not be flawless (like all other stats), but they aren't useless either.

bucksfan2
05-16-2008, 03:51 PM
For example:

Game 1: Jeff Keppinger hits a single with two outs in the bottom of the 9th, driving in 2 runs to win the game, 5-4.

Game 2: Adam Dunn hits a solo HR in the top for the 6th, extended the Reds lead to 8-2.

In a retrospective sense, Keppinger's single was more "productive" than Dunn's HR. It did more to further the team's chances of winning that game. However, the differential "run creation" value of Keppinger's single versus Dunn's HR was due to context. While we can celebrate the event of Keppinger’s single winning the game, we cannot fairly give Keppinger credit for the facts of the situation. He didn’t put those runners on 2nd and 3rd. He didn’t put the Reds down 1. (If they had been down/up 5, would that make Keppinger a less productive player?) The event was great – but the credit for the event should be shared across all of those players who “created” it. And that those specific runs were the “game winning” ones, they weren’t any more or less important in the final outcome than the first 3.


RMR I see your points. But to not give Keppinger in the first example as much credit for driving in the runs is something I don't understand. He wasn't the reason that two runners were one base but he was the reason they scored. Any on us who have watched the Reds play as of late realized the frustrating aspect of not being able to knock runners in. With the exception of a HR and very few other examples any time a run scores it is usually a creation of two or more people. But I don't feel the need to discount the 2 RBI single because Keppinger wasn't the one responsible for getting on base.

I can see the usefulness of a RC stat. If you are looking to build a team you may use this to compare and contrast certain players against their peers. I would wonder if a GM would use such a start to decide which FA to sign or whether he would go back and look at game by game logs and situation hitting in order to assess value.

BCubb2003
05-16-2008, 03:51 PM
The crux of the debate might be the heartfelt belief by some people that some players create their own context or at least are more flexible in a variety of contexts ... the proverbial player who finds a way to beat you.

Actually, a slugger like Dunn is able to create his own context (a runner is always in scoring position when he's up), although it happens less often than the walk or single. But the walk or single is team-dependent -- all that on base goodness goes for naught if weak hitters hit behind you.

Raisor
05-16-2008, 03:55 PM
all that on base goodness goes for naught if weak hitters hit behind you.

Except it doesn't.

Let's not forget, an offense is a chain. Every time you make a non-out, you're actually adding a link to the chain at the end.

Even if something good doesn't happen immediatly following a walk (let's say Dunn gets stranded without getting home), you're adding an extra plate appearance later, where something else good can happen.

BCubb2003
05-16-2008, 04:05 PM
Except it doesn't.

Let's not forget, an offense is a chain. Every time you make a non-out, you're actually adding a link to the chain at the end.

Even if something good doesn't happen immediatly following a walk (let's say Dunn gets stranded without getting home), you're adding an extra plate appearance later, where something else good can happen.

Which might be another two-out walk followed by a strikeout.

I'm not resistant to the concept. In fact, I favor it, especially if the alternative for someone like Dunn is to swing at more bad pitches and strike out even more than he does. I just feel for the fan who sees it happen so often and knows what's going to happen next. As you know, it's hard to tell that guy that in some other game down the line, maybe in a 6-0 game, maybe not, Dunn will add to his runs created total.

jojo
05-16-2008, 04:12 PM
ERA is useless? Tell me, when it is all said and done would you rather have a pitcher on your team have an ERA of 3 when other stats say it should have been 5 or a pitcher post an ERA of 5 when other stats say it should be 3?

Give me the guy who actually finds success on the mound over a guy who should have had success.

Lets assume it's a contract year. I'd rather pay the latter pitcher a few million a year for 3 years than the former $127M over 7.....

edabbs44
05-16-2008, 04:24 PM
Lets assume it's a contract year. I'd rather pay the latter pitcher a few million a year for 3 years than the former $127M over 7.....

Which has zero to do with the conversation at hand.

But is very true.

jojo
05-16-2008, 04:27 PM
Which has zero to do with the conversation at hand.

But is very true.

It has everything to do with the conversation at hand.

*BaseClogger*
05-16-2008, 04:43 PM
Please allow me to enter the conversation and pose a question: which correlates better with runs scored, OBP or SLG? According to BuckeyeRedLeg's work it is SLG, but I remember reading in Baseball Between the Numbers that OBP correlates better and that one point of OBP is worth three points of SLG. :help:

*BaseClogger*
05-16-2008, 04:46 PM
ERA is useless? Tell me, when it is all said and done would you rather have a pitcher on your team have an ERA of 3 when other stats say it should have been 5 or a pitcher post an ERA of 5 when other stats say it should be 3?

Give me the guy who actually finds success on the mound over a guy who should have had success.

The first pitcher had less talent but more success. The second pitcher had more talent, but less success. Players with more skill are more likely to have success in the future, however, due to the randomness of the game less talented players sometimes have more success, but the Law of Large Numbers will even things out over time...

BCubb2003
05-16-2008, 04:47 PM
It has everything to do with the conversation at hand.

Which pitcher would you rather have on the team ... for the next game, probably the one with the better supporting stats ... for this season, maybe the one who's having a lucky year ... for the next three years, the one with the better supporting stats.

RedsManRick
05-16-2008, 05:19 PM
I'm not saying that RBI is the be all, end all stat. Far from it. But why would someone discount it? Because one may have had more opportunities than someone else? Cool. Add that to the equation.

Player A had 115 RBI in 250 ABs with RISP.
Player B had 100 RBI in 150 ABs with RISP.

Pretty relevant info. But I think it is easier to add that to the equation rather than discounting a statistic just because there may be other information skewing the data.

Relevant to what? It might be "easier" to add, given the universe of metrics with which you are currently comfortable. But that doesn't mean that there isn't a better way to get at the point you're trying to make -- which is that Player B is more productive.



ERA is useless? Tell me, when it is all said and done would you rather have a pitcher on your team have an ERA of 3 when other stats say it should have been 5 or a pitcher post an ERA of 5 when other stats say it should be 3?

Give me the guy who actually finds success on the mound over a guy who should have had success.

Ok, on balance, I'll take the guys with the lower "should be" ERAs and you take the group with the lower current ERAs. I'll most likely do better. This study has been done. FIP is a better predictor of a player's future ERA than is his current ERA. Crazy huh? But it's true.


Now, solely using ERA to project what the pitcher will do next year may be a bit distorted and is a different story. But what's wrong with listening to ERA and then disputing it with other data if that is the case? Don't throw out ERA, supplement it if necessary.

But why use ERA in the first place? What start someplace I know to be only roughly accurate and adjust if I can start someplace is that is generally more accurate?

The only argument here which makes sense is convenience, but looking up things like FIP simply isn't hard. No, the problem is that we are educated about stats like ERA early on. We get comfortable with them. We know what they mean. For you, and many, many others, the cost is one of comfort. If the goal is to measure the effectiveness of a pitcher, the likelihood that he will do the pitching things which lead to the fewest runs, there are easily accessible measures of that better than is ERA.


There's a reason why pitchers like Webb and Johan are typically near the top of the ERA leaderboard and why hitters like ARod and Pujols are typically driving in the most runs.

The stats may not be flawless (like all other stats), but they aren't useless either.

You're absolutely right. ERA is a pretty decent measure of pitcher effectiveness. And RBI is fairly reflective of player performance too. (especially using a recursive argument where we select the best players using those stats and then use those stats to prove their excellence). The point is not that, in a vacuum, ERA and RBI are useless as measures of performance and skill. It's that there are better ways to measure the same things that are not really any more difficult to obtain or use.

Let's say I want to know how much a box weighs so I can mail it with proper postage. Well, I could pick it up, feel it, and probably get in the ball park with an educated guess. Wanting to be more precise, I could take it in to the bathroom and throw it on my scale and get it right down the pound (though my scale is sort of wonky and seems to be off by a pound or two depending on the humidity). That's probably good enough for most circumstances. I could slap some postage on it, ship it off, and be done with it.

But what if I have to be absolutely sure about using the right postage because I'll lose my job if that package gets returned (or if somebody finds out I wasted postage). I'm going to want an even more precise measure. So I go down to the post office; They use a fancy scale which goes down to the ounce and has much less random variance. That's a no-brainer.

You might say that it's a hassle to walk down to the post office. I'd agree. But now let's pretend that you have one of those in your bathroom too. Why would you use the cheaper, less accurate scale and try to adjust it for its problems? Because you're comfortable with it? Because the press uses it? Because you're used to it?

What really matters is getting to the right weight of the package, right? And if you have to take 15 minutes to read the instructions first, so be it. The investment is small. It just takes a willingness to learn.

Let me try another analogy. It's long. Stop reading now if you're one of those people who hate it when I ramble. I should probably just delete it, but I wasted the last 20 minutes writing it, so here it is.

I could build a sky scraper with 20 foot thick walls made of brick. Bricks are useful for small structures, but simply not strong enough to bear their own weight at great heights. And their rigidity limits their efficacy in many situations. Now, I could buttress every wall, reinforce every ceiling. I could spread the load across a massive base. Given that I'm comfortable with masonry and have a huge supply of bricks, this makes sense to me. I can do better by adding more bricks. If I want to make a large structure in honor of something, I could build a massive pyramid. At one point in history, that was quite the achievement.

But I'm not tied to masonry. I'm just interested in building a really tall building that serves a specific purpose. Sure, brick does a pretty decent job in general. I can build a nice building with it. I could probably build a shorter building that would generally get at my intent. But then I learn of this thing called steel. It's way stronger and lighter than brick. It allows me to design my building to fit my purpose more precisely. I can build a much taller building, make it better fit what I need it to do, do it more quickly, and with less labor.

But there's a catch. I personally don't know how to work with steel and there's a big pile of bricks right over there. Furthermore, the market produces a lot more brick than steel. I'd have to look around a bit to find steel and I'd have to figure out how exactly to use it effectively. Sure, I hear that other people have built great things with it, but I like bricks. I understand bricks. Bricks work! So I have a choice. I can either learn about steel working and where to obtain it, or I can just keep adding more and more bricks.

I won't begrudge you if you want to keep adding bricks. Bricks aren't useless. If you have enough of them and combine them in the right way, you can build a pretty darn impressive structure. Plus, they make a pretty facade and have a rich historical tradition. There are even building design contests where the quality of the masonry is the prime factor.

But if your goal is to build a tall, functional building with reasonable effort, bricks simply aren't the way to go. At one point in time, they were the best we had. They're still way better than sticks and mud and will do in a pinch or for small buildings. So maybe you build a steel building and cover it bricks to satisfy the masses. But building the thing with brick from the foundation up is a pretty poor solution.

But when you want to tackle something big, like a skyscraper (or the entire offensive value of a MLB player), you're starting from scratch, you've been told where to find steel, and you've been directed to study after study showing how to work it, there's not much of a case for simply getting a bigger pile of bricks.

RedsManRick
05-16-2008, 05:43 PM
RMR I see your points. But to not give Keppinger in the first example as much credit for driving in the runs is something I don't understand. He wasn't the reason that two runners were one base but he was the reason they scored. Any on us who have watched the Reds play as of late realized the frustrating aspect of not being able to knock runners in. With the exception of a HR and very few other examples any time a run scores it is usually a creation of two or more people. But I don't feel the need to discount the 2 RBI single because Keppinger wasn't the one responsible for getting on base.

I can see the usefulness of a RC stat. If you are looking to build a team you may use this to compare and contrast certain players against their peers. I would wonder if a GM would use such a start to decide which FA to sign or whether he would go back and look at game by game logs and situation hitting in order to assess value.

Let's try this other angle. Why do you record "2 RBI" as a stat all; For what purpose? What does it tell you about Jeff Keppinger?

Yes, he gets credit, generically speaking, for being the guy who had the hit that drove them in. If I were writing a story about the game, I'd certainly highlight that event. It is a historical fact that Jeff Keppinger hit a ball that resulted in 2 runners scoring. I don't deny that fact and it deserves to be recorded.

But if at the end of the season I ask the question "What guy on the Reds is most likely to do the things that lead to our team scoring runs in the future?", RBI can be misleading. Because of the context, something you cannot predict for the future, Keppinger looks like a more productive player than Adam Dunn. Jeff Keppinger got 2 RBI worth of credit for hitting a single. Adam Dunn got 1 RBI worth of credit for hitting a HR. If we tried to project that forward, we might make the mistake that hitting singles leads to more runs than hitting homers. If you interpreted RBI to be a proxy for overall "run production" you could easily be mislead because of a quirk in context distribution.

It's not taking credit away from anybody. It's about properly defining what you mean by "credit" in the first place and what you do with that credit down the road. It's not that RBI is bad. It's that it has limited function. It's simply not a great measure of "run production". It's not horrible, but we have better -- and should use them.

RedsManRick
05-16-2008, 06:01 PM
Please allow me to enter the conversation and pose a question: which correlates better with runs scored, OBP or SLG? According to BuckeyeRedLeg's work it is SLG, but I remember reading in Baseball Between the Numbers that OBP correlates better and that one point of OBP is worth three points of SLG. :help:

Covariance. The players who get on base a lot tend to be the guys who SLG a lot. That is to say that they overlap. When you just look at one when running a correlation between OBP and Runs, the effect of the other one is still there, but you aren't giving it any credit.

But if those two things correlate with each other (teams who get OBP tend to SLG more - or less - than teams who don't), and OBP and SLG do correlate, then you are still capturing a lot of the other one's effect even though you aren't measuring it directly. So if you want to look at them at the same time, you have to get rid of the part that overlaps or you'd count it twice.

Let's say:
- A vs C have a correlation of .7
- B vs C have a correlation of .5

If you look at A & B together, your correlation isn't going to be 1.2 (that's impossible!). Instead, you are going to have to divide up that overlap.

Let's say that you re-run your analysis look at A and B at the same time to try and predict C.
- A and B together vs C has a correlation of .8

However, when you run a multivariate regression, you also get regression coefficients. Essentially, these tell you the component pieces of that overall model correlation. The process of running the regression tells you were the overlap is and gets rid of it so that the pieces fit together squarely.

So your new overall correlation of .8 is explained by taking 3 parts of A to every 1 part of B. Tthat's not .6 correlation and .2 correlation, mind you. That is to say that a change in OBP has 3 times the effect on run scoring as a similarly scaled change in SLG.

I don't know if that's actually the right ratio for runs and slugging (I thought it was more like 1.5:1), but that's covariance in a nutshell.

edabbs44
05-16-2008, 06:07 PM
Let's say I want to know how much a box weighs so I can mail it with proper postage. Well, I could pick it up, feel it, and probably get in the ball park with an educated guess. Wanting to be more precise, I could take it in to the bathroom and throw it on my scale and get it right down the pound (though my scale is sort of wonky and seems to be off by a pound or two depending on the humidity). That's probably good enough for most circumstances. I could slap some postage on it, ship it off, and be done with it.

But what if I have to be absolutely sure about using the right postage because I'll lose my job if that package gets returned (or if somebody finds out I wasted postage). I'm going to want an even more precise measure. So I go down to the post office; They use a fancy scale which goes down to the ounce and has much less random variance. That's a no-brainer.

You might say that it's a hassle to walk down to the post office. I'd agree. But now let's pretend that you have one of those in your bathroom too. Why would you use the cheaper, less accurate scale and try to adjust it for its problems? Because you're comfortable with it? Because the press uses it? Because you're used to it?

What really matters is getting to the right weight of the package, right? And if you have to take 15 minutes to read the instructions first, so be it. The investment is small. It just takes a willingness to learn.

So you feel that RC is as accurate as weighing a package on a fancy scale?

Cyclone792
05-16-2008, 06:13 PM
Thanks for the responses guys.

Let me throw another one at you -- when evaluating player decisions (lineups, contracts, etc.) should we put any weight on team context or, as RMR said above should we strictly "remove context, find the guy who did the best given the actions for which he's responsible?"

Also, jojo mentioned WPA above. I know just enough about WPA to say something stupid about it, so instead, I'll defer to those more well-versed than me. What are your thoughts on it as a useful tool?

Contracts: Definitely remove all team context to find each player's true value. Teams like the Reds get burned when they fail to do so because they think a guy like Eric Milton is a good pitcher since he had a decent number in the wins column.

Lineups: Remove all team context to find each player's true value in your lineup and determine where each player should hit in an average lineup, then use team context to make slight adjustments given your team's personnel (Ex: the Reds don't have an ideal leadoff hitter but somebody has to leadoff, and unfortunately that somebody shouldn't be Corey Patterson).

WPAs uses are for determining past situational success in specific individual games. It assigns different weights to events based on how much win probability was added after each event occurs. That means a single in the 7th inning of a tie game has more value than a single in the 7th inning of a four run game.

The key here, and this is what people always, always, always forget is that WPA has no predictive value. A player can produce an identical amount of runs in back-to-back seasons, but his WPA can be quite different.

Thinking in terms of basics, there are largely two groups of stats: stats that have value for past events, and stats that have value for future events. Some stats (RC, for example) have value for both. WPA is strictly a stat that has value for past events; it's not designed for predictive value. DIPS ERA is an example of a stat that largely has value for future events rather than past events.

If you're confused about which stat to use then sometimes it's best to first ask: what do I want to know ... actual value in past production, probable value for future production, or both?

RedsManRick
05-16-2008, 06:14 PM
So you feel that RC is as accurate as weighing a package on a fancy scale?

I'm talking conceputally here, edabbs, clearly. RC isn't perfect; It's not exact. When it comes to player production, there's nothing comprable to the precision of finely calibrated scale for weight.

The point is that you seem unwilling to consider the idea that anything could be better than your familiar statistics. If I want to know how productive an offensive player is over a given period of time -- or how productive he's likely to be in the future, I'd use RC in a heartbeat over RBI.

Cyclone792
05-16-2008, 06:14 PM
So you feel that RC is as accurate as weighing a package on a fancy scale?

I'm confused. What precisely have you added to this topic in a meaningful and insightful manner?

edabbs44
05-16-2008, 08:02 PM
I'm talking conceputally here, edabbs, clearly. RC isn't perfect; It's not exact. When it comes to player production, there's nothing comprable to the precision of finely calibrated scale for weight.

The point is that you seem unwilling to consider the idea that anything could be better than your familiar statistics. If I want to know how productive an offensive player is over a given period of time -- or how productive he's likely to be in the future, I'd use RC in a heartbeat over RBI.

My familiar statistics? That's a good one.

I never said that "my familiar statistics" are better than any other statistic. What I have said, however, is that I think that some stats are used in a manner that isn't very fair to others. I remember when, on this very board, Morneau won the MVP and some were using Win Shares as "evidence" that Jeter should have won the MVP. That, to me, is extremely short sighted and a closeminded approach to genuinely trying to evaluate the "valuableness" of a player.

Plugging the numbers in an equation is nice sometimes, but it isn't the be all end all.

edabbs44
05-16-2008, 08:03 PM
I'm confused. What precisely have you added to this topic in a meaningful and insightful manner?

Nice.

jojo
05-16-2008, 08:26 PM
So you feel that RC is as accurate as weighing a package on a fancy scale?

I think this is a non pejorative way of classifying different metrics. Things like ERA and BA/RBIs are very blunt ways of describing performance. Thus reliance upon them as indices of a player's true performance level often leads to mistakes especially as one moves away from the margins and toward the group of players clumped closer to the mean. Metrics which tend to isolate a player's performance from forces that are out of his control or which correlate more strongly to run scoring tend to be more precise.

edabbs44
05-16-2008, 08:40 PM
Metrics which tend to isolate a player's performance from forces that are out of his control or which correlate more strongly to run scoring tend to be more precise.

I think that all of these stats have value. But I'm just not a fan of defining one's performance on a single stat such as Win Shares or Runs Created while just discounting other stats as having zero value like BA and ERA.

jojo
05-16-2008, 08:48 PM
I think that all of these stats have value. But I'm just not a fan of defining one's performance on a single stat such as Win Shares or Runs Created while just discounting other stats as having zero value like BA and ERA.

The thing is though, while a rope that is exactly ten foot long has value as a measuring instrument, carpenters opt for tape measures.....

edabbs44
05-16-2008, 09:01 PM
The thing is though, while a rope that is exactly ten foot long has value as a measuring instrument, carpenters opt for tape measures.....

If you are looking for an estimation for how many runs a player has "created", you should look at the equation for "runs created."

If you are looking for the value that a player brings to a ballclub, you should look at a number of statistics.

jojo
05-16-2008, 09:09 PM
If you are looking for the value that a player brings to a ballclub, you should look at a number of statistics.

You're preaching to the choir. Still this is often best accomplished without relying upon raw counting stats.

edabbs44
05-16-2008, 09:28 PM
You're preaching to the choir. Still this is often best accomplished without relying upon raw counting stats.

Not relying on them...using them as part of the discussion.

I think we're running around in circles.

jojo
05-16-2008, 09:49 PM
Not relying on them...using them as part of the discussion.

I think we're running around in circles.

I don't think we're running around in circles. I'm suggesting metrics like ERA or RBI can actually cloud the discussion rather than clarify it. You're arguing otherwise. It's not semantics-it's a fundamental difference and I'd argue it's one which underpins many of the disagreements the occur with player valuation.

Sure this thread has discussed RC in great length but I don't think any one is arguing that RC is the only metric needed (frankly I only use it because it's hard enough having discussions like this and most people are familiar with RC- I hate winshares BTW). The point is that when given the choice, gravitate toward better metrics.

IMHO, it's a tough sell to argue that eschewing ERA and RBIs is a sign of myopia...

SteelSD
05-17-2008, 12:29 AM
I think that all of these stats have value. But I'm just not a fan of defining one's performance on a single stat such as Win Shares or Runs Created while just discounting other stats as having zero value like BA and ERA.

Batting Average has it's uses. We couldn't calculate Isolated Discipline (IsoD) and Isolated Power (IsoP) without it and both those metrics are predictive in nature, particularly IsoD. When an analysis incorporates BA, IsoD, and IsoP we can get a good idea about whether or not a player has performed or is performing at a potentially-sustainable level. We also couldn't type BABIP without "BA", and that's an important variable in determining whether or not a hitter or pitcher might be affected by poor luck or defense when coupled with DER and/or BIP type rates.

More information, rather than less, is always good for an analysis as long as we understand the accuracy and relevance of every piece being used. The latter item is the real issue with using Batting Average as player valuation metric. It's useful as a component of an analysis, but when compared to other more complete statistical measurements, its relevance as a stand-alone metric fades to black. Batting Average may suggest that someone may be good, but other more complete statistics are just as accessable and far more relevant from a performance evaluation standpoint. Things like Runs Created or even Linear Weights (which jojo is hinting at) are simply far far better tools to use when actually evaluating player performance with things like OBP, SLG, and OPS sitting a tier below the most complete metrics.

ERA and RBI? Now those two are truly pointless. Way way WAY too much noise involved with those.

*BaseClogger*
05-17-2008, 12:33 AM
Covariance. The players who get on base a lot tend to be the guys who SLG a lot. That is to say that they overlap. When you just look at one when running a correlation between OBP and Runs, the effect of the other one is still there, but you aren't giving it any credit.

But if those two things correlate with each other (teams who get OBP tend to SLG more - or less - than teams who don't), and OBP and SLG do correlate, then you are still capturing a lot of the other one's effect even though you aren't measuring it directly. So if you want to look at them at the same time, you have to get rid of the part that overlaps or you'd count it twice.

Let's say:
- A vs C have a correlation of .7
- B vs C have a correlation of .5

If you look at A & B together, your correlation isn't going to be 1.2 (that's impossible!). Instead, you are going to have to divide up that overlap.

Let's say that you re-run your analysis look at A and B at the same time to try and predict C.
- A and B together vs C has a correlation of .8

However, when you run a multivariate regression, you also get regression coefficients. Essentially, these tell you the component pieces of that overall model correlation. The process of running the regression tells you were the overlap is and gets rid of it so that the pieces fit together squarely.

So your new overall correlation of .8 is explained by taking 3 parts of A to every 1 part of B. Tthat's not .6 correlation and .2 correlation, mind you. That is to say that a change in OBP has 3 times the effect on run scoring as a similarly scaled change in SLG.

I don't know if that's actually the right ratio for runs and slugging (I thought it was more like 1.5:1), but that's covariance in a nutshell.

I follow everything except the multivariat regression. How is that performed?

BTW-thanks

Yachtzee
05-17-2008, 12:52 AM
If you are looking for an estimation for how many runs a player has "created", you should look at the equation for "runs created."

If you are looking for the value that a player brings to a ballclub, you should look at a number of statistics.

While that is true, I think looking at RC is a pretty good indicator. Statistics like RC and even Bill James' win shares are attempts to find out a player's true contribution to the things necessary for winning ballgames. They come about from the endeavors of people seeking to find the true value of a baseball player.

Yes, when looking for the value of a player, you should look at a number of statistics. However, teams have entered into quite a few bad contracts with players based on stats like BA, RBI and Win-Loss records. I don't know of any bad contracts that teams entered into because they said "Wow, that guy's Runs Created is off the charts!" As Steel said, BA has its uses. It's just not the primary indicator of offensive production that it has been held out to be.

pahster
05-17-2008, 01:10 AM
I follow everything except the multivariant regression. How is that performed?

BTW-thanks

It's a bit complicated: http://en.wikipedia.org/wiki/Regression_analysis

Basically you build a statistical model that tries to explain the variance in a single numerical dependent variable (so runs, for example) based upon the variance in a number of independent variables (take your pick: singles, doubles, HRs, RC, AVG, OBP, SLG, OPS, whatever). Of course there has to be some logic to the variables you choose; you can't just toss everything in. You also must ensure that there isn't too much covariation among the independent variables or you're model will end up screwy. The results of a regression will tell you what effect (positive or negative) each independent variable exerts upon the dependent variable, the size of the effect, and whether the effect is significant or not.

*BaseClogger*
05-17-2008, 01:19 AM
It's a bit complicated: http://en.wikipedia.org/wiki/Regression_analysis

Basically you build a statistical model that tries to explain the variance in a single numerical dependent variable (so runs, for example) based upon the variance in a number of independent variables (take your pick: singles, doubles, HRs, RC, AVG, OBP, SLG, OPS, whatever). Of course there has to be some logic to the variables you choose; you can't just toss everything in. You also must ensure that there isn't too much covariation among the independent variables or you're model will end up screwy. The results of a regression will tell you what effect (positive or negative) each independent variable exerts upon the dependent variable, the size of the effect, and whether the effect is significant or not.

And some sort of stats software is necessary? I checked and I don't think this is possible on my TI-83... :)

RedsManRick
05-17-2008, 01:41 AM
I follow everything except the multivariat regression. How is that performed?

BTW-thanks

You can do it in excel as well. Under the Tools section, there should be a Data Analysis option which includes all sorts of analytical functions. You just want "regression".

Your Y-axis is your dependent variable (runs) and X-axis is your independent variables -- all of them.

No need to select any of the options. The results will come out in a new worksheet. You'll have all the results you need, including the coefficients.

The coefficients give you the basic formula for the line of best fit.

Y = (Coefficient1)X1 + (Coefficient2)X2... etc.

pahster
05-17-2008, 02:01 AM
And some sort of stats software is necessary? I checked and I don't think this is possible on my TI-83... :)

A lot of people use Excel. I use Stata, SAS, SPSS, and R.

SteelSD
05-17-2008, 02:27 AM
BTW, while this thread has continued to 33 pages, does anyone want to guess who currently leads the Reds in RBI?

MWM
05-17-2008, 02:37 AM
BTW, while this thread has continued to 33 pages, does anyone want to guess who currently leads the Reds in RBI?

And who got the winning RBI tonight with a .... breace yourself.... a walk? Unreal how that happens.

edabbs44
05-17-2008, 03:23 PM
BTW, while this thread has continued to 33 pages, does anyone want to guess who currently leads the Reds in RBI?

In the land of the blind, the one eyed man is king.

pahster
05-17-2008, 03:28 PM
BTW, while this thread has continued to 33 pages, does anyone want to guess who currently leads the Reds in RBI?

Time to move him down to the 9-hole. :p:

KronoRed
05-17-2008, 04:54 PM
Time to move him down to the 9-hole. :p:

That's where RBI guys go.

MWM
05-17-2008, 07:59 PM
So walks in the run to win the game yesterday; hits a GW 3 run homer today. Now if we can just figure out a way for him to win the game on a strikeout tomorrow, it would complete the trifecta. :evil:

jojo
05-17-2008, 08:00 PM
So walks in the run to win the game yesterday; hits a GW 3 run homer today. Now if we can just figure out a way for him to win the game on a strikeout tomorrow, it would complete the trifecta. :evil:

Bases loaded, score tied, bottom of the 9th, full count.....swings and misses....Martinez lets it go through his legs....Dunn to first...game over.