PDA

View Full Version : MERGED: Streakiness / I Wish There Was a "Consistency" Stat...



nate
06-11-2008, 10:26 AM
For the stat-minded "dudes" here, I was wondering if there is a way to measure "streakiness" in a player.

Let's say you have two hitters that go 15 for 40 at the plate. But hitter A gets all of his 15 hits over the course of a 4-5 games while hitter B gets roughly one hit per game.

Is there a way to measure this? Is it worth measuring? Does it tell you something important?

RedsManRick
06-11-2008, 10:57 AM
I highly recommend The Book: Playing the Percentages in Baseball (http://www.insidethebook.com/) by Tom Tango et. al.

Chapter 2 discusses the nature of streakiness. The extremely abridged version is that streaks tend to be random fluctuations more than a real change in actual player performance. A guy who is on a cold or hot streak may see his expected OPS moving forward go up or down by 20 points or so -- statistically significant, but not such that it justifies benching a guy or moving him in the lineup.

Obviously, due to the nature of both randomness and reality, there will be some guys who continue to streak and there will be some guys for whom the streak represents the beginning of a new skill level. But your best bet is that guys are who are hot/cold will start performing commensurate to their expected level of performance prior to the streak starting with their next at bat.

flyer85
06-11-2008, 11:05 AM
Phillips OPS against lefties is an unreal 1.206 and is countered by an abysmal 670 against RHP. That certainly could contribute to "streakiness".

HokieRed
06-11-2008, 01:41 PM
A good place to begin with Phillips would be getting him out of the 4 hole, where he has no business. 7th is as high in the order as he would hit on a good team. He is a classic example of a bad organization's tendency to overvalue its players by failing to evaluate them vis-a-vis the talent on other teams. He's a good second baseman with a little better than average power who is also terribly streaky and has a huge platoon differential. He's not what you want in the middle of the lineup on a contending team.

BCubb2003
06-11-2008, 02:00 PM
Streakiness should be easier to measure than clutch. You can look at whether a player's performance comes in bunches or over a wider distribution. And it might begin to give you some tangible insight into the psychology of the game.

nate
06-12-2008, 07:00 PM
I highly recommend The Book: Playing the Percentages in Baseball (http://www.insidethebook.com/) by Tom Tango et. al.

Chapter 2 discusses the nature of streakiness. The extremely abridged version is that streaks tend to be random fluctuations more than a real change in actual player performance. A guy who is on a cold or hot streak may see his expected OPS moving forward go up or down by 20 points or so -- statistically significant, but not such that it justifies benching a guy or moving him in the lineup.

Obviously, due to the nature of both randomness and reality, there will be some guys who continue to streak and there will be some guys for whom the streak represents the beginning of a new skill level. But your best bet is that guys are who are hot/cold will start performing commensurate to their expected level of performance prior to the streak starting with their next at bat.

Thanks, RMR...I'll check that out!

*BaseClogger*
06-12-2008, 09:39 PM
Is Streakiness important? I don't think it matters at all...

bucksfan2
06-13-2008, 10:34 AM
For the stat guys out there......

One of the reasons that I struggle with statistical analysis is that it seems to take an aggregate sum with out looking at each individual data point. I am wondering if anyone has measured something alongs the lines of standard deviation of certain numbers?

For example lets assume that someone has an obp of .350. Wouldn't it be better to look at the frequency of deviation from .350 that a certain hitter was? You can look at a final number and say wow he had a pretty good season but if you look closer at the data you may find that without a torrid month the player was pretty poor in general.

I was thinking about this looking at Encarnacion's game log and without a good 10 game span he had been pretty horriable. But with that 10 game span he worked his average up from the interstate to right around the .300 level. I would wonder if you could look at 5 or 10 game splits and look at those numbers and where they are in comparison to the final numbers.

I think that streakiness is an important measure in a players overall year. I would wonder if you want a guy who is more "streaky" or more "consistant"? Woud you want a guy who goes 10-20 but then 2-20 for an average .300 or would you want a guy to go 6-20 and then 6-20 for the same average?

nate
06-13-2008, 10:49 AM
I think that streakiness is an important measure in a players overall year. I would wonder if you want a guy who is more "streaky" or more "consistant"? Woud you want a guy who goes 10-20 but then 2-20 for an average .300 or would you want a guy to go 6-20 and then 6-20 for the same average?

Yeah, that's kind of why I'm asking. Is one guy preferable to the other? How can you measure it? Is it important to measure?

I guess maybe I'm asking something like, after each game, how often does stat "X" go _up_ versus _down_?

GAC
06-13-2008, 10:54 AM
When does a streak stop being labeled as a streak? ;)

You obviously want consistency out of a player; but the fact is, all players experience streaks of good and bad. It is such a random fluctuation that I don't know if it can be accurately charted.

And I understand people like to look at an individual player's splits, such as vs lefties/righties; but can't that lead being over-analytical in some sense when it comes to constructing a consistent (set) batting order? Or maybe some don't advocate that. What I mean is, you're then expecting a manager to be moving practically all of your players daily all around in this batting order based on who is pitching, what their splits are vs them, etc.

Even your better players have their "weaknesses" in their game. So do you then move guys like Dunn, Bruce, and/or whoever all around based on their splits vs that particular opposition?

Isn't that a little too much to ask?

bucksfan2
06-13-2008, 11:20 AM
When does a streak stop being labeled as a streak? ;)

You obviously want consistency out of a player; but the fact is, all players experience streaks of good and bad. It is such a random fluctuation that I don't know if it can be accurately charted.

And I understand people like to look at an individual player's splits, such as vs lefties/righties; but can't that lead being over-analytical in some sense when it comes to constructing a consistent (set) batting order? Or maybe some don't advocate that. What I mean is, you're then expecting a manager to be moving practically all of your players daily all around in this batting order based on who is pitching, what their splits are vs them, etc.

Even your better players have their "weaknesses" in their game. So do you then move guys like Dunn, Bruce, and/or whoever all around based on their splits vs that particular opposition?

Isn't that a little too much to ask?

Ok but my question with streaks are when does a streak player hurt you. When do need a lesser yet more consistant player. Lets assume you have two players players A and B. Player A hits at a rate of .300 and usually sees fluctuations that stay within a .050 range. Player B has the same rate of .300 but sees fluctuations of .100-.150. Player A would give you much more consistant production but Player B has the ability for great weeks but awful weeks.

Streaks are a normal part of baseball. You will see fluctuations because of the nature of the game. I think at some point it becomes important to dive into the given games and try to find a common occurance. I wonder if streaks are really just a reflection of a given players ability.

Kc61
06-13-2008, 11:22 AM
A good place to begin with Phillips would be getting him out of the 4 hole, where he has no business. 7th is as high in the order as he would hit on a good team. He is a classic example of a bad organization's tendency to overvalue its players by failing to evaluate them vis-a-vis the talent on other teams. He's a good second baseman with a little better than average power who is also terribly streaky and has a huge platoon differential. He's not what you want in the middle of the lineup on a contending team.

I think this is good point, Phillips is not a cleanup hitter, but I think the point is a bit exaggerated here.

Phillips' last two seasons against righties were .737 and .743. He is usually better againt righties than this year and it is likely that he needs an adjustment or is pressing righty or is in a slump righty. These numbers are not great but there are other good players who do much better against either lefties or against righties.

There is no doubt that the Reds could use a top notch righty hitter to bat cleanup, but Phillips isn't a seventh place hitter.

I don't think the Reds overvalue Phillips bat at all. I think they know what they have. But they want a righty bat in the middle of the lineup and don't see better options right now. I don't either unless they want EE to hit cleanup. He's going ok right now so maybe it makes sense against righties.

nate
06-13-2008, 11:24 AM
When does a streak stop being labeled as a streak? ;)

You obviously want consistency out of a player; but the fact is, all players experience streaks of good and bad. It is such a random fluctuation that I don't know if it can be accurately charted.

And I understand people like to look at an individual player's splits, such as vs lefties/righties; but can't that lead being over-analytical in some sense when it comes to constructing a consistent (set) batting order? Or maybe some don't advocate that. What I mean is, you're then expecting a manager to be moving practically all of your players daily all around in this batting order based on who is pitching, what their splits are vs them, etc.

Even your better players have their "weaknesses" in their game. So do you then move guys like Dunn, Bruce, and/or whoever all around based on their splits vs that particular opposition?

Isn't that a little too much to ask?

I wouldn't use "streakiness" as a lineup construction tool...at all.

I'm just wondering, in historical context, which player was good _and_ minimized their slumps. IOW, which players had the least variation in stat "X" while keeping that stat "good."

GAC
06-13-2008, 11:45 AM
Ok but my question with streaks are when does a streak player hurt you. When do need a lesser yet more consistant player. Lets assume you have two players players A and B. Player A hits at a rate of .300 and usually sees fluctuations that stay within a .050 range. Player B has the same rate of .300 but sees fluctuations of .100-.150. Player A would give you much more consistant production but Player B has the ability for great weeks but awful weeks.

Streaks are a normal part of baseball. You will see fluctuations because of the nature of the game. I think at some point it becomes important to dive into the given games and try to find a common occurance. I wonder if streaks are really just a reflection of a given players ability.

based on the illustration you are showing me.... I'd go with Player A because I want more consistent play out of my players. If looking at a graph.... I'd want a player who shows much better consistency that stays along the plane, with mild fluctuations up/down. Rather then a player whose fluctuations look like an earthquake on a seismograph.

GAC
06-13-2008, 12:11 PM
I wouldn't use "streakiness" as a lineup construction tool...at all.

I'm just wondering, in historical context, which player was good _and_ minimized their slumps. IOW, which players had the least variation in stat "X" while keeping that stat "good."


It's definitely a sound and viable question you've asked nate.

Far better then the Corey Patterson and Dusty Baker is an Idiot threads. ;)

It's good to spark debate.

I think there is too much random fluctuation involved that it prevents those who study it from setting firm parameters as to what defines streakiness. It may be a lot like trying to define what clutch hitting is/is not.

It's obvious that a player can get hot and carry a team for a short spell, just like a slumping player can hurt the team. Managers play that "hot hand" all the time, and sit players that are slumping. But it's hard to predict when that hot spell is ending or going to end, or when a slump is going to be over.

Many would say Adam Dunn's April, in certain areas, contributed to our poor start offensively. Adam picked it up in May, and then EE had a terrible May. Now we got Phillips, so far in June, in slump.

I guess the bigger question can be asked.... are certain types of ballplayers more prone to streakiness then others, and can it be accurately tracked?

I've always believed, and it's based on a larger window not a small one, that once a player has established himself, as far as performance goes, and you basically know what you're going to get from them and what type of player they are.... then when they slump you let them work (play) their way out of it. Unless there is some other contributing factor (injury, etc) benching a guy that has shown consistent performance in the past, IMHO, only complicates matters and messes with their heads more.

And another factor to consider too is what is that player's work habits? I'm not trying to point fingers at anyone, or insinuate anything; but if a player, when they slump, doesn't show that extra effort in-between games... and I'm talking about possibly doing even more then your daily routine... that may make it harder.

SteelSD
06-13-2008, 12:14 PM
Ok but my question with streaks are when does a streak player hurt you. When do need a lesser yet more consistant player. Lets assume you have two players players A and B. Player A hits at a rate of .300 and usually sees fluctuations that stay within a .050 range. Player B has the same rate of .300 but sees fluctuations of .100-.150. Player A would give you much more consistant production but Player B has the ability for great weeks but awful weeks.

Fluctuations to those extremes don't happen over the short term when we reach a large enough data sample. As a data set grows, it becomes more difficult for single or even multiple-game performance to affect things like OBP/SLG/OPS past the point of tenths of percentage points.

To view any potentially reliable trend information, one would first have to reach a large enough data sample in order to confirm actual performance level, use that as a starting point, and then continue to track performance from there while not re-setting the percentage data with each new season. But that brings us back to point one again- as the data set continues to grow, our graph line will fluctuate less and less and even 10-game data sets will show as nothing more than raindrops in the ocean of a player's career performance.

And here's a good example of how wild early-season fluctuations can affect game-by-game standard deviations. Two players and their game-by-game standard deviations (0 PA games subtracted) for 2007:

Player A:

BA: 0.035
OBP: 0.031
SLG: 0.071
OPS: 0.096

Player B:

BA: 0.025
OBP: 0.019
SLG: 0.125
OPS: 0.137

Which one is Edwin Encarnacion and which is A-Rod?

Now let's remove the first 249 AB from each player in order to avoid some of the interference caused by seasonal AB/PA "reset" (which causes rate stats to fluctuate wildly at the beginning of a season):

Player A:

BA: 0.012
OBP: 0.007
SLG: 0.027
OPS: 0.034

Player B:

BA: 0.009
OBP: 0.007
SLG: 0.021
OPS: 0.027

That's a much different snapshot for both players and we no longer see the "false positive" 100+ point swings caused by seasonal reset.

RedsManRick
06-13-2008, 01:32 PM
Touching on what Steel has said, one of the issues with observing streakiness is that in large measure, it involves a selection bias whereby we "start" the streak at the moment the thing we're measuring occurs. Thus, the player is bound to appear to be struggled because you have purposefully biased your sample through the inclusion of a set starting point. How you define the intervals makes all of the difference.

I'll illustrate with a small example.

Game 1
Out
Out
Out
Out

Game 2
Out
Out
Out
Out
Out

Game 3
Out
Out
Out
Hit

Game 4
Hit
Out
Out
Out
Out

Game 5
Out
Out
Out
Hit

Game 6
Hit
Hit
Hit
Hit

So, here we have 6 games and a total of 26 AB. There are many ways we can descrbe this series of events.
- The player has a 4 game hit streak
- The player has a 5 at bat hit streak
- The player is hitting .269 over the last 6 games.
- The player is hitting .318 over the last 5 games.
- The player is hitting .412 over the last 4 games.
- The player is hitting .500 over his last 14 AB.
- The player is extremely streaky, never alternating hit-out-hit or out-hit-out.

Is this good? Is this bad? What happened in Game 0? Where should we draw the line to describe his recent performance? The typical answer is that we draw the line based on what conclusion we'd like to reach. This is hardly an intellectually honest way to approach player evaluation. Though in the desperate need to put a name, a story on the randomness of events for the purposes of entertainment, we (announcers in particular) end up slicing the past up in to chunks that fit a certain narrative.

The one thing I will say is that there is an argument on behalf consistency. In short, it's that you will win more games scoring 5 runs every game than you will scoring 0 runs in half your games and 10 runs in the other half. I wish I could find the study on which this assertion is based, but I couldn't seem to dig it up.

I think it's only reasonable to extrapolate this to players, but in a limited manner. A player who contributes on a more evenly distributed basis provides his production more efficiently, meaning he is more likely to provide production when that production has a significant effect on the outcome of games.

What I don't know is the extent of this effect and how to describe it in the context of other values. Even if we were to establish that 1) consistency is a desirable trait and 2) a given player is extremely inconsistent, so what? How do we account for that in our valuation of that player. My gut says that we grossly overestimate the effect, due to the human tendency of biased selection of events. We imagine events happening in an different order and grossly underestimate the amount of random variation going on in every facet of the game.

bucksfan2
06-13-2008, 02:38 PM
Unfortuantly streakiness is hard to define because as RMR has shown its hard to find a starting or stoping point in order to define the streak. To take it even further it is very easy to manipulate the days in order to create the point you want to make. I wonder if someone has ever examined streaks or variance in different spans. I wonder if you went by every 10 games, every week, every series, every road trip, and broke all that down into data points if it would show you any sort of worthwile variance.

Since it is difficult to discern when a streak starts or stops it seems that there is an overexageration due to playing time during streaks. I would assume that many managers are likely to buy into a "hot" bat over a "cold" bat. I think more than anything there needs to be a strong visual aspect concerning playing time and streaks. One could simply be unlucky or one could be lost at the plate or one could be Corey Patterson and it doesn't matter.

nate
06-13-2008, 06:21 PM
The one thing I will say is that there is an argument on behalf consistency. In short, it's that you will win more games scoring 5 runs every game than you will scoring 0 runs in half your games and 10 runs in the other half. I wish I could find the study on which this assertion is based, but I couldn't seem to dig it up.

I think this is the jist of what I'm asking. That is, a guy who gets, say, 4 hits every 5 games throughout a season may contribute more to winning than a guy who goes 4 for 4 one night and then disappears for the rest of the week.

Edskin
06-24-2008, 01:16 PM
IMO, Adam Dunn is one of those guys who's impact on a team doesn't quite match his numbers. In his case, and many similar cases, it is very difficult to articulate my point because the numbers ARE against me. In the end, Adam Dunn (and others) generate quality numbers.

But I've always wondered about "streakiness, consistency, and clutchness" and how they all interact with overall numbers.

I've always maintained that "clutchness" goes WAY beyond hitting a HR in the 9th inning.

Very interesting stat on Sunday about Andy Pettite and his career with the Yankees... In his career there, he has had 20 starts when his team was facing a sweep in a series....in those 20 starts, he was 15-1 with an EAR under 2.00

To me, that is a "clutch" performer. Stepping up when your team needs you.

It was also one of the first statistics I've ever seen that adds even a tad of actual analysis to one of these theories.

I've also always wondered about how streakiness affects a player and his team.

Let's say you play a double-header and you go 4-4 in game 1 and then go 0-4 in game 2 and leave a bunch of runners on base.

At the end of the day, your numbers are phenomenal, and you absolutely helped your team win game 1, but your failures in game 2 may have equally contributed to a loss.

My question is...would it be better to go 2-4 in both games?

Of course, this is an exxagerated example...but you can play this out over the course of an entire season.

I hate the fact that this far into his career Adam Dunn is still prone to incredibly long slumps at the plate. Everyone has ups and downs, but it's always bothered me that Dunn will have weeks where he looks inept. Of course, he normally offsets that numbers-wise by going on long streaks of high OPS.

But to me, the most frustrating players to go against are always the consistent ones-- the ones that you know you can't hold down over the course of a series.

I've always thought that the ability to avoid the prolonged slump was something that made good players great.

Curious if anyone knew of any stats out there that focused on this?

westofyou
06-24-2008, 01:17 PM
I hate the fact that this far into his career Adam Dunn is still prone to incredibly long slumps at the plate.

Johnny Bench 1976 says hello

SunDeck
06-24-2008, 01:30 PM
I hate the fact that this far into his career Adam Dunn is still prone to incredibly long slumps at the plate. Everyone has ups and downs, but it's always bothered me that Dunn will have weeks where he looks inept.



Those are the weeks when his mind is elsewhere- on things like bass fishing, his kids, other stuff. Clearly, if he only liked the game more, he'd be more consistent. ;)

edabbs44
06-24-2008, 01:38 PM
I have wondered the same thing...nice post.

Someone blowing up for 4 HRs in a game is great, but if it was a 14-1 whitewashing, 4 HRs spread over 4 games might be more valuable.

nate
06-24-2008, 01:53 PM
I recently (http://www.redszone.com/forums/showthread.php?t=69197&highlight=streakiness) asked this same question. I'd always wondered which players were most prone to streaks and slumps and what league "consistency" is.

PuffyPig
06-24-2008, 02:05 PM
Let's say you play a double-header and you go 4-4 in game 1 and then go 0-4 in game 2 and leave a bunch of runners on base.

At the end of the day, your numbers are phenomenal, and you absolutely helped your team win game 1, but your failures in game 2 may have equally contributed to a loss.

My question is...would it be better to go 2-4 in both games?



I think the answer i generally no.

4 hits is 4 hits.

if you go 4-4 in one game, I would think that your chances of winning that game increase. Just like a 0-4 decreases your chances. Going 2-4 in each game makes less difference in each game.

I would expect that your team will win just as many games if you go 4-4 in one game, and 0-4 in the next 3, than if you went 1-4 in each game.

It's really a matter of when you get your hits. You need to get them when your teamates get them. And that's probably a matter of random variance more than anything else.

But Pettite going 15-1 when his team is facing a sweep. That's a large enough sample size to maybe mean something. I except that it's possible for a pitcher to be clutch, more than a hitter, because a pitcher is more in control of the game than a hitter is.

Edskin
06-24-2008, 02:22 PM
But Pettite going 15-1 when his team is facing a sweep. That's a large enough sample size to maybe mean something. I except that it's possible for a pitcher to be clutch, more than a hitter, because a pitcher is more in control of the game than a hitter is.

Agreed. I also believe it's more possible for a pitcher to "prove" his "clutchness."

With Dunn, it's not so much that I don't think he isn't "clutch" or that he doesn't love the game enough.

With Dunn, for me, it's all about the fact that I do NOT believe he is the best player he could possibly be. I do not have any hard evidence to support this-- it's just based on what I've seen over the years. And I don't claim to know why or what is in his heart or his head. But I just don't think a guy with his physical gifts and his obvious abilities should sometimes struggle so mightily to do some fairly small things.

Whether or not he's the best player on the team is besides the point, IMO, when discussing whether or not he is as good as he could be.

Paul Bako is a MUCH, MUCH, MUCH bigger "problem" for the Reds than Adam Dunn will ever be, but Bako is what he is-- this is the best he'll ever be, and IMO, this is the best he COULD be.

With Dunn, I've always felt there was a superstar somewhere inside of there instead of a solid borderline all-star (which is nothing to sneeze at of course).

Red Heeler
06-24-2008, 02:45 PM
But Pettite going 15-1 when his team is facing a sweep. That's a large enough sample size to maybe mean something. I except that it's possible for a pitcher to be clutch, more than a hitter, because a pitcher is more in control of the game than a hitter is.

Maybe it means Pettite is a slacker? If he is darn near unbeatable when his team is facing a sweep, why doesn't he pitch that way every start?

bucksfan2
06-24-2008, 02:46 PM
I think the answer i generally no.

4 hits is 4 hits.

if you go 4-4 in one game, I would think that your chances of winning that game increase. Just like a 0-4 decreases your chances. Going 2-4 in each game makes less difference in each game.

I would expect that your team will win just as many games if you go 4-4 in one game, and 0-4 in the next 3, than if you went 1-4 in each game.

It's really a matter of when you get your hits. You need to get them when your teamates get them. And that's probably a matter of random variance more than anything else.

But Pettite going 15-1 when his team is facing a sweep. That's a large enough sample size to maybe mean something. I except that it's possible for a pitcher to be clutch, more than a hitter, because a pitcher is more in control of the game than a hitter is.

I disagree. You could assume that if you go 4-4 or have 4 hits in a game your team stands a chance but that really has no bearing on what the other team does or what you do. What happens if you have 4 hits but the reds of your team has 2. On the converse if you have 4 hits but the other team scores 10 runs they may be meaningless.

Also if you go 0-4 in a game that is a 1 or 2 run game that is worse than going 4-4 in a game you win 8-2. My problem with Dunn is you can look at the season ending stats and say he is the best offensive player because of his 40 HR and 100 RBI and 100 R but it doesn't tell you how those runs occured. It doesn't tell you when or the importance or in what situation. It tells you that they happened. IMO a problem with baseball stats are that when you look at the aggregate stat you lose a sense of the situational aspect.

PuffyPig
06-24-2008, 03:02 PM
I disagree. You could assume that if you go 4-4 or have 4 hits in a game your team stands a chance but that really has no bearing on what the other team does or what you do. What happens if you have 4 hits but the reds of your team has 2. On the converse if you have 4 hits but the other team scores 10 runs they may be meaningless.



What part of my post did you disagree with?

It sounds like you just confirmed everything i said.

RedsManRick
06-24-2008, 03:03 PM
When it comes to Dunn and "consistency", I think the fact that's he on base so darn much is often overlooked. I think that our gut measure of consistency focuses on how often we see the player do something exciting (putting the ball in play), as opposed to how often they do something which positively contributes to run scoring.

That is to say if we have two guys in that double header, one goes 1-3 with a BB and 0-2 with 2 BB, the other goes 2-4 and 1-4, we're inclined to think of the 2nd guy as more productive.

Also, as I discussed in Nate's thread, I think we have MAJOR selection bias problems. We choose to define the time frame based on end points that confirm the conclusion we are trying to express.

For example, take a guy who alternates games of 0-4 and 4-4 and let's say he just finished an 0-4 game. Depending on what point we wanted to make, we could say he's 4 for his last 8 or that he's 4 for his last 12. After the next game we could say that he's 8 for his last 12, or that he's 4 for his last 8. End point selection allows us to create streaks at almost any point in time. All you have to do is go back far enough.

BCubb2003
06-24-2008, 03:04 PM
Your heart might say so, but the thing is that there's not much of a way for Dunn to say, "I can go 0-4 today, but tomorrow we're going to need four hits from me."

I do think it's legitimate to try to measure some of the intangibles. And it's possible that if you have a team full of steady performers you might do better than a team full of streaky players (can't get everybody hitting at the same time).

But it might be like run differential. For any given game, it's meaningless, but it tracks very well at the end of the season.

AtomicDumpling
06-24-2008, 04:37 PM
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.

flyer85
06-24-2008, 05:08 PM
IIRC, I believe "The Book" dealt with these issues. Every player has ups and downs and they don't mean much. The best you can hope for is a somewhat consistent level of production on a yearly basis though the journey will be different from year to year.

bucksfan2
06-24-2008, 05:15 PM
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.

Good post. I also think it sheds some light on the problem that the fans have with Dunn. I think it would be safe to assume that Dunn's high OBP with runners on is largely due to his high walk totals. While walks are fine and dandy they hardly ever drive in a run and force the player/s hitting next to drive in the run. I would also assume that Dunn takes more of a passive approach to the plate rather than an attacking approach. Again this is fine except if you are looking at Dunn to drive in runners. It is especially disconcerting when the players who have hit behind Dunn have been Edwin and Bako for a large portion of the year. I think I would be correct in thinking that a lot of Reds fans would want Dunn to attept to drive in the runner than walk down to first and put it in the hands of Edwin.

Dunn possesses such a unique skill set that it is hard to place the correct value to him. He is very frustrating because of his lack of contact. His career obp is around .140 points higher than his average. The Reds wanted Jay Burce's difference to be .070 points, half of what Dunn's is. Is that discrepancy good, bad, or both? IMO if you have Dunn hitting 6th or lower you are losing productivity from him. You hit him 2nd or 3rd and you will increase his productivity. At the same time what do you do with his power? Does that drop him in the lineup because of his power and make him less valuable. A statline like Dunns tends to defy all logic which makes him an enigma.

Ltlabner
06-24-2008, 05:24 PM
Far too often in baseball people use a time reference that is way to short.

Almost any player can be made to look good, or look bad, by examining at an at-bat, inning, game, week or even month. Normal random variations in performance are suddenly called streaks or slumps when in relality they don't mean much.

A players real value to the team is accumulated over the corse of an entire season.

If Brandon Phillips goes wild until the end of the season and puts up another 30/30 year will anybody remember his various struggles at the beginning of the season? Conversly, if Paul Bako continues to be colder than an iceberg will anybody remember his strong start to the year come October?

Also doesn't help that baseball is a team sport, but people want to micro-examine a players individual stats and extrapolate them into why a team is/isn't doing well.

AtomicDumpling
06-24-2008, 05:27 PM
Good post. I also think it sheds some light on the problem that the fans have with Dunn. I think it would be safe to assume that Dunn's high OBP with runners on is largely due to his high walk totals. While walks are fine and dandy they hardly ever drive in a run and force the player/s hitting next to drive in the run. I would also assume that Dunn takes more of a passive approach to the plate rather than an attacking approach. Again this is fine except if you are looking at Dunn to drive in runners. It is especially disconcerting when the players who have hit behind Dunn have been Edwin and Bako for a large portion of the year. I think I would be correct in thinking that a lot of Reds fans would want Dunn to attept to drive in the runner than walk down to first and put it in the hands of Edwin.

Dunn possesses such a unique skill set that it is hard to place the correct value to him. He is very frustrating because of his lack of contact. His career obp is around .140 points higher than his average. The Reds wanted Jay Burce's difference to be .070 points, half of what Dunn's is. Is that discrepancy good, bad, or both? IMO if you have Dunn hitting 6th or lower you are losing productivity from him. You hit him 2nd or 3rd and you will increase his productivity. At the same time what do you do with his power? Does that drop him in the lineup because of his power and make him less valuable. A statline like Dunns tends to defy all logic which makes him an enigma.

OK, that makes sense.

My thoughts are the data show that Dunn does not get good pitches to hit with runners on base and 1st base open. If he were to swing at balls outside the zone he is much more likely to make an out than to drive in a run. Unless you are Vlad Guerrero bad things happen when you swing at balls. Good things happen when you swing at strikes.

You are correct that people perceive Dunn as an RBI guy, but in reality he is actually better at scoring runs than driving them in. He has more runs scored than RBIs in his career despite batting down in the lineup with poor batters behind him. His OBP skills would be much better utilized near the top of the lineup with good hitters behind him. Moving him up in the lineup in order to enable him to score more runs would also have the fringe benefit of enabling him to see better pitches to hit, which ironically would increase his slugging percentage.

I agree with you regarding Dunn having an unusual set of hitting skills. It requires a manager with good thinking skills to take proper advantage of his abilities. Unfortunately Dusty Baker likes to manage in an old-school style that no longer works in today's game of baseball.

westofyou
06-24-2008, 05:28 PM
Far too often in baseball people use a time reference that is way to short.

I call it the NFL effect

Little moments of data and the need to quantify it as important in the big picture.

Sure it works with one game a week, it fails miserably in baseball most of the time.

Ltlabner
06-24-2008, 05:33 PM
I call it the NFL effect

Little moments of data and the need to quantify it as important in the big picture.

Sure it works with one game a week, it fails miserably in baseball most of the time.

Good point.

Not to mention that in the NFL if you make the play that wins the game you've helped your team win 6% of their overall schedule. If you help your baseball team win one game, you've just acomplished the mighty task of securing .006% of your teams schedule.

Far too often people say, "if only Dunn could be more consistent". Yea, that's true. But it overlooks that other forces impact whether he's consistent, most of which are outside of his controll. It's not like he strolls to the plate and thinks, "gee, I don't think I feal like hitting this ball for a double right now".

nate
06-24-2008, 05:43 PM
Good post. I also think it sheds some light on the problem that the fans have with Dunn. I think it would be safe to assume that Dunn's high OBP with runners on is largely due to his high walk totals. While walks are fine and dandy they hardly ever drive in a run and force the player/s hitting next to drive in the run.

I'm totally fine with this because:

1. He hasn't made an out
2. He's created the possibility to score more runs


I would also assume that Dunn takes more of a passive approach to the plate rather than an attacking approach.

I think his approach is to wait for a pitch he can hit and mash it.


Again this is fine except if you are looking at Dunn to drive in runners. It is especially disconcerting when the players who have hit behind Dunn have been Edwin and Bako for a large portion of the year.

I think he should bat 1-3 myself. He's cemented into that 5-spot by the Duster.


I think I would be correct in thinking that a lot of Reds fans would want Dunn to attept to drive in the runner than walk down to first and put it in the hands of Edwin.

I agree that a lot of fans think that, yes.


Dunn possesses such a unique skill set that it is hard to place the correct value to him. He is very frustrating because of his lack of contact. His career obp is around .140 points higher than his average. The Reds wanted Jay Burce's difference to be .070 points, half of what Dunn's is. Is that discrepancy good, bad, or both?

I think it's great!


IMO if you have Dunn hitting 6th or lower you are losing productivity from him. You hit him 2nd or 3rd and you will increase his productivity. At the same time what do you do with his power? Does that drop him in the lineup because of his power and make him less valuable. A statline like Dunns tends to defy all logic which makes him an enigma.

I think if you have him hitting 5th or lower, you're not getting him the opportunities to both drive in and score a lot of runs. A guy who's on base that much should be at bat as many times a game as possible.

Johnny Footstool
06-24-2008, 05:44 PM
But Pettite going 15-1 when his team is facing a sweep. That's a large enough sample size to maybe mean something. I except that it's possible for a pitcher to be clutch, more than a hitter, because a pitcher is more in control of the game than a hitter is.

The question is, why wasn't Pettite able to perform like that when his team *wasn't* facing a sweep? Was he just slacking off, sandbagging, holding his best effort until he felt he needed it? ;)

Stormy
06-24-2008, 05:45 PM
I'm not sure that such a concept exists in the game of baseball, Edskin. Given the duration, and the variables, which inherently factor into a 162 game schedule, it is very difficult to maintain a measurable consistency during every segment of the season. Although certainly high contact 'hitting machine' prototypes normally exhibit more consistency than most on a month in, month out basis.

Charting any players individual season production is like tracking the progress of a ship crossing the Atlantic: from a distance the course may appear to be a straight line, but upon closer inspection that straight line consists of innumerable ebbs and flows. A player of Adam Dunn's relatively low average, low contact skill set, certainly demonstrates more volatility in his month to month performance (and is more prone to prolonged droughts) than are more elite contact driven hitters, but some of that is also mitigated by his ability to still get OB during those dry spells via the walk.

Regardless, even when you look at many MVP candidate type seasons, you'll see the erratic production of players who OPS 1100 in April, and 850 in May, or who follow a .330 12HR month with a .250 7HR month, only to continue those hot and warm variations throughout their season. For example, just looked at Griffey circa 1996 (49HR, 140RBI, 1020OPS season): He basically enfolded multiple .260-.270BA, 860-890OPS months, within multiple .350-.360BA, 1200OPS type months). Huge, volatile variations between productive and wildly productive over the course of a year, even within an elite season.

Consistency certainly exists, and even moreso for players from certain molds, but in baseball consistency never equals constancy, in the way it can be measured in so many other sports.

Stormy
06-24-2008, 06:04 PM
I hate the fact that this far into his career Adam Dunn is still prone to incredibly long slumps at the plate. Everyone has ups and downs, but it's always bothered me that Dunn will have weeks where he looks inept. Of course, he normally offsets that numbers-wise by going on long streaks of high OPS.

I've always thought that the ability to avoid the prolonged slump was something that made good players great.

I just looked at Mike Schmidt's 1980 MVP season, one of the most productive of any MLB thirdbaseman in history, and en route to compiling his 48HR 121RBIs, Schmidt spent the entire months of June and July hitting between .231 and .244, with a paltry 11HR, and a middling .800s OPS during that span. That is an extraordinarily 'clutch' performer in the midst of an elite season, and yet there are still protracted struggles.

I don't disagree with you at all regarding the frustrating nature of Adam Dunn's sporadic performance, en route to his exceptional and consistent production. However, I think it is somewhat commonplace, even among many elite players of similar high power, low contact profile.

johngalt
06-24-2008, 10:10 PM
If you're just looking at consistency, couldn't you take someone's AVG, OBP, SLG or OPS over the course of the season (or their career) and chart the standard deviation?

bucksfan2
06-25-2008, 03:22 PM
Question for the stat guys. Has anyone tried to measure a given players productivity of an out? Yes I know the idea of baseball is to avoid outs at all cost but in reality each team has 27 outs in a game. I have never heard of a game being called because one team was unable to use all the aloted outs.

The way I look at it each team is given 27 (24) outs and you are to avoid or use those outs the most productive way you can. A tripple play is obviously the worst but that is more a freak occurance than anything. A double play is damning because it consist of two outs. A strike out is also awful because the runner is not advanced by a strike out. Because this topic is largely due to Dunn I checked out Dunn's 07 season. 26% of the time ( K/(ab+bb+hbp)) Dunn struck out last season. I am making an assumption here, but that would mean around 1/4 of the time Dunn came to the plate with runners on he automatically left without doing anything to move the runner along.

westofyou
06-25-2008, 03:43 PM
Has anyone tried to measure a given players productivity of an out?

Yeah and it crashed an burned like the Hindenberg.

http://sports.espn.go.com/mlb/columns/story?columnist=olney_buster&id=1792101

http://www.hardballtimes.com/main/article/yet-another-productive-outs-article/

http://www.baseballmusings.com/cgi-bin/mt/mt-tb.cgi?__mode=view&entry_id=6584

AtomicDumpling
06-25-2008, 04:04 PM
Question for the stat guys. Has anyone tried to measure a given players productivity of an out? Yes I know the idea of baseball is to avoid outs at all cost but in reality each team has 27 outs in a game. I have never heard of a game being called because one team was unable to use all the aloted outs.

If the home team wins they often use only 24 of their outs. :cool:

I see your point though. The flip side of the issue is the best way to look at it. Using all the outs is not the problem -- the goal is to continue batting by not making outs. It drives me crazy when the Reds send a batter to the plate to intentionally make an out. Even if a sacrifice is successful (and they only succeed about 60% of the time), it rarely improves your chances of winning the game. Only terrible hitters (like pitchers and Corey Patterson) should be bunting, good hitters should try to get on base or at least drive the runners in. Even if the bunt is successful it only slightly increases your chances of scoring one run, but severely hurts your chances of scoring more than one run.



The way I look at it each team is given 27 (24) outs and you are to avoid or use those outs the most productive way you can. A tripple play is obviously the worst but that is more a freak occurance than anything. A double play is damning because it consist of two outs. A strike out is also awful because the runner is not advanced by a strike out. Because this topic is largely due to Dunn I checked out Dunn's 07 season. 26% of the time ( K/(ab+bb+hbp)) Dunn struck out last season. I am making an assumption here, but that would mean around 1/4 of the time Dunn came to the plate with runners on he automatically left without doing anything to move the runner along.

The analyses I have seen show that on average a strike out is no more harmful than hitting into an out. The chances of advancing the runner while making an out are canceled out by the chances of hitting into a double play. There are many kinds of double plays to consider -- traditional ground ball double play, line drive out with the runner getting doubled off, hitting a fly ball and having the runner thrown out after tagging up and a couple of others.

We should focus 100% of our attention on how often a hitter makes an out rather than how he makes an out.

Mario-Rijo
06-27-2008, 01:38 AM
I think it's important to differentiate between a player who is a good or potentially good player who slumps at times as all players do and a player who isn't that good but has streaks of excellent play. I think extremely talented players who lack some specific skill or overall polish are capable if given the opportunity to have an extremely good week or so, 4-5 times a year which makes the end #'s look respectable or even good. The rest of the time the guy has below average overall production. How does that affect the team in the win loss column?