Why no love for BA w/RISP

[kaldaniels]

There may be an easy answer to this...but why is that stat ridiculed so much.

Looking at it from a football perspective, there are some QB's who I want on my team with less than 2 minutes and my team trailing...doesn't it stand to reason there are some players who step up more than others in critical situations in baseball as well?

It seems as this stat is assumed to be bogus by many, but I haven't seen the explanation why. Sorry if this is a repetitive issue.

[edabbs44]

http://www.archangelfireworks.com/eMerchantPro/pc/catalog/Powder%20Keg.jpg

[_Sir_Charles_]

I'm on your side on this one Kal. I'm no sabermetric stat freak, so I'll leave the debate to others. But I do feel that ba/risp and simple ba both get overlooked by saber statheads.

[Tony Cloninger]

It has something to do with some guy named Dunn...and then the rest becomes a real love fest between the people who think it's rabbit season and those who think it's duck season.

[wolfboy]

There's been plenty of bait planted in this thread, so I'll do my best to side step that and give the original poster my take on this topic. I'll preface this by saying that I favor statistical analysis when it comes to baseball, but I'm barely a novice. To be sure, your very question has been debated on here many times in the past. There are better arguments out there than I can offer.

For me, BA RISP is defective in the same manner that BA is defective. BA gives an incomplete and often misleading indication of the value of a hitter. At any point in the game, I want a hitter who is not making outs. That's even more important when runners are on. BA or BA RISP does not accurately show you which hitters are able to avoid outs.

A player like Adam Dunn is a prime example why BA RISP is flawed. Ideally you'd like to have a guy that has a high BA, OBP and SLG in any situation. Those guys are rare. A guy like Dunn is pretty good in the second and third category, but not necessarily good in the 1st. People point out a low BA RISP and use that as the sole measure of his worth in the lineup. However, that only paints a partial picture of what is going on. The BA might be low, but the OBP and SLG are high. What this tells you is that the player is not getting the number of hits (BA) you might want with RISP, but he's avoiding outs (high OBP) and when he does hit, it's for power (high SLG).

Under just one measure (BA), it looks like the guy is worthless. When you factor in the other measures (OBP and SLG), the picture of performance doesn't look so bleak.

So, to answer your question, I think BA RISP gets a lot of negative reaction around here because it's often used as a critically important measure of a player. A lot of posters on the board strongly disagree. When a debate occurs about Dunn's value to the team, it begins with BA RISP and goes from there. Some view it as the be all end all, and others take it with a grain of salt.

I think the great misunderstanding in all of this is that the alleged "stat freaks" would dismiss BA RISP if it were used to discredit Dunn or any other player in baseball. Yet, some people tend to think that the only reason we don't like BA RISP is because of Dunn. It's curious.

[kaldaniels]

There's been plenty of bait planted in this thread, so I'll do my best to side step that and give the original poster my take on this topic. I'll preface this by saying that I favor statistical analysis when it comes to baseball, but I'm barely a novice. To be sure, your very question has been debated on here many times in the past. There are better arguments out there than I can offer.

For me, BA RISP is defective in the same manner that BA is defective. BA gives an incomplete and often misleading indication of the value of a hitter. At any point in the game, I want a hitter who is not making outs. That's even more important when runners are on. BA or BA RISP does not accurately show you which hitters are able to avoid outs.

A player like Adam Dunn is a prime example why BA RISP is flawed. Ideally you'd like to have a guy that has a high BA, OBP and SLG in any situation. Those guys are rare. A guy like Dunn is pretty good in the second and third category, but not necessarily good in the 1st. People point out a low BA RISP and use that as the sole measure of his worth in the lineup. However, that only paints a partial picture of what is going on. The BA might be low, but the OBP and SLG are high. What this tells you is that the player is not getting the number of hits (BA) you might want with RISP, but he's avoiding outs (high OBP) and when he does hit, it's for power (high SLG).

Under just one measure (BA), it looks like the guy is worthless. When you factor in the other measures (OBP and SLG), the picture of performance doesn't look so bleak.

So, to answer your question, I think BA RISP gets a lot of negative reaction around here because it's often used as a critically important measure of a player. A lot of posters on the board strongly disagree. When a debate occurs about Dunn's value to the team, it begins with BA RISP and goes from there. Some view it as the be all end all, and others take it with a grain of salt.

I think the great misunderstanding in all of this is that the alleged "stat freaks" would dismiss BA RISP if it were used to discredit Dunn or any other player in baseball. Yet, some people tend to think that the only reason we don't like BA RISP is because of Dunn. It's curious.

Thats a fair enough start...and please don't think my question was baited at all...I realize Dunn's name was brought up but I don't want this thread to concern him at all. I'm truly looking for some understanding on this. I for one do realize that BA is flawed and overused to evalute hitters, so along with that it stands to reason BA w/ RISP does the same...because it doesn't account for "not making outs"

What about then, a stat that is "OBP with RISP or OPS with RISP". Couldn't there be something that analyzes whether a player is better or worse with RISP.

I don't look at batting stats with runners on as the end all to evaluate hitters by any measure. But shouldn't it make up a piece of the puzzle???

[AtomicDumpling]

The reason BA w/RISP is not considered a meaningful stat is because it is not a predictive statistic.

You are much better off using the player's overall batting average to predict his performance w/RISP than just using his actual batting average w/RISP.

It has been shown in many studies that over their careers players will hit the same batting average with and without runners on base once a meaningful sample size has been reached. When you start evaluating batting averages in various situational scenarios you are reducing the sample size to the point where the results of your evaluation are meaningless.

Only about 25% or less of a players plate appearances come with runners in scoring position. That percentage varies widely based on a hitter's spot in the lineup and the ability of his teammates to get on base in front of him. For example, when Brandon Phillips comes to the plate he has RISP about 30% of the time, while Jerry Hairston has RISP only 18% of the time.

Of course, we know that batting average is not a good barometer of a player's offensive performance in the first place, but that is another story.

[wolfboy]

I was never saying that you use BA with RISP as a be all end all. Some have over the years, and I think that's why it gets such negative treatment on this board. To answer your question, there is a measure of OPS with RISP. It's often pointed to as a response to Dunn's poor showing in the BA with RISP column. I'm going to re-post something AtomicDumpling made in another thread that involved a debate over Dunn.
http://www.redszone.com/forums/showpost.php?p=1675206&postcount=32

Even when Dunn is in one of his slumps in terms of batting average, he is still making fewer outs than just about anybody else on the team. His slumps don't look so bad when you look beyond batting average towards OBP, SLG, RBIs and Runs.

Regarding Dunn's "clutchiness" maybe these career numbers will shed some light:

Overall - .898 OPS
2 outs, RISP - .854 OPS
Late & Close - .875 OPS
Tie Game - .908 OPS
Within 1 R - .895 OPS
Within 2 R - .914 OPS
Within 3 R - .914 OPS
Within 4 R - .910 OPS
Margin > 4 R - .833 OPS

Compare those numbers to the Reds' overal team OPS of .726 this year (which includes Dunn's OPS score remember).

You can see that Dunn is fantastic no matter what the score is. He does not choke in clutch situations as some would have you believe. The only time his OPS goes down noticeably from his normal rate is when the game is already out of hand. Even then though his OPS is still excellent.

One thing I noticed when viewing his situational stats is that his OPS goes up considerably when there is a man on first base. Presumably that runner on first discourages the opponent from pitching around him, so he gets better pitches to hit.

When there are runners on base, but not on 1st base, his OBP goes through the roof because the other team just walks him intentionally or semi-intentionally to take the bat out of his hands. Dunn's OBP with runners on just 2nd and 3rd is an ungodly .520! With a runner on 3rd only his OBP is .466 and with a runner on 2nd only it is .423, while with nobody on base at all it is "only" .358 for his career. Why would you pitch to Dunn in the clutch if you could pitch to a much inferior hitter instead?

I strongly believe Dunn's production would increase dramatically if he had an excellent hitter batting after him. He has never had much protection in the lineup, which just makes it too easy for the opponent to pitch around him. He just doesn't get that many true hitter's pitches. Most of the strikes he sees are on the corners. The pitchers are content to nibble on the corners and are not upset if he walks. He needs somebody behind him to make the pitcher pay for those walks. His career OBP of .381 would score a lot more runs if there were good hitters after Dunn in the lineup.

[wolfboy]

The reason BA w/RISP is not considered a meaningful stat is because it is not a predictive statistic.

You are much better off using the player's overall batting average to predict his performance w/RISP than just using his actual batting average w/RISP.

It has been shown in many studies that over their careers players will hit the same batting average with and without runners on base once a meaningful sample size has been reached. When you start evaluating batting averages in various situational scenarios you are reducing the sample size to the point where the results of your evaluation are meaningless.

Only about 25% or less of a players plate appearances come with runners in scoring position. That percentage varies widely based on a hitter's spot in the lineup and the ability of his teammates to get on base in front of him. For example, when Brandon Phillips comes to the plate he has RISP about 30% of the time, while Jerry Hairston has RISP only 18% of the time.

Of course, we know that batting average is not a good barometer of a player's offensive performance in the first place, but that is another story.

As I stated before, there are those who are much more knowledgeable and can articulate the argument much better than I.

[sonny]

I just want the Reds to score more than their opponents in each game. How it's done really doesn't bother me. Lets keep it simple! :thumbup:

[wolfboy]

I just want the Reds to score more than their opponents in each game. How it's done really doesn't bother me. Lets keep it simple! :thumbup:

I think we can all agree with that one. :beerme:

[kaldaniels]

The reason BA w/RISP is not considered a meaningful stat is because it is not a predictive statistic.

You are much better off using the player's overall batting average to predict his performance w/RISP than just using his actual batting average w/RISP.

It has been shown in many studies that over their careers players will hit the same batting average with and without runners on base once a meaningful sample size has been reached. When you start evaluating batting averages in various situational scenarios you are reducing the sample size to the point where the results of your evaluation are meaningless.

Only about 25% or less of a players plate appearances come with runners in scoring position. That percentage varies widely based on a hitter's spot in the lineup and the ability of his teammates to get on base in front of him. For example, when Brandon Phillips comes to the plate he has RISP about 30% of the time, while Jerry Hairston has RISP only 18% of the time.

Of course, we know that batting average is not a good barometer of a player's offensive performance in the first place, but that is another story.

For most players I would assume that sounds correct, and I'm sure it is. But are there any players out there who over a large enough sample size do have a significant difference? If so, wouldn't it help to identify those players.

[AtomicDumpling]

I just want the Reds to score more than their opponents in each game. How it's done really doesn't bother me. Lets keep it simple! :thumbup:

Agreed.

Some people have the urge to focus on certain situational scenarios while ignoring the other situational scenarios.

There are dozens of different offensive situations that can occur. For the purposes of an illustration, let's suppose there were only four scenarios: A, B, C and D.

Brutus Sledgehammer is an awful player in scenario A, coming through only 10% of the time. Many people conclude that Brutus is a poor player because he fails so often in that situation. What those people fail to consider is that Brutus is average in scenario B and excellent in scenarios C and D. When you add up his performance in all four scenarios, Brutus' overall production is well above average. So the people that focus on scenario A are developing an inaccurate appraisal of Brutus Sledgehammer's hitting ability.

You can't focus on situational statistics like BA w/RISP because those stats represent a small portion of the offensive scenarios. Some players have low batting averages w/RISP, but because they are so good in other situations they can overcome that one poor scenario to deliver excellent overall offensive production.

[wolfboy]

Agreed.

Some people have the urge to focus on certain situational scenarios while ignoring the other situational scenarios.

There are dozens of different offensive situations that can occur. For the purposes of an illustration, let's suppose there were only four scenarios: A, B, C and D.

Brutus Sledgehammer is an awful player in scenario A, coming through only 10% of the time. Many people conclude that Brutus is a poor player because he fails so often in that situation. What those people fail to consider is that Brutus is average in scenario B and excellent in scenarios C and D. When you add up his performance in all four scenarios, Brutus' overall production is well above average. So the people that focus on scenario A are developing an inaccurate appraisal of Brutus Sledgehammer's hitting ability.

You can't focus on situational statistics like BA w/RISP because those stats represent a small portion of the offensive scenarios. Some players have low batting averages w/RISP, but because they are so good in other situations they can overcome that one poor scenario to deliver excellent overall offensive production.

Well said AD. I might add that I don't see any problem with analyzing, criticizing or praising a player based upon one scenario. However, a common trap I've seen is constructing an overall analysis of a player's worth to the team based upon one scenario combined with some anecdotal evidence.

[cincrazy]

There's been plenty of bait planted in this thread, so I'll do my best to side step that and give the original poster my take on this topic. I'll preface this by saying that I favor statistical analysis when it comes to baseball, but I'm barely a novice. To be sure, your very question has been debated on here many times in the past. There are better arguments out there than I can offer.

For me, BA RISP is defective in the same manner that BA is defective. BA gives an incomplete and often misleading indication of the value of a hitter. At any point in the game, I want a hitter who is not making outs. That's even more important when runners are on. BA or BA RISP does not accurately show you which hitters are able to avoid outs.

A player like Adam Dunn is a prime example why BA RISP is flawed. Ideally you'd like to have a guy that has a high BA, OBP and SLG in any situation. Those guys are rare. A guy like Dunn is pretty good in the second and third category, but not necessarily good in the 1st. People point out a low BA RISP and use that as the sole measure of his worth in the lineup. However, that only paints a partial picture of what is going on. The BA might be low, but the OBP and SLG are high. What this tells you is that the player is not getting the number of hits (BA) you might want with RISP, but he's avoiding outs (high OBP) and when he does hit, it's for power (high SLG).

Under just one measure (BA), it looks like the guy is worthless. When you factor in the other measures (OBP and SLG), the picture of performance doesn't look so bleak.

So, to answer your question, I think BA RISP gets a lot of negative reaction around here because it's often used as a critically important measure of a player. A lot of posters on the board strongly disagree. When a debate occurs about Dunn's value to the team, it begins with BA RISP and goes from there. Some view it as the be all end all, and others take it with a grain of salt.

I think the great misunderstanding in all of this is that the alleged "stat freaks" would dismiss BA RISP if it were used to discredit Dunn or any other player in baseball. Yet, some people tend to think that the only reason we don't like BA RISP is because of Dunn. It's curious.

Awesome post

[Spring~Fields]

It has something to do with some guy named Dunn...and then the rest becomes a real love fest between the people who think it's rabbit season and those who think it's duck season.

:clap::clap:

:lol::lol:

That was good, :)

[OldRightHander]

Once again, any runner on base is a RISP.

[paintmered]

There's no love for BA w/RISP because it plays only a minor role in the Runs Scored equation. Advocating for the stat is like advocating for the actor playing Boy #2 to win an Oscar.

[SteelSD]

The reason BA w/RISP is not considered a meaningful stat is because it is not a predictive statistic.

Well, that and the fact that Batting Average isn't really what drives RBI acquisition (and Run scoring for that matter). Over the past three years, here are the correlations for MLB:

BA to RBI: 0.71
OBP to RBI: 0.83
SLG to RBI: 0.86
OPS to RBI: 0.93

As usual, what you want are a group of hitters who're able to avoid Outs and acquire a lot of Bases while they're avoiding Outs.

[flyer85]

this is all that matters

http://www.cookiemadness.net/Rice%20Krispie%20Treat.jpg

[westofyou]

Why?

Just because it represents 24% of all at bats is a great start.

[Spring~Fields]

There may be an easy answer to this...but why is that stat ridiculed so much.

Looking at it from a football perspective, there are some QB's who I want on my team with less than 2 minutes and my team trailing...doesn't it stand to reason there are some players who step up more than others in critical situations in baseball as well?

It seems as this stat is assumed to be bogus by many, but I haven't seen the explanation why. Sorry if this is a repetitive issue.

I think on base percentage is more fair to evaluate a batter by than batting average alone. The short, check the BA w RISP, and I think that you will find that those are the ones with the higher OBP, the highest percentage of time. OBP with RISP is what I want to see from the batters.

OBP, it has BA in the formula to give them credit for a good batting average, and it gives them credit for getting on base in other ways. The batter that is hitting well with runners in scoring position probably are the ones with the higher OBP to begin with.

OBP
On-base percentage
(H + BB + HBP) divided by (AB + BB + HBP + SF)

Avg or BA is short, it does not give the batter full credit for when he is doing his job of getting on base in other ways, or other times.

Batting average
H divided by AB

Each batter is doing more than just H/AB in each of his at bats over the season.

The batter that has a high BA w RISP is probably the one with the higher OBP to begin with, no not always, there is no always in baseball, it is a game of failures, situations, and variables that have it successes measured in percentages. The other night when Pittsburgh was scoring against the Reds it was the players with the higher SLG that were doing the damage of getting on base and driving the others in.

BA won’t tell you the full total story on a batter percentage wise of getting on base, which is his job and indicates whether he is good at his job or not while performing the role of batter.

Dunn for example is pitched around a lot by the various pitchers as evidenced by his league leading walks. 68 walks, Dunn will not get credit for getting on base, which is the job of a batter, in his BA, neither will other batters around the league, that’s not right, all positives or negatives must be accounted for, before one can say player x or player y is this or that, or was this or that etc. Each act with the AB or PA must be measured to asses fair judgments of value and worth too.

The numbers or stats reflect a percentage of the time, a batter is doing or not doing something a percentage of the times in his PA or AB

I even judge the managers decisions and choices as graded by the various batters OBP plus or minus, of course the crowd does not like any input that implies a manager can effect runs scored by his decisions, but that is another story.



Those clutch quarterbacks are going to be the ones with the higher percentages of positive passing stats, more often than not, with solid line and running game behind them too. Just like certain football stats don't tell the whole story, also we can get a false read from observations alone, and faulty memory or filtering of the minds perceptions.

[Spring~Fields]

this is all that matters

http://www.cookiemadness.net/Rice%20Krispie%20Treat.jpg

Looks like "Bakers Choice" lineup on the buffet table to me. :) sorely lacking in OBP and SLG :all_cohol

[Spring~Fields]

I just want the Reds to score more than their opponents in each game. How it's done really doesn't bother me. Lets keep it simple! :thumbup:

Dusty is that you? :D

Well, just keep giving your lowest on base percentage guys the most PA and AB, and well, as Hank was telling me just the other day. Low scoring affairs are good, if your the one on top.

[Spring~Fields]

For me, BA RISP is defective in the same manner that BA is defective. BA gives an incomplete and often misleading indication of the value of a hitter. At any point in the game, I want a hitter who is not making outs. That's even more important when runners are on. BA or BA RISP does not accurately show you which hitters are able to avoid outs.

A player like Adam Dunn is a prime example why BA RISP is flawed. Ideally you'd like to have a guy that has a high BA, OBP and SLG in any situation. Those guys are rare. A guy like Dunn is pretty good in the second and third category, but not necessarily good in the 1st. People point out a low BA RISP and use that as the sole measure of his worth in the lineup. However, that only paints a partial picture of what is going on. The BA might be low, but the OBP and SLG are high. What this tells you is that the player is not getting the number of hits (BA) you might want with RISP, but he's avoiding outs (high OBP) and when he does hit, it's for power (high SLG).

Under just one measure (BA), it looks like the guy is worthless. When you factor in the other measures (OBP and SLG), the picture of performance doesn't look so bleak.

So, to answer your question, I think BA RISP gets a lot of negative reaction around here because it's often used as a critically important measure of a player.


I think this is an excellent explanation.

The Dunn example always troubles for the reasons that you have addressed and the other is that Dunn is not getting cited for or receiving credit from the fans, for the fear or respect that opposing pitchers have of Dunn, they simply avoid throwing him strikes to hit as evidenced by his highest walk total in the league.

That implies to me that the pitchers believe that if they did throw him an increase of strikes that Dunn would in probability hit the pitches, and that would hurt them, simultaneously it would increase the BA part of Dunn’s catalogue in hitting.

[Wheelhouse]

BA w/RISP is an important offensive stat because a base hit generally means at least one run scores. A walk rarely drives in a run. And a strikeout with RISP almost never scores a run. A base hit w/RISP is by far the most productive result of the at-bat.

[Spring~Fields]

BA w/RISP is an important offensive stat because a base hit generally means at least one run scores. A walk rarely drives in a run. And a strikeout with RISP almost never scores a run. Team with the most runs wins.

Don't you already have that if your highest OBP batters are stacked in an order? Granted they might not be the leaue leading OBP hitters, but none the less aren't the highest OBP ones going to be the ones with BA w RISP most of the time?

Tonight the guy with the highest BA w RISP is 0-3 now what ?

[OldRightHander]

BA w/RISP is an important offensive stat because a base hit generally means at least one run scores. A walk rarely drives in a run. And a strikeout with RISP almost never scores a run. A base hit w/RISP is by far the most productive result of the at-bat.

How often do we see a batter take a two out walk and then the next guy up gets the hit? The guy who drew the walk didn't drive the run in, but made it possible for the next guy to because he didn't make an out.

[Spring~Fields]

BA to RBI: 0.71
OBP to RBI: 0.83
SLG to RBI: 0.86
OPS to RBI: 0.93

As usual, what you want are a group of hitters who're able to avoid Outs and acquire a lot of Bases while they're avoiding Outs.

So we want the batters or players on our team with the higher OBP and SLG on their resumes.

[Wheelhouse]

Don't you already have that if your highest OBP batters are stacked in an order? Granted they might not be the leaue leading OBP hitters, but none the less aren't the highest OBP ones going to be the ones with BA w RISP most of the time?

Tonight the guy with the highest BA w RISP is 0-3 now what ?

Sure a good deal of the time there is a correlation. But quite often there are deviations. Extreme deviations.

[Wheelhouse]

How often do we see a batter take a two out walk and then the next guy up gets the hit? The guy who drew the walk didn't drive the run in, but made it possible for the next guy to because he didn't make an out.

Em, well, on this Reds team, rarely, sad to say...

[jojo]

So we want the batters or players on our team with the higher OBP and SLG on their resumes.

Only if you want to score a lot. Otherwise, not so much. :cool:

[Spring~Fields]

Only if you want to score a lot. Otherwise, not so much. :cool:

I had to look twice, got me .........:D

[jojo]

I had to look twice, got me .........:D

And the beautiful thing about offense is that generally it's the least risky commodity that a GM can spend his money on...

[Spring~Fields]

Sure a good deal of the time there is a correlation. But quite often there are deviations. Extreme deviations.

Yes the variables that come into play. Those variables are what get's us almost every time when we write or say something positive or negative about another or a player, situations have variables, and those individuals or players will go out do or achieve the opposite of what we were praising or being negative about in the next set of situations that they face, and no not always, it’s variable.

[Reds Nd2]

Over the past three years, here are the correlations for MLB:

BA to RBI: 0.71
OBP to RBI: 0.83
SLG to RBI: 0.86
OPS to RBI: 0.93

The correlations between OBP and SLG are fairly close. I always thought that OBP was more valuable than SLG when it came to the acquisition of runs. I understand that your using RBI's, but are the correlations between the two that close when using a teams total runs scored over the past three seasons? I guess my real question is this. Has SLG overtaken OBP in matters of scoring runs on the team level?


As usual, what you want are a group of hitters who're able to avoid Outs and acquire a lot of Bases while they're avoiding Outs.

Agreed.

[Spring~Fields]

And the beautiful thing about offense is that generally it's the least risky commodity that a GM can spend his money on...

comparison to pitching as in cost/benefit analysis, or risk analysis?

Risk is hard to measure, all ventures with the players are going to have risk of known and unknown.

I don't know if that is true. Let's say the general manager happens to get lucky and is able to have a stable of solid pitchers that look good to the market place looking in the window, the value of pitching has changed such that the value of return might well exceed perceived risk as in the past of developing pitching over a offensive player, one is a slow investment with long term growth, the offensive player is more like a liquidity type investment.

The market will pay more for a good pitcher than a good batter. Market pressures by supply and demand.

[Spring~Fields]

And the beautiful thing about offense is that generally it's the least risky commodity that a GM can spend his money on...

Risk adversity increases if there are limited resources, or if one has less operating capital than another, even with GMs.

[RedsManRick]

Just to add fuel to the fire...

"RISP" includes only runners on 2nd and 3rd base. However, there are often runners on first base and the reality of the game is that these runners can score on a hit too. The "RISP" concept is more relevant, if still not terribly useful, in a station to station game with fewer extra base hits - a game that hasn't been around in quite some time.

All runners, including the batter, are in scoring position; You just need the right sort of hit to plate them. Any stat that ignores a good portion of runners, ignores the types of hits,and further ignores where the batter leaves himself for the next batter really misses the boat on assessing run production or clutch.

We can't infer anything about a player's general abilities from BA w/ RISP and it hardly summarizes how well the guy performs in situations where a run could be driven in. It's not that it's bad, it's just that it really doesn't tell us very much, and could even be misleading.

[Spring~Fields]

I will be the bad guy. The jerk with the bad question.

So what do we do with those batters that invariably come to the plate when our Reds need a score, and they have Patterson-itus, and make outs, and our team ends up with runners left on base and our team goes on to lose, and we get, well you know................kind of upset when it seems to happen over and over again with those batters?

[RedsManRick]

I will be the bad guy. The jerk with the bad question.

So what do we do with those batters that invariably come to the plate when our Reds need a score, and they have Patterson-itus, and make outs, and our team ends up with runners left on base and our team goes on to lose, and we get, well you know................kind of upset when it seems to happen over and over again with those batters?

Not surprisingly, the guys who tend to make outs in RBI situations with regularity tend to be the guys who make outs regularly in all situations. The answer is to get them out of the lineup and/or bury them at the bottom where they'll have the fewest RBI opportunities and PA in general.

[wolfboy]

I will be the bad guy. The jerk with the bad question.

So what do we do with those batters that invariably come to the plate when our Reds need a score, and they have Patterson-itus, and make outs, and our team ends up with runners left on base and our team goes on to lose, and we get, well you know................kind of upset when it seems to happen over and over again with those batters?

I think that's the anecdotal evidence I had talked about earlier. Dunn strikes out with runners on in a "clutch" situation. It happens another handful of times. These instances really stick in your head because they really hurt as a die hard fan. You also see that the guy doesn't register well on the BA with RISP measure. All of a sudden, you start to feel like Dunn is a bum.
At the same time, what's forgotten is the time he belts the ball out with two guys on (high SLG) or when he takes a walk and Encarnacion or someone else drives the runners in right after him (high OBP). These are just as important to note as the strikeouts, etc...
Unfortunately, it's that dangerous combination of a non-predictive statistical measure combined with anecdotal evidence that gets people down on guys like Dunn. I'm not saying that's the only reason people don't care for him. It's a pretty common one though.

[kaldaniels]

May I ask this....If you had a choice....

Would you rather....
1) Have a player with a higher OPS with RISP
2) " " " " " less " " "
3)Same OPS in all situations

Or does it not matter at all...

[Mario-Rijo]

I think this has been a really good thread with some easy to understand explanations on the subject. I would say I have a better understanding of the situation.

That said I would ask this question w/ regards to Dunn. It's presumed that if Dunn had..let's just say Ted Williams hitting behind him that his #'s overall would improve. Does anyone that believes that to be a fact give any consideration to the idea that his overall #'s would perhaps get worse? Might for example his BA and Slg get slightly better but his OBP get marginally worse and therefore his overall OPS be worse?

Let me be very clear here, I am not stating or even inferring that his numbers would get worse. What I am asking is if anyone else think it's possible, I am actually on the fence about it.

[Spring~Fields]

May I ask this....If you had a choice....

Would you rather....
1) Have a player with a higher OPS with RISP
2) " " " " " less " " "
3)Same OPS in all situations

Or does it not matter at all...

I think that you are going to have to ask that question more clearly and directly to get the responses that you may be looking for. Personally I am not sure how to interpret what you are saying exactly in your illustration there.

“Would you rather.…” ?
1) Have a player with a higher OPS with RISP ?
2) " " " " " less " " " ?
3)Same OPS in all situations ?

Or does it not matter at all…?

I personally would rather have very high on base percentage all through the entire lineup, if we were allowed to cherry pick and have our way, that is what I would prefer. Of course I am assuming I have the better batters and that a higher SLG will accompany those with higher OBP, which both combine to make OPS

You do realize that these guys I want, may also strikeout, or make ground ball or fly outs a lot too with RISP? Even though they have the higher percentages working in their favor? Of course there are times when fly out or ground out is superior to a strikeout, runner on second or third might be a case.

[kaldaniels]

I think that you are going to have to ask that question more clearly and directly to get the responses that you may be looking for. Personally I am not sure how to interpret what you are saying exactly in your illustration there.

“Would you rather.…” ?
1) Have a player with a higher OPS with RISP ?
2) " " " " " less " " " ?
3)Same OPS in all situations ?

Or does it not matter at all…?

I personally would rather have very high on base percentage all through the entire lineup, if we were allowed to cherry pick and have our way, that is what I would prefer. Of course I am assuming I have the better batters and that a higher SLG will accompany those with higher OBP, which both combine to make OPS

You do realize that these guys I want, may also strikeout, or make ground ball or fly outs a lot too with RISP? Even though they have the higher percentages working in their favor? Of course there are times when fly out or ground out is superior to a strikeout, runner on second or third might be a case.

I am talking about analyzing a single player...not the whole lineup.

Lets say you have a .275 BA and .800 OPS for argument sake

Would you rather have him hit...

.300 BA and .900 OPS w/RISP and .250 and .700 with no runners on

.250 BA and .700 w/RISP and .300 and .900 with bases empty

or .275 and .800 equally in all situations

[AtomicDumpling]

I think this has been a really good thread with some easy to understand explanations on the subject. I would say I have a better understanding of the situation.

That said I would ask this question w/ regards to Dunn. It's presumed that if Dunn had..let's just say Ted Williams hitting behind him that his #'s overall would improve. Does anyone that believes that to be a fact give any consideration to the idea that his overall #'s would perhaps get worse? Might for example his BA and Slg get slightly better but his OBP get marginally worse and therefore his overall OPS be worse?

Let me be very clear here, I am not stating or even inferring that his numbers would get worse. What I am asking is if anyone else think it's possible, I am actually on the fence about it.

I think his OPS would go up a fair amount, but his run production would increase even more. Since he would walk less his opportunities to hit doubles and home runs would increase, which would lead to extra runs and RBIs.

If he cut his walk rate by half while keeping his SLG at his current rate, he would hit 5-7 more home runs per season.

Generally the better the hitter batting after you, the better the pitches you will get to hit.

The reasoning is because the pitcher can't afford to pitch around Dunn because he will be facing a tough hitter in the next slot. It is better to face that tough hitter with the bases empty rather than having Dunn on first base because you pitched around him and walked him.

Conversely, if Dunn is at the plate with a weak hitter up next, the pitcher will just nibble at the corners without giving Dunn anything very good to swing at. He figures either Dunn will chase a bad pitch and get himself out, or Dunn will walk and and the pitcher gets to face the easier out instead.

So if Dunn were followed in the lineup by Albert Pujols he would have the opportunity to swing at better hitting pitches than if he were followed by Corey Patterson.

It only makes sense that your batting average and slugging percentage would rise if you have meatier pitches to swing at instead of being forced to swing at pitches that only graze the plate if they are strikes at all. So in the end, Dunn would probably strike out less, have a better BA, same or slighty lower OBP and a higher SLG. It is all theory of course. It would be nice to test it by moving Dunn up to the 2nd or 3rd slot in the lineup, followed by Keppinger, Votto or Phillips instead of having Bako behind him like he has when hitting 7th most of the season.

[Spring~Fields]

That said I would ask this question w/ regards to Dunn. It's presumed that if Dunn had..let's just say Ted Williams hitting behind him that his #'s overall would improve. Does anyone that believes that to be a fact give any consideration to the idea that his overall #'s would perhaps get worse? Might for example his BA and Slg get slightly better but his OBP get marginally worse and therefore his overall OPS be worse?


I haven’t given any consideration to a reduction in his numbers with the types of batters who perform to such a level that it makes the opposing manager and his pitcher reconsider walking or pitching around Dunn.

I believe that the number of strikes that the opposing pitcher has to throw would go up to Dunn with the better hitters behind him, and that with the better number of pitches for Dunn to hit, a greater proportion of base hits, doubles and homeruns, and a reduction in strikeouts would follow, while he would still have a high number of walks, not as many but still a high number.

Not to be nasty or sarcastic, but why does anyone think that Griffey really wants to bat in the three hole, because his daddy did? Griffey is smart and learned the game well from daddy and others.

I believe that Griffey thinks or knows that he will get more pitches to drive if he has sluggers behind him and on base guys on in front of him, where better than the 3 slot can that happen for him?

The pitcher has to throw him a greater portion of strikes for a fear that if they walk him, the sluggers will make them pay, if there are runners already on base ahead of him, they have to pitch to him also. On a greater percentage, not all or nothing absolutes.

I would be shocked if Dunn batted in front of Griffey the rest of the way and his numbers did not improve across the board, but I feel that way about Encarncion and Votto too, but they are not going to give us a high enough sample size to find out, one way or another. Now I am not talking about just one or two nights, I am talking season.

[Spring~Fields]

I am talking about analyzing a single player...not the whole lineup.

Lets say you have a .275 BA and .800 OPS for argument sake

Would you rather have him hit...

.300 BA and .900 OPS w/RISP and .250 and .700 with no runners on

.250 BA and .700 w/RISP and .300 and .900 with bases empty

or .275 and .800 equally in all situations

I think Rick already answered that one for us, when he said



Not surprisingly, the guys who tend to make outs in RBI situations with regularity tend to be the guys who make outs regularly in all situations.

The opposite occurrence with the guys that hit, will continue to hit with or without runners on also.

In another words the batters that you have a preference for with the stats or numeric symbols that you are presenting will tend to hit along those same lines once they have reached their norms, that they will hit the same with runners on or runners off the bases. So again I am going to want to see my batter with the best OBP coming up. I want to avoid outs at the highest percentage of times possible.

[bucksfan2]

I like BA w/RISP. I like it because it gives you some insight on each singular AB in a certain situation. OBP can be important but at the same time a walk with RISP can be a negative occurance due to who is on deck batter. It also may put the defensive in an easier defensive position. I don't like OPS so much because it is a more weighted stat. In a given situation a single, double, or triple may provide the same outcome but are substantially valued differently. On a particular atbat a single provides 2, double 3, and triple, 4 to OPS but at the same time they all drive in the same amount of runs. The key so much isn't seing how many bases you can get rather driving in the runs on the bases.

[OldRightHander]

I like BA w/RISP. I like it because it gives you some insight on each singular AB in a certain situation. OBP can be important but at the same time a walk with RISP can be a negative occurance due to who is on deck batter. It also may put the defensive in an easier defensive position. I don't like OPS so much because it is a more weighted stat. In a given situation a single, double, or triple may provide the same outcome but are substantially valued differently. On a particular atbat a single provides 2, double 3, and triple, 4 to OPS but at the same time they all drive in the same amount of runs. The key so much isn't seing how many bases you can get rather driving in the runs on the bases.

Yes, with a runner on third any kind of hit short of a homer will give the same result in that individual situation, but what kind of hit puts the offense in better shape to score another run? I'll take the extra base hit any day. BA w/RISP might give you an indication of what to expect just in that at bat, while OPS is a better indicator of what you can expect over the course of the inning, or an entire game, series, or season.

[RedsManRick]

I like BA w/RISP. I like it because it gives you some insight on each singular AB in a certain situation. OBP can be important but at the same time a walk with RISP can be a negative occurance due to who is on deck batter. It also may put the defensive in an easier defensive position. I don't like OPS so much because it is a more weighted stat. In a given situation a single, double, or triple may provide the same outcome but are substantially valued differently. On a particular atbat a single provides 2, double 3, and triple, 4 to OPS but at the same time they all drive in the same amount of runs. The key so much isn't seing how many bases you can get rather driving in the runs on the bases.

Bucks, I understand your perspective. But isn't the definition of RISP, excluding runners on 1B, what makes all the various hit types more or less equal? Shouldn't runners on first base count? And if so, would BA be a poor way to measure a player's ability to drive them in?

I guess my question is this: What exactly is BA w/ RISP trying to measure?

Perhaps a better metric would be this:

Runners driven in from 3B / Runners on 3B
Runners driven in from 2B / Runners on 2B
Runners driven in from 1B / Runners on 1B

So it might look like: .58/.32/12.

That would really show how good of an "RBI" guy a player is, without rewarding empty base walks, without excluding meaningful runners, and counting productive outs.

[bucksfan2]

Bucks, I understand your perspective. But isn't the definition of RISP, excluding runners on 1B, what makes all the various hit types more or less equal? Shouldn't runners on first base count? And if so, would BA be a poor way to measure a player's ability to drive them in?

I guess my question is this: What exactly is BA w/ RISP trying to measure?

Perhaps a better metric would be this:

Runners driven in from 3B / Runners on 3B
Runners driven in from 2B / Runners on 2B
Runners driven in from 1B / Runners on 1B

So it might look like: .58/.32/12.

That would really show how good of an "RBI" guy a player is, without rewarding empty base walks, without excluding meaningful runners, and counting productive outs.

Here is the way I look at it. OPS while an offensive measure is the combination of two different stats. While this may be important the frequency of hits, BA, is lost within the formula. I am not really talking about the value of the hitters hit rather the value of getting that runner to score. There are no situations where I wouldn't want an XBH over a single but on the same hand I would want the higher frequency of hits with RISP.

OPS has such a high value for the HR that it tends to discount or even mask the other atbats. I think this becomes a bigger issue when you take the sample size down futher to atbats w/RISP. I think putting a value to a runner on 1b is tricky. How often does the runner on 1b score on a double, especially a double with less than 2 outs? Is a two out double that scores two that much greater than a two out single that scores 2?

[RedsManRick]

Here is the way I look at it. OPS while an offensive measure is the combination of two different stats. While this may be important the frequency of hits, BA, is lost within the formula.

I understand that you would rather have 4 singles with RISP than 1 HR and 3 outs. But BA is the primary component on OBP. And while a walk sometimes leaves a worse batter at the plate, it always creates an even better RISP opportunity -- and sometimes leaves a better batter at the plate!


I am not really talking about the value of the hitters hit rather the value of getting that runner to score. There are no situations where I wouldn't want an XBH over a single but on the same hand I would want the higher frequency of hits with RISP.

Fair enough. But you still haven't addressed the issue of not counting runners on first base. Why don't they count? To me, the "we only want to count times when any hit would score the runner" is fairly disingenuous because it ignores infield hits, shallow singles, and slow runners at 2B.

Again, if what we're really interested in measuring, as you indicate, is the ability of the batter to turn base runners in to runs, then why not just measure that directly with a % of runners driven in (by base) metric? Sure a metric would seem to better isolate that thing which you're trying to measure, including the RBI value of SLG, but without obscuring the value of the simple single when a single gets the job done just as well.

[westofyou]

Batting average is a great metric, but a flawed one that ignores secondary skills and fails to to count abstractly.

It's an abacus in a calculator world, it has its use but in the end it's just a using your fingers and toes system.

[edabbs44]

How often do we see a batter take a two out walk and then the next guy up gets the hit? The guy who drew the walk didn't drive the run in, but made it possible for the next guy to because he didn't make an out.

I bet we see it less than the next guy making an out.

[SteelSD]

OPS has such a high value for the HR that it tends to discount or even mask the other at bats.

What?

[RedsManRick]

What?

2 batters get 4 PA with RISP. For the sake of emphasis, let's say that all four times there are men on 2nd and 3rd.

Batter A hits 2 singles and drives in 4 runners.
Batter B hits 1 HR walks once, driving in 2 runners, excluding himself.

Batter A had a 1.500 OPS. Batter B had 1.833 OPS. Using OPS, Batter B was more productive. However, Batter A drove in more runs.

His point is that you can have a high OPS and not be that great at driving in runs compared to somebody with a lower OPS who doesn't walk and gets fewer XBH.

The question, of course, comes back to: what's the value in assessing just how well somebody plates base runners? And if this is truly our aim, why are we ignoring runners on first base?

Even a simple RBI/ROB percentage would seem to be a better metric than BA w/ RISP.

The bottom line for me with both batting average and the "RISP" designation (and all of their derivative stats) is the answer to the basic question: If this didn't already exist, would we recreate it? The answer, in both cases, is no. They were borne out of a different era, with less access to information and less understanding about the game.

A horse and buggy isn't a horrible way to get from point A to point B, but when you have a car in the driveway filled with gas (for the sake of analogy, the gas was free and causes no pollution), it's pretty hard to argue that it makes sense to use the buggy.

[edabbs44]

The bottom line for me with both batting average and the "RISP" designation (and all of their derivative stats) is the answer to the basic question: If this didn't already exist, would we recreate it? The answer, in both cases, is no. They were borne out of a different era, with less access to information and less understanding about the game.

A horse and buggy isn't a horrible way to get from point A to point B, but when you have a car in the driveway filled with gas (for the sake of analogy, the gas was free and causes no pollution), it's pretty hard to argue that it makes sense to use the buggy.

And round and round we go...:)

The problem with the car is that sometimes the owner doesn't want to stop for directions, taking away from the impact of having such a great mode of transportation. The horse and buggy might use Mapquest and get there quicker.

Player W's OPS is "x"? Great...then that OPS is equal to player Y's "x" OPS, even though player Y drove in more runs and performed better in more high leverage situations. That is the wrong way of evaluating a performance.

Some players get more out of their OPS.

[OldRightHander]

I bet we see it less than the next guy making an out.

Probably, but it still gave the offense a better chance to score than if he had gotten out, which was my whole point.

Often in a plate appearance you have the best possible outcome and the worst possible outcome and in between you have other outcomes with varying degrees of acceptability depending on the number of outs, where the runners are, etc. With a runner at third and a couple outs, a walk might not be as good as a hit, but it's certainly better than an out and shouldn't be discredited.

[RedsManRick]

And round and round we go...:)

The problem with the car is that sometimes the owner doesn't want to stop for directions, taking away from the impact of having such a great mode of transportation. The horse and buggy might use Mapquest and get there quicker.

Player W's OPS is "x"? Great...then that OPS is equal to player Y's "x" OPS, even though player Y drove in more runs and performed better in more high leverage situations. That is the wrong way of evaluating a performance.

Some players get more out of their OPS.

Eedabbs, you'll notice that I'm not making the case for OPS here as the car. My point is simply that if what you want to measure is how well a guy drives in runners, then there are better ways to do it than BA w/ RISP.

Start from scratch. The question is: what guys are the best at driving in runners who are on base? How you would measure that? Personally, I would look at the guys on base and then see how often the player drives them in. BA w/ RISP does not do this.

Further, you've clearly illustrated the point that many others here have made.


performed better in more high leverage situations
You've defined "performed better" in a very specific way -- namely, driving in runners -- completely ignoring the value of where the batter himself ends up. This is fine, if that's the question you want to ask, by all means. But in that case, BA w/ RISP is a pretty rough way to do it. If you meant some broader measurement of "doing things which generally lead to runs being produced", then BA w/ RISP is a very poor way of doing it as any assessment of overall performance must include the breadth of things a better can do productively and in proper proportion. Also, you've defined "high leverage" in a way that isn't really internally consistent. Why only count runners on 2B and 3B? Why should they count equally? Doesn't a walk with the bases loaded drive in a run? What if Albert Pujols is behind you and your walk moves guys from 1B and 2B to 2B and 3B? (among other questions).

BA w/ RISP implicitly asks a very specific question about converting base runners to runs and then attempts to answer it in a very roundabout manner which lacks outcome specificity (which can be isolated using a better crafted metric) and lacks the broad applicability of a more robust measure of general production. It simply doesn't do anything that we can't do better, easily, using some other method.

[SteelSD]

2 batters get 4 PA with RISP. For the sake of emphasis, let's say that all four times there are men on 2nd and 3rd.

I understand that you're just using that scenario for emphasis, but one of the main problems with arguments for the value of BA/RISP (not that you're producing one of those) is that they generally break down to hypothetical discussions of low-sample size near-impossible scenarios. Most hitters aren't going to see as many as 30 PA with runners on 2nd and 3rd (only) during a single season, much less four times in a single game.


Batter A hits 2 singles and drives in 4 runners.
Batter B hits 1 HR walks once, driving in 2 runners, excluding himself.

Batter A had a 1.500 OPS. Batter B had 1.833 OPS. Using OPS, Batter B was more productive. However, Batter A drove in more runs.

Ah, but Batter "B" only had three true opportunities due to the BB and we certainly can't exclude the hitter driving himself in. In that unlikely scenario, each hitter accounts for one RBI for 50% of their true opportunities. Batter A has an opportunity to plate 8 runners, and produces 4 RBI. Batter B has opportunity to plate 6 runners and produces 3 RBI. Both hitters produce the same percentage of RBI, and Batter B is more productive from a RC standpoint as well. While Batter B may have plated a lower percentage of the runners on in front of him, the fact that he substituted himself for one of those runners is of great value.


The question, of course, comes back to: what's the value in assessing just how well somebody plates base runners? And if this is truly our aim, why are we ignoring runners on first base?

Yep, and for that matter, why are we ignoring the hitter at the plate? And you're right of course. If BA w/RISP didn't already exist, there'd be absolutely no reason to build it.

[RedsManRick]

I understand that you're just using that scenario for emphasis, but one of the main problems with arguments for the value of BA/RISP (not that you're producing one of those) is that they generally break down to hypothetical discussions of low-sample size near-impossible scenarios. Most hitters aren't going to see as many as 30 PA with runners on 2nd and 3rd (only) during a single season, much less four times in a single game.

Preaching to the choir.



Ah, but Batter "B" only had three true opportunities due to the BB and we certainly can't exclude the hitter driving himself in. In that unlikely scenario, each hitter accounts for one RBI for 50% of their true opportunities. Batter A has an opportunity to plate 8 runners, and produces 4 RBI. Batter B has opportunity to plate 6 runners and produces 3 RBI. Both hitters produce the same percentage of RBI, and Batter B is more productive from a RC standpoint as well. While Batter B may have plated a lower percentage of the runners on in front of him, the fact that he substituted himself for one of those runners is of great value.

Outside of IBB, I hate the idea that walks don't get counted as at bats. I think this basic choice in defining "At Bats" separately from "Plate Appearances" was one of the worst decisions made in the history of baseball record keeping.

I understand that walks don't as at bats today, but the idea (conceptually) that a walk doesn't count as an opportunity to drive in those runs is BS. I understand that the player may have been pitched around, but the opportunity was there nonetheless, and that is an extremely slippery slope.

I still think the real challenge in this conversation is more clearly defining the "thing" we're trying to measure. Once we do that, we should be able to have a more productive discussion about how to go about measuring it.

I would love to see responses to the question(s): Why do you look at BA w/ RISP? What does it tell you? What are you hoping to learn?

There doesn't need to be any animosity or defensiveness. Let's figure this thing out together.

[bucksfan2]

The more and more I think about what is better BA/RISP, OPB/RISP, or OPS/RISP the more and more useless I think they all are.

I think that BA/RISP is a useful stat but what is it useful for? In all reality you don't set a lineup for certain situations. When you begin to do that you are overmanaging. You can't predict who is going to be on base and who is at the plate. If I had RISP w/ 2 outs I would want Keppinger up instead of Dunn for example. But in all reality how many times is Dusty going to have Keppinger and Dunn at his disposal as pinch hitters during the season? One or two games?

[RedsManRick]

The more and more I think about what is better BA/RISP, OPB/RISP, or OPS/RISP the more and more useless I think they all are.

I think that BA/RISP is a useful stat but what is it useful for? In all reality you don't set a lineup for certain situations. When you begin to do that you are overmanaging. You can't predict who is going to be on base and who is at the plate. If I had RISP w/ 2 outs I would want Keppinger up instead of Dunn for example. But in all reality how many times is Dusty going to have Keppinger and Dunn at his disposal as pinch hitters during the season? One or two games?

Exactly. Sometimes we get caught up in what we can measure and forget to think that the only reason we have metrics at all is because they help us make decisions better or understand the past differently.

Any time these sorts of debates about stats come up, I find it useful to take that step back and ask what we're really trying to accomplish. Often, simply answering that question will highlight a different approach that we might not have considered previously because we're so beholden to the way things have been done historically.

[edabbs44]

The more and more I think about what is better BA/RISP, OPB/RISP, or OPS/RISP the more and more useless I think they all are.

Those numbers are useless until you see people getting left on base. Then the game threads complain about people being left on base. But a month later, BA/RISP makes no difference.

[SteelSD]

Preaching to the choir.

Oh, I know.


I understand that walks don't as at bats today, but the idea (conceptually) that a walk doesn't count as an opportunity to drive in those runs is BS. I understand that the player may have been pitched around, but the opportunity was there nonetheless, and that is an extremely slippery slope.

Actually, the slippery slope is a position that essentially assumes that players should be able to turn Walks into Hits as a situation dictates.

[RedsManRick]

Oh, I know.



Actually, the slippery slope is a position that essentially assumes that players should be able to turn Walks into Hits as a situation dictates.

I never said the player should be able to convert them. Not all all. You assume too much. Performance with men on base isn't an answer of should have or shouldn't have. And that's really the core problem here. We want to tweak our stats to match the way people tend to interpret them rather than get people to simply interpret the stats they have correctly. RBI, or any derivative measure, isn't a judgment of how one should have approached the at bat to achieve the most productive outcome. It's merely a record of what has happened in a given situation.

That a player has walked tells us only that he had four balls called before either having three strikes or putting the ball in play. It does not tell us that he did or did not have an especially good opportunity to put the ball in play and advance the runners. Similarly, had he put the ball in play or strike out, that outcome does not suggests the player could not have worked a walk had he reacted differently to the pitches he saw.

If a pitcher throws every pitch 2 feet off the plate, how the batter reacts to those pitches doesn't change the nature of the opportunity. Why should we count the plate appearance as an at bat if he swings each time, but ignore it if he watched the pitches and takes a walk? The nature of the opportunity has not changed. If a player fouls of 10 straight pitches down the middle, and manages to see 4 balls in the process, how can we possibly pretend like that was not an opportunity? We can't. We can't tell the quality of the opportunity from the outcome of the opportunity with any reliability -- and should stop conflating the two.

I agree with your basic premise that when people walk, it is often the best outcome that could have been expected given the pitches they saw. However, that does negate the reality that a plate appearance is an opportunity to advance the runners home. We should first measure what happened and then assess why. Give Dunn a low RBI conversion rate and then explain the nature of his opportunities. But manipulating of the definition of an opportunity based on the outcome of said opportunity is simply misguided in my judgment.

The "Dunn problem" of having a relatively low RBI total (and conversion rate) relative to his overall level of productivity tells us just how limited RBI is in the first place in assessing player performance. RBI, or any derivative thereof, can only be used to assess how well the player advances runners given his opportunities to do so.

There is no purpose of sugarcoating or obscuring it. Dunn gets hits very rarely compared to most other players and this is especially true in RBI opportunities. The trick is properly understanding what we're precisely measuring so that we use the information appropriately. We shouldn't see Dunn's low RBI and conclude he's a poor run producer. Rather we should conclude that Dunn is not particularly adept at advancing runners already on base (due both to his natural skill set and the way he's pitched) and thus it would not be wise to utilize him such that the offense relies in his "RBI" ability.

[kaldaniels]

Thanks for all the interesting discussion on this thread.

Lets flip it around though...does it stand to reason that "inherited runners scored" is not a good stat for relievers then?

[SteelSD]

I never said the player should be able to convert them. Not all all. You assume too much. Performance with men on base isn't an answer of should have or shouldn't have. And that's really the core problem here. We want to tweak our stats to match the way people tend to interpret them rather than get people to simply interpret the stats they have correctly. RBI, or any derivative measure, isn't a judgment of how one should have approached the at bat to achieve the most productive outcome. It's merely a record of what has happened in a given situation.

I'm with you...


That a player has walked tells us only that he had four balls called before either having three strikes or putting the ball in play. It does not tell us that he did or did not have an especially good opportunity to put the ball in play and advance the runners. Similarly, had he put the ball in play or strike out, that outcome does not suggests the player could not have worked a walk had he reacted differently to the pitches he saw.

...and now I'm not. There's a reason Walks are counted as Plate Appearances and not At-Bats, and rightly so. It's because we can't assume the hitter actually had a reasonable opportunity to achieve a Base Hit.


If a pitcher throws every pitch 2 feet off the plate, how the batter reacts to those pitches doesn't change the nature of the opportunity. Why should we count the plate appearance as an at bat if he swings each time, but ignore it if he watched the pitches and takes a walk? The nature of the opportunity has not changed. If a player fouls of 10 straight pitches down the middle, and manages to see 4 balls in the process, how can we possibly pretend like that was not an opportunity? We can't. We can't tell the quality of the opportunity from the outcome of the opportunity with any reliability -- and should stop conflating the two.

We should count those AB opportunities because of hitter intent. Hitter A swings at bad pitches because he's attempting to plate a runner. Hitter B takes bad pitches because he's not willing to give into the pitching and defense in order to produce low-quality BIP events that aren't likely to plate a runner.

And exactly how many times have you seen a hitter foul off 10 pitches?


I agree with your basic premise that when people walk, it is often the best outcome that could have been expected given the pitches they saw. However, that does negate the reality that a plate appearance is an opportunity to advance the runners home. We should first measure what happened and then assess why. Give Dunn a low RBI conversion rate and then explain the nature of his opportunities. But manipulating of the definition of an opportunity based on the outcome of said opportunity is simply misguided in my judgment.

Misguided? You're actually the one attempting to manipulate the number of "opportunities" by adding BB to the mix. We know, with certainty, that the majority of the most dangerous hitters receive a lower percentage of Strikes per PA than less dangerous hitters across all situations. Do we somehow assumes that they receive more Strikes with ducks on the pond? Opposing teams take these guys out of play quite often- to the point that the opposition either won't pitch to them at all (IBB) or will attempt non-intentional Intentional Base on Balls. The best hitters don't give in.


The "Dunn problem" of having a relatively low RBI total (and conversion rate) relative to his overall level of productivity tells us just how limited RBI is in the first place in assessing player performance. RBI, or any derivative thereof, can only be used to assess how well the player advances runners given his opportunities to do so.

The "Dunn problem" is only a "problem" if we assume that he's supposed to be driving in runs when he's being pitched around. Otherwise, there's no "problem" at all.


There is no purpose of sugarcoating or obscuring it. Dunn gets hits very rarely compared to most other players and this is especially true in RBI opportunities. The trick is properly understanding what we're precisely measuring so that we use the information appropriately. We shouldn't see Dunn's low RBI and conclude he's a poor run producer. Rather we should conclude that Dunn is not particularly adept at advancing runners already on base (due both to his natural skill set and the way he's pitched) and thus it would not be wise to utilize him such that the offense relies in his "RBI" ability.

The fact that Dunn accurately takes pitches outside of the strike zone has nothing to do with being "adept" at converting RBI opportunities. He's just fine in converting given actual opportunity (see: AB). We should actually be looking at the guy as being someone who isn't going to receive a plethora of actual opportunity with RISP due to the low volume of Strikes he receives because pitchers simply don't want to face him. Due to his overall SLG, Dunn is exceptional at advancing runners on base ahead of him as well as himself when he's pitched to.

I honestly think, excepting our different takes on how BB should be viewed, that we're actually agreeing with each other in a somewhat unusual fashion. But I think that it hinges on the idea that an Adam Dunn type player actually isn't bad at RBI production, but would better serve a team smart enough to position him in a place that includes the idea that he's better able to avoid hitting poor pitches into play.

[RedsManRick]

Steel, I had a big response laid out and the board ate it, so I'll try to summarize it quickly.

The outcome of the PA is a relatively weak indicator of the quality of the hitting opportunity the hitter was given. There is simply too much noise in the distribution of the composition of possible plate appearances to establish with any reliability that one plate appearance merits being counted as a sufficient to hit and one does not (IBB excluded). It's a very large categorization difference imposed upon events which may often be quite similar in character. Plate appearances with good hitting opportunities can, and do, result in a walk while plate appearances with few opportunities can do results in strikeouts or balls in play. The distinction is clear in the extremes. But most plate appearances live somewhere in the murky middle.

Further, the use of the PA/AB distinction does not provide it's intended value, as evidenced by the fact that we're having this conversation in the first place and the ongoing debates about BA vs OBP. There are certainly times when we'd like to consider only a subset of plate appearances. However, the PA/AB is simply too rough, and perhaps more to the point is so well established that it causes additional confusion as people accept the default calculation, failing to think critically about their chosen denominator. It's a shortcut enabling laziness in a time when we have such easy, quick access to good information that it a shortcut is no longer required.

The "Dunn" problem is that not all walks are created equal. Many of his walks are times when the pitcher avoided him entirely. Others are the result of Dunn failing to capitalize on the opportunities he was given, but still working the count sufficiently to get 4 balls. This non-distinction is at the heart of the debate and the core reason why the conventional types deny your premise. You are right in that Dunn sees many fewer good hitting opportunities than most other hitters. But he also has problems capitalizing on the ones he has. Until and unless we can isolate those things, we're not going to get anywhere.

The question of why Dunn walks so often in RISP situations is completely relevant. But it must be separated from the basic description of the situations he faced. Let's first establish that Dunn came to the plate X number of times in Y circumstance. Let's then agree on the outcomes of those X times. Then, and only then, can we together look at why and how those outcomes came to be and what inferences we should take from that.

[SteelSD]

The "Dunn" problem is that not all walks are created equal.

Actually, no. The issue is that not all PA or "opportunities" are created equal. If a hitter sees only 44% strikes during his PA, can we really hold him to the same standard as someone who sees 50% or 54% strikes? Absolutely not.


Many of his walks are times when the pitcher avoided him entirely. Others are the result of Dunn failing to capitalize on the opportunities he was given, but still working the count sufficiently to get 4 balls. This non-distinction is at the heart of the debate and the core reason why the conventional types deny your premise.

I have absolutely no interest as to why "conventional types" might deny that concept. Seriously. I have no interest because I've analyzed Dunn's PA backwards and forwards over the last few seasons and what you're presenting as a hypothesis simply doesn't exist. He isn't a guy who just sits back and is content to work the count to draw a walk. He's a guy who sees very VERY few strikes he can deal with unless he's willing to swing at pitches he can't do anything with. Those PA aren't real "opportunity". They're the opposing pitcher trying to get him to chase crap; which I would think we'd both agree that he shouldn't.


You are right in that Dunn sees many fewer good hitting opportunities than most other hitters. But he also has problems capitalizing on the ones he has. Until and unless we can isolate those things, we're not going to get anywhere.

And you've isolated that propensity exactly how?


The question of why Dunn walks so often in RISP situations is completely relevant. But it must be separated from the basic description of the situations he faced. Let's first establish that Dunn came to the plate X number of times in Y circumstance. Let's then agree on the outcomes of those X times. Then, and only then, can we together look at why and how those outcomes came to be and what inferences we should take from that.

As long as you're considering BB as actual opportunity without also considering the rate of Strikes (and quality Strikes) he sees in those situations, I don't think we can go any further. You want to isolate RBI production to actual RBI per Runner per AB, then I'm with you. Otherwise, this conversation is dead in the water for pretty much any MLB hitter.

[kaldaniels]

Just a thought as I was overlooking Josh Hamiltons season stats tonight. Assuming BA w/RISP is not a real useful stat...can you say RBI's are an overrated stat as well??? Surely they are more important than BA w/RISP...but are they overrated in general?

Like I mentioned before OPS w/RISP would be an interesting stat.

[Ron Madden]

Just a thought as I was overlooking Josh Hamiltons season stats tonight. Assuming BA w/RISP is not a real useful stat...can you say RBI's are an overrated stat as well??? Surely they are more important than BA w/RISP...but are they overrated in general?

Like I mentioned before OPS w/RISP would be an interesting stat.

RBI are important to Runs Scored but they are team dependent.

[Raisor]

.can you say RBI's are an overrated stat as well??? .

On a player level, absolutly.

[RedsManRick]

Just a thought as I was overlooking Josh Hamiltons season stats tonight. Assuming BA w/RISP is not a real useful stat...can you say RBI's are an overrated stat as well??? Surely they are more important than BA w/RISP...but are they overrated in general?

Like I mentioned before OPS w/RISP would be an interesting stat.

Kal, one thing that you might find useful when examining a stat is to ask the following question:

Am I just trying to assess how well a guy has done something in the past or am I (also) trying to assess how well a guy is likely to do in the future?

A stat like RBI is very useful in telling you how many runners that batter was directly responsible for advancing to the plate. But it actually isn't all that useful in telling you how many runners that batter is likely to drive in in the future. That's based on a combination of opportunity and performance, neither of which are very well measured by RBI itself.

Most of the time, we're using stats to assess a player's actual ability, what he's likely to do in the future, not to just measure his past production.

The idea of something being predictive of itself is very crucial and can help differentiate between meaningful stats and meaningless ones. For example, most situational hitting statistics aren't predictive. Take monthly splits, for example. What a guy has done in every April in his career is a worse predictor of how he's like to hit next April than is his prior full season. It might be interesting to note that a guy struggled, or has struggled, in April, but that doesn't tells us much about future Aprils. Most "RBI" based metrics are the same way. Players who perform well in RBI situations are those who perform well outside of RBI situations. So if you want to identify the players who have done the best in RBI situations past, that's fine. But be careful in extrapolating that forward, suggesting that those players are going to do similarly well in the future. Chances are, if they are great hitters otherwise, they'll keep doing well in RBI situations. And if they aren't, they won't.

Taking Josh Hamilton for a second, he's hit a superb .320/.359/.588 with runners on base. But he's also hit a very good .301/.376/.514 without runners on base. Hamilton's RBI numbers are in part because he's hit so well, but they're also because of opportunity. He's had 220 PA with runners on base. Dunn, by contrast, has had just 142 PA with runners on base. Dunn has had just 65% as many PA with runners on base. Dunn has 62% as many RBI. Obviously that's counting PA, not total runners, so it's not completely accurate. But you can see the massive impact of opportunity -- something not likely to be repeated in the future.

[bucksfan2]

Kal, one thing that you might find useful when examining a stat is to ask the following question:

Am I just trying to assess how well a guy has done something in the past or am I (also) trying to assess how well a guy is likely to do in the future?

A stat like RBI is very useful in telling you how many runners that batter was directly responsible for advancing to the plate. But it actually isn't all that useful in telling you how many runners that batter is likely to drive in in the future. That's based on a combination of opportunity and performance, neither of which are very well measured by RBI itself.

Most of the time, we're using stats to assess a player's actual ability, what he's likely to do in the future, not to just measure his past production.

How many more runners on base has Hamilton had than the other RBI leaders? I agree that RBI is dependent on multiple players but that doesn't mean the batter should have anything taken away from him. I mean if Hamiton had 10 more or 10 less opportunities than say the #2 guy that would be a better evaluation.

[Sea Ray]

It has been shown in many studies that over their careers players will hit the same batting average with and without runners on base once a meaningful sample size has been reached.


Well if that's the case then Adam Dunn must be exceptionally poor with RISP because his career avg with RISP is .024 lower than his overall career BA.

[Chip R]

BA w/RISP is an important offensive stat because a base hit generally means at least one run scores. A walk rarely drives in a run. And a strikeout with RISP almost never scores a run. A base hit w/RISP is by far the most productive result of the at-bat.


"Generally" is a key word here. But there are a few scenarios where the run does not score and the hit is just a hit. There are also examples of where the batter does not get a hit but, just by making contact, a run scores. In the former situation the batter's BA w/RISP increases but the run didn't score. Perhaps the runner on 2nd was slow and they didn't want to send him home and risk an out. Perhaps it was an infield single. Maybe it was a Texas Leaguer where the runner had to hold at 2nd in case the ball was caught. It could be that the runner was thrown out at home.

In the latter example, we all know that sacrifices are a way of scoring someone without getting a hit as is a bases loaded walk or HBP. Of course those do not decrease BA w/RISP but they don't increase it either. Double plays to the middle infield with a runner on 3rd can score a run as can an error. In those cases the batter's BA w/RISP decreases even though the run scored. Does it make what the batter did any less important? Of course not. That's the key to a lot of stats. It's not that the stat is worthless but just looking at it in general without looking into it can be misleading.

[nate]

Well if that's the case then Adam Dunn must be exceptionally poor with RISP because his career avg with RISP is .024 lower than his overall career BA.

But he has a .189 IsoD in those situations and slugs .473. Doesn't that say that pitchers really don't want any part of him with runners on? Doesn't that say they'd rather take their chances with the guy hitting behind him?

Heck, even when they've HAD TO pitch to him (with the bases loaded) he's STILL walked 14 times in 102 plate appearances. BA in those situations: .256.

[RedsManRick]

How many more runners on base has Hamilton had than the other RBI leaders? I agree that RBI is dependent on multiple players but that doesn't mean the batter should have anything taken away from him. I mean if Hamiton had 10 more or 10 less opportunities than say the #2 guy that would be a better evaluation.

Here's the top 10 of OBI, "others batted in" or guys besides the batter himself. Hamilton, not surprisingly, is way ahead of the crowd.



# NAME TEAM YEAR PA PA_ROB R1 R2 R3 R1_BI R2_BI R3_BI ROB OBI R1BI% R2BI% R3BI% OBI%
1. Josh Hamilton TEX 2008 425 220 143 95 65 13 26 35 303 74 9.1% 27.4% 53.8% 0.24422
2. Ryan Howard PHI 2008 417 210 127 111 58 10 29 17 296 56 7.9% 26.1% 29.3% 0.18919
3. Carlos Lee HOU 2008 393 206 131 103 47 13 21 21 281 55 9.9% 20.4% 44.7% 0.19573
4. Justin Morneau MIN 2008 412 229 160 101 45 13 24 17 306 54 8.1% 23.8% 37.8% 0.17647
5. David Wright NYN 2008 432 220 138 94 70 12 11 30 302 53 8.7% 11.7% 42.9% 0.1755
6. Jose Guillen KCA 2008 380 181 125 88 49 13 22 17 262 52 10.4% 25.0% 34.7% 0.19847
7. Mark Teixeira ATL 2008 407 225 170 110 44 15 17 20 324 52 8.8% 15.5% 45.5% 0.16049
8. Carlos Beltran NYN 2008 411 209 135 100 51 12 21 18 286 51 8.9% 21.0% 35.3% 0.17832
9. Lance Berkman HOU 2008 397 200 122 93 39 16 17 18 254 51 13.1% 18.3% 46.2% 0.20079
10. Bengie Molina SFN 2008 340 186 113 89 44 10 16 24 246 50 8.9% 18.0% 54.5% 0.20325

However, when we control for opportunity, and look at the % other runners batted in, the picture changes a bit.

# NAME TEAM YEAR PA PA_ROB R1 R2 R3 R1_BI R2_BI R3_BI ROB OBI R1BI% R2BI% R3BI% OBI%
1. David Dejesus KCA 2008 343 116 71 55 35 5 16 20 161 41 7.0% 29.1% 57.1% 0.25466
2. Josh Hamilton TEX 2008 425 220 143 95 65 13 26 35 303 74 9.1% 27.4% 53.8% 0.24422
3. Alexi Casilla MIN 2008 228 96 59 47 27 5 12 15 133 32 8.5% 25.5% 55.6% 0.2406
4. Nate Mclouth PIT 2008 417 153 100 70 35 14 16 16 205 46 14.0% 22.9% 45.7% 0.22439
5. Carlos Quentin CHA 2008 395 165 105 79 47 8 17 23 231 48 7.6% 21.5% 48.9% 0.20779
6. Cody Ross FLO 2008 269 108 82 46 32 10 10 13 160 33 12.2% 21.7% 40.6% 0.20625
7. Jerry Hairston CIN 2008 213 70 42 37 18 2 7 11 97 20 4.8% 18.9% 61.1% 0.20619
8. Bengie Molina SFN 2008 340 186 113 89 44 10 16 24 246 50 8.9% 18.0% 54.5% 0.20325
9. Casey Blake CLE 2008 341 153 95 81 38 5 21 17 214 43 5.3% 25.9% 44.7% 0.20093
10. Lance Berkman HOU 2008 397 200 122 93 39 16 17 18 254 51 13.1% 18.3% 46.2% 0.20079

After Hairston, the remaining qualified Reds are:
91. Dunn (.154)
93. Phillips (.152)
116. Keppinger (.147)
136. Griffey (.142)
149. Votto (.138)
178. Bako (.127)
239. Encarnacion (.097)
244. Patterson (.092)

Bruce is at .135. Suffice it to say, the Reds have not been very good as a team at driving in runners on base. And we know that, as a team, the Reds have not been particularly good at getting on base. Voila, a below average offense. For all his walks, strikeouts, and low batting average, Dunn still has driven in a greater percentage of runners on base than any other Red besides Hairston.

The one caveat I would make is that this does not control for the difference in driving in somebody from 3B or 1B -- a weighted system would be a more precise measure. But in general, it does show pretty well who the best "RBI guys" have been for the Reds this year.

[Raisor]

How many more runners on base has Hamilton had than the other RBI leaders? I agree that RBI is dependent on multiple players but that doesn't mean the batter should have anything taken away from him. I mean if Hamiton had 10 more or 10 less opportunities than say the #2 guy that would be a better evaluation.

Since the Hamilton/Dunn comparisions are always made:

As of 7/13/08, Hamilton has has had 89 more runners on base for his PA's then Dunn this season:
303 vs 214.

[Highlifeman21]

Since the Hamilton/Dunn comparisions are always made:

As of 7/13/08, Hamilton has has had 89 more runners on base for his PA's then Dunn this season:
303 vs 214.

Hold on. Hoooooooooooooooold on.

You're telling me that Dunn has less RBI opportunities b/c his teammates suck at getting on base in front of him?

You're also telling me that Hamilton has more RBI than Dunn b/c he's had more guys on when he's up?

That kinda crazy talk is pure nonsense.

[RichRed]

But he has a .189 IsoD in those situations and slugs .473. Doesn't that say that pitchers really don't want any part of him with runners on? Doesn't that say they'd rather take their chances with the guy hitting behind him?

Heck, even when they've HAD TO pitch to him (with the bases loaded) he's STILL walked 14 times in 102 plate appearances. BA in those situations: .256.

Not only that, but Dunn's overall career BABIP is .287. With RISP, his BABIP drops to .255. Maybe there's some explanation for that but it just shows that there's more to it than simply saying "Dunn can't hit with runners in scoring position."

[RedsManRick]

Not only that, but Dunn's overall career BABIP is .287. With RISP, his BABIP drops to .255. Maybe there's some explanation for that but it just shows that there's more to it than simply saying "Dunn can't hit with runners in scoring position."

Maybe the low BABIP with RISP is an indicator that when Dunn tries to expand his zone and put the ball in play because the situation dictates it, he hits the ball with less authority. An effect of "situational hitting" from a guy who generally has problems making contact.

[bucksfan2]

Maybe the low BABIP with RISP is an indicator that when Dunn tries to expand his zone and put the ball in play because the situation dictates it, he hits the ball with less authority. An effect of "situational hitting" from a guy who generally has problems making contact.

Nice knock on "situation hitting"

I was wondering if it could have anything to do with pitching changes. If Dunn is up in the 6th, 7th, 8th, 9th inning with RISP or ROB then the manager may be mroe apt to bringing in another pitcher to face Dunn.

Another possibility is a pitcher concentrates more when Dunn is up with ROB. If the pitcher hits his spots then Dunn is an easy out. If he doesn't Dunn can plant a ball in the stands in the blink of an eye.

[RichRed]

Maybe the low BABIP with RISP is an indicator that when Dunn tries to expand his zone and put the ball in play because the situation dictates it, he hits the ball with less authority. An effect of "situational hitting" from a guy who generally has problems making contact.

I wondered about that too. One thing that might offer a clue is how his LD% with RISP compares to what it is in other situations. I'm not sure where to find those numbers.

[Sea Ray]

After Hairston, the remaining qualified Reds are:
91. Dunn (.154)
93. Phillips (.152)
116. Keppinger (.147)
136. Griffey (.142)
149. Votto (.138)
178. Bako (.127)
239. Encarnacion (.097)
244. Patterson (.092)

Bruce is at .135. Suffice it to say, the Reds have not been very good as a team at driving in runners on base. And we know that, as a team, the Reds have not been particularly good at getting on base. Voila, a below average offense. For all his walks, strikeouts, and low batting average, Dunn still has driven in a greater percentage of runners on base than any other Red besides Hairston.

The one caveat I would make is that this does not control for the difference in driving in somebody from 3B or 1B -- a weighted system would be a more precise measure. But in general, it does show pretty well who the best "RBI guys" have been for the Reds this year.

All that really tells us is none of the Reds are really very good at clutch hitting, but that doesn't help Adam Dunn that much. Seems to me he's part of an overall team problem with RISP.

[RedsManRick]

All that really tells us is none of the Reds are really very good at clutch hitting, but that doesn't help Adam Dunn that much. Seems to me he's part of an overall team problem with RISP.

Dunn isn't part of the problem. He's above league average at driving in runs when given the opportunity to do so. I suppose if you expect him to be a league leader in that, then you would consider him a disappointment. But anybody who expects Dunn to be a superb RBI guy clearly doesn't get the nature of his production, which is undeniably substantial on balance.

To nip the retort in the bud, any suggestion that he gets paid a ton of money to drive in runs simply isn't accurate. He gets paid to produce runs and he does that quite well -- commensurate to his salary.

[Sea Ray]

But anybody who expects Dunn to be a superb RBI guy clearly doesn't get the nature of his production, which is undeniably substantial on balance.



Exactly right. Dunn is not an RBI guy although he does have other very good offensive skills. A lot of folks do expect him to be a superb RBI guy; Marty Brennaman is one of them. And afterall, RBIs are what this thread is all about.

[bucksfan2]

Dunn isn't part of the problem. He's above league average at driving in runs when given the opportunity to do so. I suppose if you expect him to be a league leader in that, then you would consider him a disappointment. But anybody who expects Dunn to be a superb RBI guy clearly doesn't get the nature of his production, which is undeniably substantial on balance.

To nip the retort in the bud, any suggestion that he gets paid a ton of money to drive in runs simply isn't accurate. He gets paid to produce runs and he does that quite well -- commensurate to his salary.

Yea but there are 90 other players in baseball better at driving Dunn in. I wouldn't exactly call that good.

[BuckeyeRedleg]

Exactly right. Dunn is not an RBI guy although he does have other very good offensive skills. A lot of folks do expect him to be a superb RBI guy; Marty Brennaman is one of them. And afterall, RBIs are what this thread is all about.


It's hard to knock in runs when nobody is on base in front of you.

Dunn could solve this problem if he'd just hit 70-80 HR's every year. Darn him.

[RedsManRick]

Yea but there are 90 other players in baseball better at driving Dunn in. I wouldn't exactly call that good.

I wouldn't call it good either. I called it above average, which it is. The point though is that RBI are merely one way in which a player can be productive, and arguably the worst part of Dunn's offensive game.

[Sea Ray]

Here's the top 10 of OBI, "others batted in" or guys besides the batter himself. Hamilton, not surprisingly, is way ahead of the crowd.



Where did you find OBI stats listed? I must say that's a new stat on me. I've been reading about baseball my whole life and I've never even seen it mentioned on this site before. I doubt I'll ever see it on a TV graphic or listed in the Sunday paper.

What is the genesis of this stat?

[RedsManRick]

Where did you find OBI stats listed? I must say that's a new stat on me. I've been reading about baseball my whole life and I've never even seen it mentioned on this site before. I doubt I'll ever see it on a TV graphic or listed in the Sunday paper.

What is the genesis of this stat?

It's from Baseball Prospectus' standard RBI Opportunities report. In a manner, you do see it in every single box score. It's just RBI-HR -- the count of guys a batter drove in besides himself.

OBI% is likewise a very simple concept. Basically, it's the inverse of the LOB. Of all the guys on base when a player came to bat, what percentage of them did he drive in? I'm not sure why they choose to display a decimal instead of the more intuitive percentage, but it's the same either way.

In fact, I would argue that OBI% is the best stat I've seen when it comes to answering the question asked in this thread: how do we measure how well a batter converts runners in to runs?

Josh Hamilton has driven in 24.4% of the runners on base when he's come to bat. That's pretty impressive. Adam Dunn has driven in just 15.4% of the runners on base for him. Brandon Phillips just 15.2% (and given that he gets on base so much less than Dunn, why does he bat ahead of him again?).

Here is the report run for all Reds with at least 50 PA. I've converted OBI% in to a percent. This highlights one of my big complaints about using RISP; it's failure to account for men on first base. Look at how well Dunn does at driving in those runners compared to Phillips, EE, Keppinger, and Votto. Why would we choose to exclude those base runners? Other than to buttress an argument about batting average, I can't see why.



# NAME TEAM YEAR PA PA_ROB R1 R2 R3 R1_BI R2_BI R3_BI ROB OBI R1BI% R2BI% R3BI% OBI%
1 Jerry Hairston CIN 2008 213 70 42 37 18 2 7 11 97 20 4.8% 18.9% 61.1% 20.6%
2 Adam Dunn CIN 2008 368 142 91 77 46 10 9 14 214 33 11.0% 11.7% 30.4% 15.4%
3 Brandon Phillip CIN 2008 403 200 137 85 60 5 14 24 282 43 3.7% 16.5% 40.0% 15.2%
4 Jeff Keppinger CIN 2008 238 96 61 41 27 1 8 10 129 19 1.6% 19.5% 37.0% 14.7%
5 Ken Griffey Jr. CIN 2008 374 170 106 76 43 10 10 12 225 32 9.4% 13.2% 27.9% 14.2%
6 Ryan Freel CIN 2008 143 53 36 22 13 0 4 6 71 10 0.0% 18.2% 46.2% 14.1%
7 Joey Votto CIN 2008 343 142 102 64 29 6 10 11 195 27 5.9% 15.6% 37.9% 13.8%
8 Scott Hatteberg CIN 2008 61 33 27 15 9 2 3 2 51 7 7.4% 20.0% 22.2% 13.7%
9 Jay Bruce CIN 2008 182 78 55 42 14 1 7 7 111 15 1.8% 16.7% 50.0% 13.5%
10 Paul Bako CIN 2008 228 106 81 56 20 3 9 8 157 20 3.7% 16.1% 40.0% 12.7%
11 Paul Janish CIN 2008 72 29 25 14 6 0 3 2 45 5 0.0% 21.4% 33.3% 11.1%
12 Javier Valentin CIN 2008 76 43 31 23 11 0 3 4 65 7 0.0% 13.0% 36.4% 10.8%
13 David Ross CIN 2008 131 65 45 31 14 1 2 6 90 9 2.2% 6.5% 42.9% 10.0%
14 Edwin Encarnaci CIN 2008 337 152 110 59 37 3 8 9 206 20 2.7% 13.6% 24.3% 9.7%
15 Corey Patterson CIN 2008 212 81 53 39 17 1 4 5 109 10 1.9% 10.3% 29.4% 9.2%
16 Norris Hopper CIN 2008 58 21 17 6 8 0 0 1 31 1 0.0% 0.0% 12.5% 3.2%

As I noted above, I would like to see a weighted version that treats runners differently based on their base. Just as with RISP, it's intuitively not "fair" to treat a guy on 3B the same as a guy on 2B or 1B. But it's still worlds better as an evaluation of both performance and ability.

A side note, I think you backed in to a very important point. Stats have no purpose outside of their usefulness in helping us understand who did what, when, why, and how. That it's widely published, used, or accept is completely besides the point. If it helps us understand things, it should be used. And if it doesn't, it shouldn't.

[red-in-la]

It has something to do with some guy named Dunn...and then the rest becomes a real love fest between the people who think it's rabbit season and those who think it's duck season.

OMG..........:clap:

[red-in-la]

When I was 25 years old, and started teach engineering using computers for the first time, I wondered why some of the older people in my classes just seemed to refuse to get what we were talking about.

You guys are making me see for the first time, what those older guys were going through. :shocked:

I know the stats shown here are real and informative to some, but I just don't get it. I grew up enjoying baseball without analysis......and I just cannot get beyond that. I KNOW when a guy is a clutch hitter, when he is the guy I hope comes up with the bases loaded.

I have learned on the board, that the term "know" as I use it, is an emotional not necessary factual point-of-view. And I have come to appreciate that.

But we used to have a saying, "paralysis by analysis" and to me, these discussions become almost pointless when there is no apparent opinion except a spreadsheet.

BTW RedsManRick, I really like the table you showed. I amost understand it. I do understand your explanation of it and OBI to me makes more sense than OPS or VORP or BABIP et al.

[Sea Ray]

I enjoy Rick's posts as well. The effort he puts into that stuff is very impressive.

I am convinced that anybody can prove a point if they keep looking for more and more obscure stats. If BA doesn't do it for you go to OPS. If not OPS then BAbip or tOPS or iso...whatever suffix you like.

There are those that espouse OPS as a very indicative stat. Well Josh Hamilton's OPS this year is less than it was last year and it is almost identical to Adam Dunn's this year. Does anybody argue that Hamilton is having a worse 2008 than he had in 2007? Or who would argue that he's having an offensive year equal to Dunn's? Stats can show you anything you want.

[RedsManRick]

BTW RedsManRick, I really like the table you showed. I amost understand it. I do understand your explanation of it and OBI to me makes more sense than OPS or VORP or BABIP et al.

I'm glad it could help. Frankly, if I understood everything I've read about sabermetrics, it wouldn't be nearly as fun. I've learned a TON over the 10 years or so I've been delving in to this stuff. It's the constant learning that keeps it interesting. I'm always being shown or finding better ways to do things, to understand what I'm seeing. The trick, for me, is approaching it with the understanding that there are always more precise ways to measure something.

Stats are just numbers and there's no reason to be emotionally attached to any of them. The ideas behind the stats are what gets me excited and the better the job the stat does at uncovering it, the more fun it is.

The eyes are a wonderful data collection tool and the brain is an incredible calculator. But they're fallible. For me, analyzing the game (it's 100% acceptable to want to watch/enjoy the game and not analyze it!) is about collecting as much information as possible and finding the most accurate ways to interpret it.

It's an extremely iterative process and many people choose to quit once they reach a certain point. Some people stop at their own eyes. Some stop at what the announcers are saying. Others real analysis online and still others dig in to the data themselves. Obviously everybody can stop where they choose to. Many choose to stop once they reach a point where they start getting paid for what they already understand. My request is that regardless of where a person chooses to draw the line, he/she appreciates that those continuing to dig deeper likely are developing a fuller understanding of certain aspects of game.

One could draw a comparison with scouting. Imagine the scout in the 1960s who didn't want to use a radar gun because he felt he could see the velocity just fine with his own eyes. Or the scouts today who don't want to use pitch f/x.

That's not to suggest anybody understands the game 100%. There are areas of the game to which statistical analysis cannot begin to speak. But when it comes to how to measuring past performance and predicting future performance for established players, statistical analysis takes the cake. I'm sure the stats community will continue to develop new, better ways to understand player performance and I'm looking forward to learning about them when the time comes.

[MWM]

I am convinced that anybody can prove a point if they keep looking for more and more obscure stats.

then you don't have even a cursory understanding of statistical evaluation, because that is a patently false statement.

[RedsManRick]

I enjoy Rick's posts as well. The effort he puts into that stuff is very impressive.

I am convinced that anybody can prove a point if they keep looking for more and more obscure stats. If BA doesn't do it for you go to OPS. If not OPS then BAbip or tOPS or iso...whatever suffix you like.

There are those that espouse OPS as a very indicative stat. Well Josh Hamilton's OPS this year is less than it was last year and it is almost identical to Adam Dunn's this year. Does anybody argue that Hamilton is having a worse 2008 than he had in 2007? Or who would argue that he's having an offensive year equal to Dunn's? Stats can show you anything you want.

Sea Ray, don't confuse the stats with the interpretation of the stats. Stats don't lie; Statisticians do.

Just because somebody has used stats to make an argument doesn't mean they are using them properly. And just because their answer is the opposite of what you think is obvious doesn't mean they're wrong. Perhaps you aren't asking the question you think you are.

You bring up a great topic. If OPS is such a great measurement, and Dunn has a greater OPS then Hamilton, can we conclude Dunn has been more productive?

Don't let other people answer your question without pushing them on it. But first, be clear on exactly what you are trying to measure. What's the idea behind the stat? Define what you mean by "productive". If you've defined productive as "driven in the most base runners", clearly Hamilton has outproduced Dunn. But shouldn't we count the batter's ability to get on base for the guys behind him too? What about the fact that they don't have the same number of guys available to drive in nor the same guys behind them advancing them?

The trick is being very explicit in what you're trying to analyze. It's very easy to get lost in the history of what we think we know and what preconceived notions we have about what counts, what doesn't, and how certain stats should be used. Wipe the slate clean. Ask your question in plain terms and then look for the ways to measure the relevant things.

And if you don't feel like going through all of that and would prefer to simply stick to what you know, then you don't really have room to call the other people liars for disagreeing with you.

[Sea Ray]

Just because somebody has used stats to make an argument doesn't mean they are using them properly. And just because their answer is the opposite of what you think is obvious doesn't mean they're wrong. Perhaps you aren't asking the question you think you are.



The trick is being very explicit in what you're trying to analyze. It's very easy to get lost in the history of what we think we know and what preconceived notions we have about what counts, what doesn't, and how certain stats should be used. Wipe the slate clean. Ask your question in plain terms and then look for the ways to measure the relevant things.

And if you don't feel like going through all of that and would prefer to simply stick to what you know, then you don't really have room to call the other people liars for disagreeing with you.


You bring up a good point. A lot of times the disagreements around here are about what's being analyzed and not the players themselves. That's when the discussions get tedious.

As for your last statement I have no idea to what you're referring. I for one always respect others' opinions as being exactly that.

[Ron Madden]

You bring up a good point. A lot of times the disagreements around here are about what's being analyzed and not the players themselves. That's when the discussions get tedious.

STATISTICS are the best way we have to analyze the players themselves.

They hold all the information about a players past production and can give us a pretty good idea of a players future production.

The key is to expand our thinking capacity beyond (the three stats shown at the bottom of the TV screen) BA, HR, RBI to judge hitters, or W's, L's and ERA to judge pitchers.

[Sea Ray]

STATISTICS are the best way we have to analyze the players themselves.

They hold all the information about a players past production and can give us a pretty good idea of a players future production.

The key is to expand our thinking capacity beyond (the three stats shown at the bottom of the TV screen) BA, HR, RBI to judge hitters, or W's, L's and ERA to judge pitchers.

Stats are a vital part of player evaluations but it can get tedious 'cause sometimes there can be no end to it.

Let's say we're having an argument between who's the better offensive player, A or B.

One fan says player A 'cause he has a higher BA and RBI totals. Another fan says no it's player B 'cause his OPS is higher. Another says no it's RC that determines the better player. Another fan counters and says no it's win shares. Another fan chimes in and says no it's RC park adjusted. Another says well the reason that park has so many HRs is because the pitchers for the home team are so horrible skewing the park adjustment factor...There's no end to it...I take that back. Here on RZ there is an end. The end is when the frustrated party says to the other "you just aren't smart enough to understand my stats." Talk about tedious.

I'll talk stats all day long with anyone but when it gets to the point that someone quits talking baseball and starts talking about who's smarter, I move on.

[bucksfan2]

then you don't have even a cursory understanding of statistical evaluation, because that is a patently false statement.

IMO there are so many diferent evaluations that one can pick and chose which stat to use at a particular time. I have a general question about a basic statistical analysis.

Player A has an obp of .400 and has gotten on base 2-2. He is now up for the 3rd time. What are his odds of getting on base the thrid time.

Is it
1. .400*.400*.400 = .064

or is it

2. .400

Example 1 says that the likelihood of player A getting on base three consecutive times is his obp times number of occurances. Example 2 takes each atbat as an independent event which is simply player A's obp.

[princeton]

IMO there are so many diferent evaluations that one can pick and chose which stat to use at a particular time. I have a general question about a basic statistical analysis.

Player A has an obp of .400 and has gotten on base 2-2. He is now up for the 3rd time. What are his odds of getting on base the thrid time.

Is it
1. .400*.400*.400 = .064

or is it

2. .400

Example 1 says that the likelihood of player A getting on base three consecutive times is his obp times number of occurances. Example 2 takes each atbat as an independent event which is simply player A's obp.

2.

number 1 is the likelihood that the player will be on base each of his next three times up.

[RedsManRick]

IMO there are so many diferent evaluations that one can pick and chose which stat to use at a particular time. I have a general question about a basic statistical analysis.

Player A has an obp of .400 and has gotten on base 2-2. He is now up for the 3rd time. What are his odds of getting on base the thrid time.

Is it
1. .400*.400*.400 = .064

or is it

2. .400

Example 1 says that the likelihood of player A getting on base three consecutive times is his obp times number of occurances. Example 2 takes each atbat as an independent event which is simply player A's obp.

If each at bat is independent, #1 is the chance a player gets on base 3 times in a row. #2 is the chance that he gets on base in any given AB, regardless of what has happened previously.

[kaldaniels]

If each at bat is independent, #1 is the chance a player gets on base 3 times in a row. #2 is the chance that he gets on base in any given AB, regardless of what has happened previously.

Spot on. :thumbup:

[Ron Madden]

Stats are a vital part of player evaluations but it can get tedious 'cause sometimes there can be no end to it.

Let's say we're having an argument between who's the better offensive player, A or B.

One fan says player A 'cause he has a higher BA and RBI totals. Another fan says no it's player B 'cause his OPS is higher. Another says no it's RC that determines the better player. Another fan counters and says no it's win shares. Another fan chimes in and says no it's RC park adjusted. Another says well the reason that park has so many HRs is because the pitchers for the home team are so horrible skewing the park adjustment factor...There's no end to it...I take that back. Here on RZ there is an end. The end is when the frustrated party says to the other "you just aren't smart enough to understand my stats." Talk about tedious.

I'll talk stats all day long with anyone but when it gets to the point that someone quits talking baseball and starts talking about who's smarter, I move on.

This is a discussion about baseball statistics. Let's say all of us are just sharing our opinions. ;) I haven't noticed anyone talking down to others.

:)

[Sea Ray]

I haven't noticed anyone talking down to others.

:)

You haven't? I've documented just a few examples in another thread. Checkout post #119 if you'd like to see a few examples

http://www.redszone.com/forums/showthread.php?t=70256&page=8

[jojo]

You haven't? I've documented just a few examples in another thread. Checkout post #119 if you'd like to see a few examples

http://www.redszone.com/forums/showthread.php?t=70256&page=8

Was that the post where you claimed you wouldn't go some where you were actually just leaving? :cool:

Just saying.....

[Chip R]

Let's keep it on topic, people.