BABIP

[TheCrew07]

This has been eating at me for well over a year, and with all the BABIP discussion lately I can't get the thought out of my head.

What logical reason is there that HRs are not included in BABIP? It's a basehit any way you slice it. Why should a batter be penalized because he actually hit the ball so well it ended up out of play? I know there's other rough ways of doing BABIP that involve simple calculations like LD% + .120 and so on, but how can a hit be totally removed from a Batting Average formula? I've bounced around read quite a bit about BABIP and most people treat it as is, without exactly describing why it is the way it is. If a sac fly is included, why wouldn't a home run be?

Educate me please.

[TooLiveBrew]

I can't educate you, but I can guess that the wonderfully-hit screamers that end up in gloves (Dillon's PH rip tonight is an example) more than cancel out the HRs. Just a guess.

[rluzinski]

People usually look at BABIP to see how lucky or unlucky a player has been finding holes in a small sample. There is some skill involved but Jones having a .414+ BABIP suggests that he's probably gotten a little lucky with balls finding holes so far this year. Of course, you can just look at his BA to know he's gotten a little lucky (when I say lucky, I mean that his performance has been better than his actual skill).

 

I don't really look at a batter's BABIP, though.

[TheCrew07]

Yes but the pitchers have HRs removed totally from the equation... like they never happened. A hit is a hit IMO, and isn't a HR basically the ultimate LD?

[logan82]

The defense doesn't have a chance to field a HR and like Russ said BABIP is used to measure how "lucky" a player has been at finding holes in the defense.

[TheCrew07]

So if you give up a triple you're unlucky, but if you give up a home run you stink? I see what you guys are saying, but the concept and the math is flawed if that's the true answer as to why a sac fly would count by a home run doesn't... A home run is still a ball in play.

[logan82]

It is a measure of what a defense did behind the pitcher. A defense, except on rare occasions, can not do anything about HR.

[TheCrew07]

logan believe me that I get what you're saying, but there are more balls that defense doesn't have an impact on... Infield Fly Rule, Ground Rule Doubles... I realize it's not perfect, no metric is, but a pitcher could theoretically have a lower BABIP than actual average against... they get credit for a sac fly being caught which lowers the average, and the HRs are totally removed, the simple truth here is that the more HRs you give up, the better your BABIP is... you can be a totally horrible pitcher who throws nothing but gopher balls and have a decent BABIP. Do you see what I'm saying?

[logan82]

I understand how BABAIP is figured and how it works. I have no idea what you are trying to get at. I think you are trying to use BABIP for something it was never meant to be used for.

[endaround]

BABIP isn't a stat to rate how a pitcher throws. It can explain why a pitcher with bad stuff has an ERA thats low or why a pitcher with good stuff has a high ERA. Thats all. The point is that pitchers have litle control over whether balls in play become hits or outs after you account for the type of pitcher (GB vs FB).

[TheCrew07]

I know it's a comparison stat. I guess I'll try and explain my issue one last time. Say a pitcher doesn't strike many guys out but gives up more than his fair share of HRs... then you factor in sacrifices as a positive result for the pitcher... It's entirely possible that a pitcher's BABIP would actually be lower than his BA against. His BA against could be .300, but his BABIP could be .280... both are really poor, but the batting average of balls in play should be always be higher regardless because even the worst pitcher is going to strike out at least one batter. If you're trying to measure if someone is "unlucky" (which I think is bunk anyway, the pitcher has control over the velocity, location, and pitch he throws), how can the result of all balls in play not be considered?

 

Again, I understand that you're theoretically trying to measure the defense, but usually a league average is used, so you compare a pitcher against the league average. For example, statements suchs as "His ERA will regress to the mean because his BABIP is too high" and so on are very common. The problem is that BABIP is not a true representation of all balls in play and it stands to reason that the harder a ball is hit, the less likely it is to be caught, which is why I personally like an LD% calculation better.

 

It doesn't matter what metric you use, good players will do well, bad players will do poorly (traditional stats such as BA, OBP, ERA all will tell you pretty much the same thing as OPS, OPS+, FIP. ERA+ for the playes on either extreme). Doesn't it stand to reason that we're really trying to determine the relative value of players around the mythical average skill level? It seems to me that whomever came up with BABIP is cherry picking stats that favor the HR pitcher. Assuming that all things are equal, the pitcher that gives up more HRs (likely more runs) would do better BABIP wise, and that's the concept I take issue with. To me, all balls in play regardless of the outcome should be considered, from a pitching standpoint, the HR is worst possible outcome of a single AB.

[logan82]

A HR is not in play unless it is an in the park HR.

[rluzinski]

It seems to me that whomever came up with BABIP is cherry picking stats that favor the HR pitcher.

 

Not at all. A pitcher's BABIP for a season tells you almost nothing about a pitcher's talents to begin. That's not because the pitcher has no control but because it's just small component. You need more than 200 innings with of balls in play to try and sort out the pitcher's BABIP skill.

[Brawndo the Thirst Mutilat]

It doesn't matter what metric you use, good players will do well, bad players will do poorly (traditional stats such as BA, OBP, ERA all will tell you pretty much the same thing as OPS, OPS+, FIP. ERA+ for the playes on either extreme).

 

The entire DIPS notion was developed because Voros McCracken noticed that BABIP does not follow the rule you describe. It isn't that there isn't any variation in single season BABIP results, it's that it has almost no stability for a particular pitcher from one year to the next. So one year, the best BABIP might be posted by Pedro Martinez, Randy Johnson, Jake Peavy, and Brandon Webb. Then, the following year, those guys are all at the bottom of the league and the very best BABIP come from Seth McClung, Glendon Rusch, Jeff Weaver, and Jaime Navarro. So for BABIP, rather than "good players do well, bad players do poorly", it looked like it was totally random who did well and who did poorly.

 

What conclusion would you draw? Doesn't it look like BABIP has nothing whatsoever to do with pitcher skill and is entirely a function of luck?

 

Homers were never counted because we knew from the beginning that allowing a lot of HR is not a function of luck. HR allowed is not quite as stable as K and BB rates for pitchers, but there is a real and obvious HR suppression skill that emerges if you track the data from year to year. BABIP was invented to answer a specific research question - When the batter makes contact with the ball and the fielders have a chance to turn it into an out, how much does the pitcher have to do with the outcome? Counting homers was never a possibility.

 

After several years of arguing, the general consensus is now that there is some genuine BABIP-suppression skill that some pitchers have and some do not. But, unless you fall into one of a couple of pitcher types where everybody's BABIP tends to be lower (e.g. knuckleballers), the signal to noise ratio on this skill is so low (93% of year to year variation is luck vs. 7% skill according to BP's Nate Silver) that you'd be a fool to begin to presume to be able to separate skill from luck with a sample size of fewer than 1000 IP or so. Better to just forecast a BABIP allowed exactly at the league average, which is usually between .290 and .300.

 

Jeff Sackmann posted a nice summary about how BABIP is useful and how it isn't, when it's luck and when it isn't, how batters are different from pitchers, etc. Those interested can find it here:

 

http://www.brewcrewball.com/2008/4/22/447297/all-about-babip

[TooLiveBrew]

I know it's pretty easy to find a given player's LD%, and his BABIP... but I wonder if, instead of just looking at BABIP, it might be more useful to look at BABIP on line-drives (which I'll call LDBABIP). Is there anywhere that provides either a LDBABIP rate, or the types of splits necessary to determine it?

 

My thought is that a player isn't necessarily unlucky for having a low BABIP, but he probably would be unlucky if he had a low LDBABIP.

[logan82]

THT has a nice xBABIP calculator. It downloads as an Excel file. I am sure is isn't perfect but it is somewhere to start.

 

I have used it for a few guys this winter. Off the top of my head here are a few.

Lopez .330

Gomez .300(ish don't remember exactly)

McGehee .305

[cwolf2012]

"It seems to me that whomever came up with BABIP is cherry picking stats that favor the HR pitcher."

 

I completely disagree with this. If you're removing HRs from the equation, you're lowering BABIP vs BA against. Since most people regress BABIP to the mean, it would suggest that the HR pitcher is actually overperforming with a .280 BABIP!

 

Either way, the stat is not used to measure pitcher skill, it's a means to regress other stats such as ERA. If anyone looks at a pitcher with a big ERA for his career because of HRs and a low BABIP, nobody is going to say that he's a good pitcher. BABIP is not something that is a measure of worth, it's a tool for future projections.