Jump to content

Providing independent coverage of the Minnesota Twins.

The Store

Subscribe to Twins Daily Email

Photo

Babip

  • Please log in to reply
23 replies to this topic

#1 Sssuperdave

Sssuperdave

    Member

  • Members
  • 80 posts

Posted 02 May 2013 - 12:21 PM

I'm a numbers guy, and I love statistics, but I haven't spent a lot of time getting comfortable with sabermetrics, pitch f/x, etc. One of the sabermetric stats that has always surprised me is BABIP. I've heard a lot of people whose sabermetric opinion I highly respect refer to a high BABIP as being "lucky" (if you're a hitter, unlucky if you're a pitcher). I'm sure there is statistical evidence that backs this up, but it has never made intuitive sense to me. For example, when a hitter talks about "seeing the ball better," wouldn't that lead to a better ability to hit the ball harder and put it where he wants, causing a higher BABIP?

Anyway, I just stumbled upon this year-old article by Steve Slowinski that I think is a great take on BABIP. I apologize if it was discussed before and I missed it.

Jeremy Hellickson and re-defining BABIP

Here's one paragraph that I particularly like:

So instead of the word “luck”, I prefer to use “random variation” to describe BABIP. If a pitcher posts an abnormal BABIP, it doesn’t necessarily matter why that player’s BABIP was so extreme — it’s still liable to regress. Based on all the above criteria, certain pitchers will regress toward different means. Strikeout pitchers will do slightly better than non-strikeout pitchers. Players pitching in front of good defenses should have lower BABIPs than players on defensively-challenged teams. These variables give each pitcher a different “expected” BABIP, with most major-league starters falling within the .275-.300 range.


#2 snepp

snepp

    Curve Hanger

  • Twins Mods
  • 4,339 posts
  • LocationSioux Falls

Posted 02 May 2013 - 12:42 PM

I've always looked at the word "luck" as being a bit of a catch-all term when discussing BABIP. Likely a poor choice, given that there could be a huge amount of luck involved, or none at all.


Where I see most of the confusion surrounding BABIP is when people that use it don't realize that players will establish their own personal baseline given their skillset. Hitters primarily, but many pitchers will as well.

#3 PseudoSABR

PseudoSABR

    Twins News Team

  • Twins News Team
  • 2,014 posts

Posted 02 May 2013 - 01:20 PM

Where I see most of the confusion surrounding BABIP is when people that use it don't realize that players will establish their own personal baseline given their skillset.

This.

It's just more likely that a BABIP diverging from a player's norm has more to do with 'random variation' than it does with an improved(or regressed) hit tool.

That's why its important to look at BABIP for any player having a career year or breakout year (or sudden regression).

#4 Willihammer

Willihammer

    ice cream correspondent

  • Members
  • 3,328 posts
  • LocationSaint Paul

Posted 02 May 2013 - 01:22 PM

For example, when a hitter talks about "seeing the ball better," wouldn't that lead to a better ability to hit the ball harder and put it where he wants, causing a higher BABIP?

The pitcher scheduled to throw against that batter, might he be described as unlucky?

#5 IdahoPilgrim

IdahoPilgrim

    Senior Member

  • Members
  • 2,424 posts

Posted 02 May 2013 - 01:30 PM

Interesting article, but it also got me thinking about statistics in general. There seems to be the assumption that eventually every facet of the game will be able to described/understood/manipulated using statistics. I wonder how realistic that really is. There was a time when it was thought that the universe could be understood with mathematics, that all physical laws would ultimately be discovered and written down, allowing us to completely predict how the universe would work. Physicists now understand that, at its core, there is an element of mystery than will never fully be able to be described analytically (although that doesn't stop them from trying:)). I wonder if baseball will ultimately be the same - that at the end of the day we will realize that the game is more than just the numbers on the piece of paper, as fun as those may be to play with.

#6 drivlikejehu

drivlikejehu

    Senior Member

  • Members
  • 534 posts

Posted 02 May 2013 - 03:09 PM

Assertions are often attributed to baseball analysts despite the fact no one has ever made such claims. Nowhere has Bill James, Tom Tango, or any prominent 'sabermetrician' ever said that baseball could be completely understood by reduction to statistics.

In fact, much work has to do explicitly with uncertainty. That is the case with BABIP (for pitchers - hitters have much more control). The problem with BABIP is that it takes a long time to have a reliable sample of actual events. Much longer than most pitchers' entire career.

Regardless, any limitations to the use of statistics do nothing to increase the validity of other means of assessment (e.g., the 'eye test').

#7 cmathewson

cmathewson

    Senior Member

  • Members
  • 2,273 posts

Posted 02 May 2013 - 04:11 PM

We often say, "His numbers look good, but he's had a lot of hard-hit outs right at people." Or, "He's hitting better than his numbers; he seems to hit into a lt of loud outs."

BABIP helps to validate these impressions.

It also helps you predict if a guy is likely to continue as he is or get better. I normally use it in conjunction with LD% for hitters and GB% for pitchers. The thinking is, if a hitter has a low BABIP and a high LD%, he's due for some better luck. If a pitcher has a high GB% and a high BABIP, he's given up a lot of seeing-eye hits and, odds are, he'll do better in the future.
"If you'da been thinkin' you wouldn't 'a thought that.."

#8 adjacent

adjacent

    Senior Member

  • Members
  • 176 posts

Posted 02 May 2013 - 04:37 PM

Interesting discussion. Related to this and the validity of parameters like BABIP, FIP, etc. Anybody knows if somebody has studied the error and variance in those parameters? For example, I have some scientific background, and in the lab, if you make a measurement, repeat it 3 times and the average is 50, that number is much more reliable if the individual measurements were 50.1, 49.9 and 50.0 than if they were 40, 50 and 60. And there are multiple procedures to calculate standard deviations, significance of differences and so on. I don't know how those analysis could be applied to baseball (for example, to calculate if the FIP of pitcher is 4.0 and of another is 3.7, if that difference is significant). I wonder if anybody has analyzed that.

#9 drivlikejehu

drivlikejehu

    Senior Member

  • Members
  • 534 posts

Posted 02 May 2013 - 05:47 PM

Using FIP as an example: it's not an experimental result. It's just a formula that incorporates events that have already transpired (e.g., strikeouts). It can't have 'error' in a statistical sense.

However, the degree to which FIP (or a similar statistic) is useful depends on the properties of the underlying events. One inning pitched tells you basically nothing regardless of how you want to express the results (ERA, FIP, xFIP, etc.). After 5000 innings, you can safely use ERA without worrying about regression issues.

Metrics that regress more heavily are better for small samples, because of the tendency for regression to the mean. In other words, xFIP is preferable to FIP for a 20 inning sample, because the former assumes pitchers have little control over HR/FB rates. But in reality, pitchers do have some control over that, and after a couple seasons it's better to look at FIP.

Stats writers have gone to very, very great lengths breaking everything down. Just have to look for it.

#10 adjacent

adjacent

    Senior Member

  • Members
  • 176 posts

Posted 02 May 2013 - 08:15 PM

I know that FIP is not an experimental result, but is a coefficient calculated from results of at bats, that, obviously, you can take as results. So, even FIP does not carry an error in itself, it carries the propagated error of the ratio of experimental results. I am not a professional statistician, but I am suspecting that there could be adistribution model that it is appropiate to treat those results of at bats.

#11 old nurse

old nurse

    Member

  • Members
  • 1,860 posts

Posted 02 May 2013 - 09:41 PM

Ervin Santana and Weaver had the same BABIP while pitching for the Angels last year. Their ERA were over 2 runs apart. I don't think BAIP would be a sole measure to use to define anything for pitchers.

#12 Kobs

Kobs

    Senior Member

  • Members
  • 301 posts

Posted 02 May 2013 - 10:41 PM

I can't see a reason to make a semantic distinction between luck and random variation.

#13 Kobs

Kobs

    Senior Member

  • Members
  • 301 posts

Posted 02 May 2013 - 10:45 PM

Ervin Santana and Weaver had the same BABIP while pitching for the Angels last year. Their ERA were over 2 runs apart. I don't think BAIP would be a sole measure to use to define anything for pitchers.


I have no idea who or what you are arguing against.

#14 PseudoSABR

PseudoSABR

    Twins News Team

  • Twins News Team
  • 2,014 posts

Posted 03 May 2013 - 01:25 AM

I have no idea who or what you are arguing against.

He's probably just making a point, irrespective of the lack of argument. Heaven forbid.

#15 old nurse

old nurse

    Member

  • Members
  • 1,860 posts

Posted 03 May 2013 - 05:24 AM

BABIP measures something. What use it is in terms of defining talent of a pitcher is not known to me. Through a season with more or less the same team behind them Santana and Weaver had the same BIAP. Also they had the same k/9 rate. A yardstick to measure the talent is ERA, as they played roughly the same amount of games in each park. They had very different ERA. 2.81 versus 5.16 The end product of their pitching was very different. A HR/9 difference of .95 versus 1.95 might be the answer to the different outcomes but not entirely

For a hitter all that BABIP measures is what happens when you do not strike out or hit a home run. I do not know the usefulness of that. It is not a measure of randomness because over time hitters would have the same BABIP. If it were luck then Mauer has no skill with the bat and should be buying lottery tickets as a proven lucky player

Edited by old nurse, 03 May 2013 - 05:31 AM.


#16 old nurse

old nurse

    Member

  • Members
  • 1,860 posts

Posted 03 May 2013 - 05:33 AM

Newsflash, Mauer did win the lottery in March 2010

#17 Sssuperdave

Sssuperdave

    Member

  • Members
  • 80 posts

Posted 03 May 2013 - 08:29 AM

This thread has been very educational - thanks everyone! Now.... let me see if I can do a little novice analysis of Joe Mauer using what I've learned in this thread. Feel free to laugh at this...

First, here are some stats from 4 selected years - 2 bad (2007 and 2011) and 2 good (2009 and 2012). My goal is to see if any of these years had some "random variation" that helped or hurt him. I looked up BABIB, AVG, and LD%, and added OPS, HR, and HR/FB% since HR is excluded from BABIP and I would think those stats would help tell the story.

[TABLE="class: grid, width: 500"]
[TR]
[TD]Year
[/TD]
[TD]BABIP
[/TD]
[TD]AVG
[/TD]
[TD]LD%
[/TD]
[TD]OPS
[/TD]
[TD]HR
[/TD]
[TD]HR/FB %
[/TD]
[/TR]
[TR]
[TD]2007
[/TD]
[TD].319
[/TD]
[TD].293
[/TD]
[TD]17
[/TD]
[TD].808
[/TD]
[TD]7
[/TD]
[TD]4.5
[/TD]
[/TR]
[TR]
[TD]2009
[/TD]
[TD].373
[/TD]
[TD].365
[/TD]
[TD]21
[/TD]
[TD]1.031
[/TD]
[TD]28
[/TD]
[TD]12.1
[/TD]
[/TR]
[TR]
[TD]2011
[/TD]
[TD].319
[/TD]
[TD].287
[/TD]
[TD]21
[/TD]
[TD].729
[/TD]
[TD]3
[/TD]
[TD]2.7
[/TD]
[/TR]
[TR]
[TD]2012
[/TD]
[TD].364
[/TD]
[TD].319
[/TD]
[TD]22
[/TD]
[TD].861
[/TD]
[TD]10
[/TD]
[TD]4.8
[/TD]
[/TR]
[TR]
[TD]Career
[/TD]
[TD].345
[/TD]
[TD].322
[/TD]
[TD]21
[/TD]
[TD].87
[/TD]
[TD]96
[/TD]
[TD]5.7
[/TD]
[/TR]
[/TABLE]

Here goes... 2007 and 2011 were two of Mauer's worst years, and both years had a BABIP that was 26 points below his yearly average. 2007 also had one of his lowest LD%'s, but 2011 had an average LD%, suggesting perhaps he was the victim of some negative "random variation." However, I'm sure we all remember the disaster that was 2011, and that he didn't seem like himself. I think the 2.7 HR/FB% gives that away. So, perhaps he was still hitting the ball hard (hence the LD% equal to career average), but not as hard as usual (the lower HR/FB%). This makes me think that the low BABIP in each of these years was not fueled by random variation, but by truly worse hitting. Of course, he had significant injuries in both of these years which is probably the real reason.

Now for the good years... as we all know 2009 was incredible. His BABIP was 28 points above his career average, yet his LD% was at career average, and the same as the awful 2011. However, the number that leaps off the page and tells the story for me is his 12.1 HR/FB%, which is more than double any other year he has had (except his rookie year with only 122 PAs). I think this means he was clearly stronger, "seeing the ball better," etc. suggesting that the BABIP was driven by truly better hitting.

Regarding 2012 - I called it a good year, but it was really average for Mauer (his average year is quite good!), Most numbers are right at career averages, but his BABIP was 19 points higher than his career average. Does this make anyone concerned that he wasn't quite as good as he seemed last year???

Okay, there's my first try. What do you think - did I go off the reservation?

#18 Willihammer

Willihammer

    ice cream correspondent

  • Members
  • 3,328 posts
  • LocationSaint Paul

Posted 03 May 2013 - 11:48 AM

Parker did a writeup a while back about how Mauer benefited from having a runner on base. Specifically, he had a high BABIP when runners were on base and he pulled balls, the theory being that the defense holding runners on 1st opened a hole on the right side for the ground balls he rolls over on.

And in fact Mauer had 148 PAs in 2012 where he batted with a runner on 1st and l/t 2 outs, compared to only 56 in 2011. Maybe that explains some of the uptick in BABIP.

#19 ashburyjohn

ashburyjohn

    Twins Daily Moderator

  • Twins Mods
  • 5,857 posts
  • LocationLake Tahoe, Nevada

Posted 03 May 2013 - 12:11 PM

I can't see a reason to make a semantic distinction between luck and random variation.


Luck is when random variation works against you. Intuitive skill is when it works in your favor.

/ or, bearing down, or gettin' after it

#20 kab21

kab21

    Senior Member

  • Members
  • 2,637 posts

Posted 04 May 2013 - 04:35 AM

For a hitter all that BABIP measures is what happens when you do not strike out or hit a home run. I do not know the usefulness of that. It is not a measure of randomness because over time hitters would have the same BABIP. If it were luck then Mauer has no skill with the bat and should be buying lottery tickets as a proven lucky player


Nobody has ever said that batters will all hit for the same BAPIP. BAPIP is simply a measure of BA on balls in play. For hitters I definitely don't consider a player lucky if he has had a high BAPIP for his entire career.

Further The Hardball Times has developed an xBAPIP (expected) calculator that uses a lot of different variables to determine an xBAPIP. A few of those include LD% (a really high BA on these), HR/FB% (hit the ball harder = higher BAPIP), speed (beat out IF singles), IF popups (always outs), FB% (lower BAPIP than GB's) and several others. I'm actually too lazy to use since I find 3+ seasons of BAPIP data to give me a pretty good idea if a hitter is lucky/unlucky.

#21 Jim H

Jim H

    Senior Member

  • Members
  • 439 posts

Posted 04 May 2013 - 09:03 AM

{Parker did a writeup a while back about how Mauer benefited from having a runner on base. Specifically, he had a high BABIP when runners were on base and he pulled balls, the theory being that the defense holding runners on 1st opened a hole on the right side for the ground balls he rolls over on. } QUOTE FROM WILLIHAMMER

This is one of the reasons that I prefer Mauer in the 3rd spot in the order, since he will have more batters on base for him in that spot rather than the 2nd spot over the course of the season. Sorry, this has nothing to do with this thread.

#22 Kobs

Kobs

    Senior Member

  • Members
  • 301 posts

Posted 04 May 2013 - 01:07 PM

He's probably just making a point, irrespective of the lack of argument. Heaven forbid.


Fine.

I don't believe that fielding percentage is the sole measure to define anything for hitters.

#23 BHtwins

BHtwins

    Member

  • Members
  • 79 posts

Posted 05 May 2013 - 12:04 PM

babip in play is a contrast stat. It tells you almost nothing out of context. Depending on the context it can tell you a whole lot of things.

#24 kab21

kab21

    Senior Member

  • Members
  • 2,637 posts

Posted 05 May 2013 - 08:28 PM

babip in play is a contrast stat. It tells you almost nothing out of context. Depending on the context it can tell you a whole lot of things.


bapip in play is a redundant usage.