PDA

View Full Version : Defending Dusty Baker by using Stats



757690
02-26-2008, 01:31 AM
Wow, it seems that Dusty has ruffled a few feathers with his opinions on how to build a line-up, and most of those feathers belong to fans of spreadsheets and stats.

I do want to say that I am a fan of spreadsheets and stats, however, I am have not drunk the kool-aid of Baseball Prospectus or any of the other major stat crazy sites. While I believe that Bill James is correct that one day, everything will be able to be explained and understood through stats, I also agree with him when he says that we are very, very far away from that day.

My main problem with Saber fans, is that many (although not all) believe that we have reached the end of our understanding of stats. Because of this they accept that the conclusions that the Saber world has reached so far are dogma and indisputable.

The reason why I have this problem is that our understanding of stats is at the very beginning stage and many mistakes have already been made. A perfect example is Voros McCracken and his theory on DIPS. That has been disproven and is no longer accepted as fact. It has been reworked by Tom Tippett and Mitchel Lichtman and a new better theory of DIPS is now available. I am certain that this will be the case with all Saber theories, they will all be reworked, re-examined and improved on in the future.

I am going to use two examples to show that we have just touched the surface of understanding stats, and these two examples also defend Dusty Baker's views on building a line-up.

I do want to add that I am not saying that Dusty absolutely is right, just that a defense can be made, by more closely examining stats.


Dusty's First View: A Speedy Leadoff Hitter Who Steals Bases Is Better Than One With A High OBP


This is rather simple, actually, although the stats only show that a base stealing leadoff hitter can be better than a high OBP hitter, depending on the numbers.

Everything else being equal, a base stealer will get to second more often than a hitter who does not steal bases. That is a tautology. And a runner on second will score more often than a runner on first. That is indisputable. Therefore, if two player's OPB are equal, than the one who steals more bases will score more often, and be a better lead off hitter (unless you believe the main purpose of a batter is to get on base instead of scoring or driving in a run.)

This also means that there will be situations where Player A could have a higher OPB, but since Player B steals a certain number of bases more than Player A, Player B is a better leadoff hitter. It all depends on the difference between the OBPs and the number of bases stolen. If you do the math, it comes out that every 10 stolen bases is worth around .005 OPB points. So if Player A has an OPB of .380 and Player B has one of .365, and Player B steals 40 more bases than Player A, then Player B is the better leadoff hitter.

Therefore, Dusty's desire to want a speedy leadoff hitter is justifiable, if the leadoff hitter steals enough bases.


Dusty's Second View: The Middle Of The Lineup Should Favor SLG Over OBP.


This is far more complicated so I will only provide a summery of my examination of the stats. Anyone who wishes to see the full work can send me a private message and I will send it to them.

Here is the general theory behind it, and I will show the stats that back up this general theory.

A middle of the batting order hitter comes to the plate with a runner on first more times than any other situation with runners on base. In fact even more often than with a runner on second and a runner on third combined. Therefore, a middle of the lineup hitter who excels at scoring the runner from first is more valuable than a one who can excel at scoring a runner from second or third.

Of course it depends on the actual stats. If Player A can score a runner from second and third five times more often than Player B, and if Player B can only score a runner from first twice as often Player A, than Player A is more valuable in the middle of the lineup.

So just to see what the ratios are, I selected two middle of the lineup hitters, one known for his OPB, J.D. Drew, and one know for his SLG, Adrian Gonzalez. I am sure people can find other examples with different results, but I am only trying to show that Baker can be right, not that he is right all the time.

Here are the numbers:

Drew: .373 OPB .423 SLG
Gonz: .347 OPB .502 SLG


These translate to AG scoring a runner from first at a .327 rate, and JD scoring a runner from first at a .208 rate. That is a huge difference. And since scoring a runner from first base is more important, since it is a more common scenario, than JD really needs to kick some serious tail at scoring runs from second and third. He does do better than AG, which makes sense considering his higher OPB, but by only .314 to .301. That is not nearly enough for him to overcome AG's huge advantage in scoring a runner in from first.

So basically, the stats show that in the middle of the lineup, is better to have a guy with a high SLG and a low OBP (unless he has a ridiculously high OBP), than the other way around. Meaning that Dusty is correct in wanting a hitter to be more worried about getting a good pitch to hit and driving the ball, than getting a walk. His goal should not be to not make an out, but to drive the ball for an extra base hit. That will help the team win more games.


Anyway, have fun ripping this apart. I had fun putting it together.

757690
02-26-2008, 01:02 PM
I just want to correct my stats in the second part, I copied the wrong numbers from my spreadsheet. Here is the correct breakdown between AG and JD

Scoring runner on first: AG - .242, JD - .223
Scoring runner on 2 or3: AG - .306, JD - .309.

Sorry about that, still it makes the same point, just not as drastically.

Bip Roberts
02-26-2008, 01:14 PM
solid stuff but this post makes me glad I stay away from deeply analyzing stats :laugh: I would drive my self nuts

Blue
02-26-2008, 01:21 PM
Remember that if a runner is caught stealing a base, he has effectively lost every base before that. So, a runner caught stealing 2B should lose the hit or walk that got them on base.

AmarilloRed
02-26-2008, 01:28 PM
Remember that if a runner is caught stealing a base, he has effectively lost every base before that. So, a runner caught stealing 2B should lose the hit or walk that got them on base.

Good point. A runner who is caught stealing or picked off base is less effective than one who simply stays on base and makes intelligent base-running decisions. Under those circumstances, a high OBP is better than a base-stealer.

757690
02-26-2008, 01:54 PM
Remember that if a runner is caught stealing a base, he has effectively lost every base before that. So, a runner caught stealing 2B should lose the hit or walk that got them on base.

That is figured into the equation, I just didn't mention it, sorry, my bad.

I did exactly what you said. I just subtracted the CS from the times on base when figuring out each players OBP. I used a 80% formula, which is about average for most base stealers. So if a player stole 40 bases, I subtracted 14 times on base in figuring out his OBP. That gets you the ratio of 10 SB = around .005 OBP.

mlbfan30
02-26-2008, 02:13 PM
I'll go through this article a little more, but here are the Linear Weights Values...

Basestealing Runs--The number of runs a batter gained through his stolen base attempts, figured as .22*SB - .38*CS.

Last year Hopper was 14-6 so that's 3.08-2.26 = 0.8. That 0.8 is Runs gained by stealing, which translates to less than 1/10 of a Win.

Freel was 15-8. so thats 3.3-3.04 = 0.26 Runs.. Even less

The point is these guys are not very good base stealer's. They could hurt the team more than help, just on the basepaths. If A different player had a higher OBP but didn't steal at all, he would be more valuable. Now for those of you who say what about baserunning, well baserunnung has a value between +.5W and -.5W. At most there would be a 1 W difference. I could get baserunning values later.

And BP 32-8 is a very good basestealer. He would be 7.04-3.04 = 4 Runs. That is still worth less than 1/2 a Win.

If you have a speedy guy and a slow guy with the same OBP to leadoff, obviously you use the speedy guy. But if that were true, it wouldn't be an augment. The augment is that they want to put a speedy guy in with a much lower OBP.

mlbfan30
02-26-2008, 02:15 PM
That is figured into the equation, I just didn't mention it, sorry, my bad.

I did exactly what you said. I just subtracted the CS from the times on base when figuring out each players OBP. I used a 80% formula, which is about average for most base stealers. So if a player stole 40 bases, I subtracted 14 times on base in figuring out his OBP. That gets you the ratio of 10 SB = around .005 OBP.

And it doesn't work like this... you just can't do this

757690
02-26-2008, 03:28 PM
I'll go through this article a little more, but here are the Linear Weights Values...

Basestealing Runs--The number of runs a batter gained through his stolen base attempts, figured as .22*SB - .38*CS.

Last year Hopper was 14-6 so that's 3.08-2.26 = 0.8. That 0.8 is Runs gained by stealing, which translates to less than 1/10 of a Win.

Freel was 15-8. so thats 3.3-3.04 = 0.26 Runs.. Even less

The point is these guys are not very good base stealer's. They could hurt the team more than help, just on the basepaths. If A different player had a higher OBP but didn't steal at all, he would be more valuable. Now for those of you who say what about baserunning, well baserunnung has a value between +.5W and -.5W. At most there would be a 1 W difference. I could get baserunning values later.

And BP 32-8 is a very good basestealer. He would be 7.04-3.04 = 4 Runs. That is still worth less than 1/2 a Win.

If you have a speedy guy and a slow guy with the same OBP to leadoff, obviously you use the speedy guy. But if that were true, it wouldn't be an augment. The augment is that they want to put a speedy guy in with a much lower OBP.


I was never arguing for Hopper or Freel to lead off. I personally think Bruce should be the centerfielder and Votto should lead off.

But you missed my point, which is that those stats that you quoted are not gospel, I was trying to show a different way to look at stats than the way that fans of Sabermetrics are looking at them now. I am not saying that my way is better, just that there is more than one way to interpret stats.

I can promise you that

"Basestealing Runs--The number of runs a batter gained through his stolen base attempts, figured as .22*SB - .38*CS"

and

"baserunnung has a value between +.5W and -.5W. At most there would be a 1 W difference."

are not definitive. A better way to figure out the effect of basestealing will be developed, and re-developed, and re-developed... I think I have come up with one way to re-develop it, which has different results. I can gladly show my work in greater detail.

That is my point.

757690
02-26-2008, 03:34 PM
And it doesn't work like this... you just can't do this

Please explain why. I am not doubting you, I just want to know why since it make perfect sense to me.

I am arguing basically that a stolen base should add a base to your total bases used for calculating SLG, and a caught stealing should be subtracted from your total bases used for calculating OBP. This seems very intuitive to me.

Newman4
02-26-2008, 04:17 PM
If I recall, Linear Weights are determined by using actual baseball data to show the probability of scoring runs. .22 per SB means that each successful SB scored a run 22% of the time and .38 per CS means that a caught stealing takes away a run 38% of the time.

Here's a good article on OPS with correlation stats BTW:

http://www.hardballtimes.com/main/article/ops-for-the-masses/

SMcGavin
02-26-2008, 05:52 PM
Dusty's First View: A Speedy Leadoff Hitter Who Steals Bases Is Better Than One With A High OBP


This is rather simple, actually, although the stats only show that a base stealing leadoff hitter can be better than a high OBP hitter, depending on the numbers.

Everything else being equal, a base stealer will get to second more often than a hitter who does not steal bases. That is a tautology. And a runner on second will score more often than a runner on first. That is indisputable. Therefore, if two player's OPB are equal, than the one who steals more bases will score more often, and be a better lead off hitter (unless you believe the main purpose of a batter is to get on base instead of scoring or driving in a run.)

This also means that there will be situations where Player A could have a higher OPB, but since Player B steals a certain number of bases more than Player A, Player B is a better leadoff hitter. It all depends on the difference between the OBPs and the number of bases stolen. If you do the math, it comes out that every 10 stolen bases is worth around .005 OPB points. So if Player A has an OPB of .380 and Player B has one of .365, and Player B steals 40 more bases than Player A, then Player B is the better leadoff hitter.

Therefore, Dusty's desire to want a speedy leadoff hitter is justifiable, if the leadoff hitter steals enough bases.



My first question is how are you getting that 10 stolen bases = .005 OBP points? As someone else pointed out you have to factor in the times caught stealing. I don't have the exact number but I think it's around 70% - if you don't steal successfully more often than that percentage you are taking away runs.

I don't think anyone would argue with you that if OBP is equal, you take the faster guy. What happens a lot in MLB is that a fast guy with a vastly inferior OBP hits leadoff instead of a slower guy with a pretty good OBP. Even if the difference was .365 with speed against .380 slow guy, I don't think many people would complain.



My main problem with Saber fans, is that many (although not all) believe that we have reached the end of our understanding of stats. Because of this they accept that the conclusions that the Saber world has reached so far are dogma and indisputable.

I honestly don't know where you are getting this - I bet there is not one person on this board who thinks that our current conclusions about baseball stats are "dogma and indisputable". People use stuff like OPS, Runs Created, VORP, etc. not because they are perfect, but because they are a lot better than the AVG/HR/RBI stuff everyone used ten years ago. You want to make the argument that the current sabrmetric stats aren't perfect - the problem is you're arguing against a position that nobody holds. We all agree they aren't perfect, but they are among the best stats we have at the moment, so that's why they get used on this board.

wlf WV
02-26-2008, 07:36 PM
What about each individual pitcher's ability or not to be distracted by base stealers and the effects of this on his pitches to batter,or individual batters distraction,or second basemans moving to bag ,or shortstop,or outfielders positioning, or managerial decisions,are these varibles factored in.

mlbfan30
02-26-2008, 07:49 PM
What about each individual pitcher's ability or not to be distracted by base stealers and the effects of this on his pitches to batter,or individual batters distraction,or second basemans moving to bag ,or shortstop,or outfielders positioning, or managerial decisions,are these varibles factored in.

Yes... What BaseRuns does is look at all occurrences in baseball for every outcome and every play. It assigns values that model numbers from real games. What each constant is, is actually coefficients that tells what is likely to happen. These values account for context such as league/park/pitchers/defense/etc. Your "nuances" of the game are already embedded in these equations because they represent the real result.

reds2221
02-27-2008, 08:17 AM
just like what SM said, I really do not know where you got the sb=.005 obp, but if that is true, then that would clearly mean that unless a player steals many bases in a season, they are clearly not as valuble as a leadoff hitter as someone with a higher OBP (in most cases)

Bip Roberts
02-27-2008, 09:54 AM
just like what SM said, I really do not know where you got the sb=.005 obp, but if that is true, then that would clearly mean that unless a player steals many bases in a season, they are clearly not as valuble as a leadoff hitter as someone with a higher OBP (in most cases)

If you remove the mental factor

757690
02-28-2008, 08:27 AM
I honestly don't know where you are getting this - I bet there is not one person on this board who thinks that our current conclusions about baseball stats are "dogma and indisputable". People use stuff like OPS, Runs Created, VORP, etc. not because they are perfect, but because they are a lot better than the AVG/HR/RBI stuff everyone used ten years ago. You want to make the argument that the current sabrmetric stats aren't perfect - the problem is you're arguing against a position that nobody holds. We all agree they aren't perfect, but they are among the best stats we have at the moment, so that's why they get used on this board.

I probably overstated my case with the terms "dogma and indisputable". However, here are a few quotes I pulled off this board (both ORG and SunDeck) recently. These may not prove that the people who wrote them think that their stats are dogma, but they sure are certain about them, and very arrogant. I feel it is worth arguing against statements like these:

"But when it comes to making out a lineup card, I have little doubt that Dusty will construct a lineup that scores less runs than it otherwise might."

"A leader that makes bad choices still makes bad choices that hurts the team...take a look at some of the players he has at the top of the order. It's just plain stupid lineup construction."

"I'll certainly care a lot less about lineup construction given it's relative impact to a bunch of other things. It's just a tad frustrating to see somebody make the same mistakes over and over again."

"I know, intellectually, that in the whole scheme of things lineup construction is not terribly important. But emotionally, I can help from wanting to lash out against a manager who is simply clueless when it comes to how the game he loves actually works."

"The game has changed a lot in the last 20 years. New methods of understanding the game have arisen. The advent of computers has allowed people to quantify which approaches to managing have been the most successful and which ones have led to failure. If Baker doesn't understand these changes I feel the Reds will not fulfill their potential under his tenure."

"(Data + Experience) > Data > Experience"

"So, if a truly great manager (pick whoever you want) had no access to stats and just went on what he saw, you would think a computer would do a better job?" "In terms of making out lineups and deciding on in-game strategy? Yes."

"This implies that he doesn't understand the value of walks. OBP is not overvalued. It is undervalued by guys like him."

hebroncougar
02-28-2008, 11:00 AM
Here is how you "defend" Dusty and some of his decisons................

LouisvilleCARDS
02-29-2008, 02:10 AM
I've lurked this board for a while, and I read a lot of hate on Baker, but I for one am happy we have ANYONE here who's gotten to the World Series, managed contender, and won manager of the year awards. Not to mention his players say nothing but good things about him.

In this age of fantasy baseball things and breaking down the most intricate stats, I think everything is scrutinized way too much. I'm sorry to say, but the Reds haven't made the playoffs in 13 years now, not because a manager decided to put a guy with a .080 higher OBP in the 6 spot in the lineup instead of the 4. We're not losing because a guy is batting leadoff who walks 10% less than some other guy.

Lets keep it simple: the Reds haven't been to the playoffs in 15 years because this team is consistently near or AT the bottom in the major leagues in ERA EVERY SINGLE YEAR without fail. For years it was the starting pitching that was horrific, with a mismesh of washed up hasbeens and never were's in the rotation, and a very good bullpen to rely on. Now the bullpen is completely horrible, and the rotation, while not good - actually has a couple of fairly reliable starters and a great ace pitcher. And you can play shuffleboard with the pitching all day and night as a manager, and dress it up as much as possible, but in the end crap is crap.

There's no need to make it more complicated than it is. While Dusty's managerial style can be critiqued to death - I like the fact he's had nine winning seasons and made the playoffs four times - which as far as the Reds organazation goes, you would have to go back to 1976 to reach 4 playoff appearances. This guy is a veteran manager, not a stopgap, low cost, interim upgrade guy. And I don't think the decisions he makes are whats going to decide the team's fate. It's going to be the consistent year in year out question of when are we actually ever going to get some decent pitching?

TN Red Fan
03-02-2008, 11:07 AM
Stealing bases also opens up the middle of the defense because one of the middle infielders has to shade over to the bag. You generally do that with a runner on first anyway, but if the guy can't run, it's just a pretense, the infielder is looking to move toward the hole.

Small ball in general keeps the team involved, and keeps guys out of the ruts that often happen when people are just sitting around waiting for a home-run.

Another point is that stolen bases don't just happen in random situations, so averaging out the numbers doesn't work. You don't steal bases when you're up by 5, you do it in the 8th inning of a tie game, so the runs created by stealing and small ball in general probably amount to more wins than the formula gives credit to.

A good manager also knows when to run. You don't run against a strong catcher or a quick pitcher, obviously. But a smart manager will know when the batter is looking at an off-speed pitch, which are really hard to throw from. On the other hand, fastballs up and away help the catcher a lot. There are other factors, as well.