Statistics & You

[Derwood]

I think it was Tim who suggested we should switch from using OBP to using 1-OBP as a proxy for how often a batter makes an out. It would put the emphasis more on what really matters - avoiding outs. Juan Pierre makes an out 67.0% his PAs, and Youkilis makes an out 62.0% of his PAs.

 

We were going to call it Suck Factor, right? I'm still on board for this.

 

SuFa? Sounds good, until Bill James adjusts it for park factors and puts out SuFA+

[Amazing_Grace]

"baseball people" overvalue the SB managing baseball teams, they allow more attempts than are really justified, and end up hurting the team.

 

I agree, but I think it works a little differently. What they fail to do is recognize the damage of the caught stealing. So instead of overvaluing the SB, they undervalue the CS. You never hear about SB%, what you hear is "This guy can steal you 40-50 bases". It's similar to how they talk about pitchers. A 15-16 pitcher means you are "a 15 game winner", while a 10-2 record means you aren't a "15 game winner". Baseball people think in terms of counting stats, and ignore the rates far too often.

 

Yeah, the counting stats are meaningless though, especially when you take the positives only like hits, SBs, wins, and ignoring the negatives like outs, CS, losses. Saying a pitcher has 15 wins like it's a good thing when they also have 16 losses is like saying Ford has 4 billion in revenues while ignoring the fact that the company's expenses were 4.5 billion and they lost half a billion dollars. Calculating value in baseball and business is not much different. If you want to know how valuable something is, you take what you invested, whether it's dollars or at-bats, and see what you got out of those dollars or at-bats. The more output you get out of your inputs, the more valuable an investment, business, or baseball player you have. It confounds me to no end how ignorant the supposedly learned "baseball people" can be.

 

prime example:

 

2006 Juan Pierre lead the league in hits. He also lead the league in outs made. The result, 200+ hits and a batting average well under .300. 700 AB's for a leadoff hitter is inexcusable

 

I was thinking of Pierre when I mentioned hits, he's the textbook example of why counting stats don't matter. His SB% is only a little above the breakeven point (around 74% last year), but a lot of people think he's one of the league's best basestealers because he gets so many.

[Magnetic Curses]

"baseball people" overvalue the SB managing baseball teams, they allow more attempts than are really justified, and end up hurting the team.

 

I agree, but I think it works a little differently. What they fail to do is recognize the damage of the caught stealing. So instead of overvaluing the SB, they undervalue the CS. You never hear about SB%, what you hear is "This guy can steal you 40-50 bases". It's similar to how they talk about pitchers. A 15-16 pitcher means you are "a 15 game winner", while a 10-2 record means you aren't a "15 game winner". Baseball people think in terms of counting stats, and ignore the rates far too often.

 

Yeah, the counting stats are meaningless though, especially when you take the positives only like hits, SBs, wins, and ignoring the negatives like outs, CS, losses. Saying a pitcher has 15 wins like it's a good thing when they also have 16 losses is like saying Ford has 4 billion in revenues while ignoring the fact that the company's expenses were 4.5 billion and they lost half a billion dollars. Calculating value in baseball and business is not much different. If you want to know how valuable something is, you take what you invested, whether it's dollars or at-bats, and see what you got out of those dollars or at-bats. The more output you get out of your inputs, the more valuable an investment, business, or baseball player you have. It confounds me to no end how ignorant the supposedly learned "baseball people" can be.

 

prime example:

 

2006 Juan Pierre lead the league in hits. He also lead the league in outs made. The result, 200+ hits and a batting average well under .300. 700 AB's for a leadoff hitter is inexcusable

 

I was thinking of Pierre when I mentioned hits, he's the textbook example of why counting stats don't matter. His SB% is only a little above the breakeven point (around 74% last year), but a lot of people think he's one of the league's best basestealers because he gets so many.

 

as i've said before, despite the popular perception, juan pierre isn't really good at anything anymore. paying him for not being good is nearly as bad as paying marquis for not being good. marquis, however, isn't paid for as long or as much.

[Rob]

Of all threads, how did I miss this one?

[KingKongvs.Godzilla]

Statheads are the Devils Spawn.

 

And how did I miss this one too? I miss the days of Stats v. Scouts threads going 90 pages.

[Amazing_Grace]

Statheads are the Devils Spawn.

 

And how did I miss this one too? I miss the days of Stats v. Scouts threads going 90 pages.

 

Is that you, Joe Morgan? :lol:

[CuseCubFan69]

I have a question for those who want to go by numbers 70% for the use of the steal of second base. Let's say Pie is called up in May. He goes 2-4 in stolen bases, do you ever send him again? He's at 50%, do you do it?

[Bull]

sample size

[CuseCubFan69]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

[Bull]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

you pick your spots. you don't let him steal at his discretion. But you don't take away that part of his game.

[Outshined_One]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

I don't think anyone on this board goes by pure numbers Cuse, you know that. Nobody is going to send Adam Dunn on a bunch of stolen base escapades over the course of the season in order to determine whether he'd be 75% successful or not.

 

Pie has an established history in the minors and the coaches know what he is capable of doing on the basepaths. Once Pie gets a reasonable sample size, then we can make the judgment as to whether or not he should be stealing with any regularity.

[CuseCubFan69]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

you pick your spots. you don't let him steal at his discretion. But you don't take away that part of his game.

 

I agree. But, for the guys that go strickly by the numbers what is a decent sample size?

[Derwood]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

you pick your spots. you don't let him steal at his discretion. But you don't take away that part of his game.

 

I agree. But, for the guys that go strickly by the numbers what is a decent sample size?

 

the sample size is too small until he reaches 70% ;)

[CuseCubFan69]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

I don't think anyone on this board goes by pure numbers Cuse, you know that. Nobody is going to send Adam Dunn on a bunch of stolen base escapades over the course of the season in order to determine whether he'd be 75% successful or not.

 

Pie has an established history in the minors and the coaches know what he is capable of doing on the basepaths. Once Pie gets a reasonable sample size, then we can make the judgment as to whether or not he should be stealing with any regularity.

 

Maybe, but that's what was debated earlier and that 70% was a solid number for them. What if a guy has been 65% effective in steals over the course of his 7 year career? What would you do then O_O?

[CuseCubFan69]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

you pick your spots. you don't let him steal at his discretion. But you don't take away that part of his game.

 

I agree. But, for the guys that go strickly by the numbers what is a decent sample size?

 

the sample size is too small until he reaches 70% ;)

 

It's like gambling, you keep trying and hope you're going to break even and you never do!

[Outshined_One]

Maybe, but that's what was debated earlier and that 70% was a solid number for them. What if a guy has been 65% effective in steals over the course of his 7 year career? What would you do then O_O?

 

As a general rule, I tend to disfavor stolen base attempts. When you have a guy who is 65% effective, that suggests to me that he either is a below average baserunner in terms of judging reads (or something comparable) or that he had a moron for a manager.

 

However, there are some cases where I'm fine with stolen bases, even if a guy does not have ideal speed. There are plenty of pitchers (Maddux) and catchers (Piazza) who are poor when it comes to holding and throwing out baserunners. These instances tend not to pop up very often.

 

More often than not, I'd rather not have guys try to steal bases.

[CuseCubFan69]

Maybe, but that's what was debated earlier and that 70% was a solid number for them. What if a guy has been 65% effective in steals over the course of his 7 year career? What would you do then O_O?

 

As a general rule, I tend to disfavor stolen base attempts. When you have a guy who is 65% effective, that suggests to me that he either is a below average baserunner in terms of judging reads (or something comparable) or that he had a moron for a manager.

 

However, there are some cases where I'm fine with stolen bases, even if a guy does not have ideal speed. There are plenty of pitchers (Maddux) and catchers (Piazza) who are poor when it comes to holding and throwing out baserunners. These instances tend not to pop up very often.

 

More often than not, I'd rather not have guys try to steal bases.

 

Interesting. I'm more aggressive than you are but I see what you're saying regarding MLB. Now, matchups can be in your favor or you see a pattern you can take advantage of, would you do both of these?

 

Some situations may call for you to make a move you normally wouldn't. Let's say it's a 1-1 game in the bottom of the 9th. You have an average pitcher and catcher steal wise and a guy that steals around 65% on first with 2 outs. You have a 250/290 hitter up with no power. The guy behind him is a 290/345 guy. What would you do?

[Mephistopheles]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

 

oh lord. Assume you have some wierd coin where the chances of it landing on heads is 70% and the chances of it landing on tails are 30%. Assume you flip it 4 times, the probability of that coin landing on heads two times or less is over 34%. That's one in three. The chances of it going two for five is still about 1 in 8. Those are pretty high. However, at the point after the fourth flip, the FIRST FOUR EVENTS HAVE NO IMPACT ON THE NEXT ONE.

 

of course stealing bases is not done in a vacuum probability matter, until the sample size is pretty very large, we're talking at least 25-30. In four attempts, what if two of them were in one game off say Carlos Zambrano (who has a great slidestep) and Hank White (who has a strong arm).

 

Theres also a belief that if you play more and run more, you're going to get better, at BP they call playing more means more production the Jeremy Giambi effect.

[Outshined_One]

Some situations may call for you to make a move you normally wouldn't. Let's say it's a 1-1 game in the bottom of the 9th. You have an average pitcher and catcher steal wise and a guy that steals around 65% on first with 2 outs. You have a 250/290 hitter up with no power. The guy behind him is a 290/345 guy. What would you do?

 

Wonder why the hell I put Neifi Perez in my lineup. ;)

[UK1679666180]

Every team knows the time of the runner to steal 2B, every team knows to pitcher's delivery time to home, and the pop time of the C to 2B. Factor in the quality of the pitcher's move to 1B and the situation of the game and ball/strike count of the AB (pitchout probability) combined with everything mentioned in the previous sentence to give you an idea of whether or not it is a good idea to steal.

 

SB% ratio tells you what HAS already happened, what I mentioned above gives you a much better idea of what WILL happen for each situation with the exact pitcher and C.

 

If you a time from pitcher to C to 2B at about 3.35 and the runner can get there at about 3.2, if the situation benefits advancing a runner and outs aren't as much of a premium (you're leading), why wouldn't you try?

Edited March 7, 2007 by UK

[CuseCubFan69]

sample size

 

Meaning? I'm aware of the small sample size but how often do you continue to steal with him? What if he gets thrown out the next time and is now 2-5? When is the sample the correct size?

 

 

oh lord. Assume you have some wierd coin where the chances of it landing on heads is 70% and the chances of it landing on tails are 30%. Assume you flip it 4 times, the probability of that coin landing on heads two times or less is over 34%. That's one in three. The chances of it going two for five is still about 1 in 8. Those are pretty high. However, at the point after the fourth flip, the FIRST FOUR EVENTS HAVE NO IMPACT ON THE NEXT ONE.

 

of course stealing bases is not done in a vacuum probability matter, until the sample size is pretty very large, we're talking at least 25-30. In four attempts, what if two of them were in one game off say Carlos Zambrano (who has a great slidestep) and Hank White (who has a strong arm).

 

Theres also a belief that if you play more and run more, you're going to get better, at BP they call playing more means more production the Jeremy Giambi effect.

 

If you're flipping coins you have no adverse effects correct? I don't think you're analogy works here.

 

I agree with what you said regarding the rest of your post. I'm just asking when a person who goes by the numbers, when would they judge those numbers to be a valid point to go by. Pie for example has struggled with his steal percentage in the minors but has improved as we have seen in the minor thread. Whe do you know a player is at their peak to judge weather or not to have him steal? Do me a favor, go into some actual scenerios and show me what you'll do those instances like I did with O_O.

[CuseCubFan69]

Some situations may call for you to make a move you normally wouldn't. Let's say it's a 1-1 game in the bottom of the 9th. You have an average pitcher and catcher steal wise and a guy that steals around 65% on first with 2 outs. You have a 250/290 hitter up with no power. The guy behind him is a 290/345 guy. What would you do?

 

Wonder why the hell I put Neifi Perez in my lineup. ;)

 

Because he's proven!

[CuseCubFan69]

Every team knows the time of the runner to steal 2B, every team knows to pitcher's delivery time to home, and the pop time of the C to 2B. Factor in the quality of the pitcher's move to 1B and the situation of the game and ball/strike count of the AB (pitchout probability) combined with everything mentioned in the previous sentence to give you an idea of whether or not it is a good idea to steal.

 

SB% ratio tells you what HAS already happened, what I mentioned above gives you a much better idea of what WILL happen.

 

If you a time from pitcher to C to 2B at about 3.35 and the runner can get there at about 3.2, if the situation benefits advancing a runner and outs aren't as much of a premium (you're leading), why wouldn't you try?

 

I agree. These are the numbers I would go by.

[Rob]

If memory serves correctly, only three teams last year had a net positive in EqSBR. Seeing as how everybody keeps posting random hypothetical situations regarding when to give the green light, I would always lean towards not letting him cost the team runs.

 

Of course, there are extreme circumstances that would change the value of a stolen base enough to make it a worthwhile risk... but those should be handled on an individual basis.

[UK1679666180]

If memory serves correctly, only three teams last year had a net positive in EqSBR. Seeing as how everybody keeps posting random hypothetical situations regarding when to give the green light, I would always lean towards not letting him cost the team runs.

 

Of course, there are extreme circumstances that would change the value of a stolen base enough to make it a worthwhile risk... but those should be handled on an individual basis.

 

I've always leaned towards more aggressive running with more information.

 

There's nothing more important as far as predicting what will likely occur if you know the times for the runner, P, and C.

 

Given current SB ratios, I can't imagine any intelligent manager wanting a player's SB ratio more than the times of the runner, P, and C.

 

Accumulative stats go in the toilet compared to play by play data, I'm glad the stat world is catching up to that idea.