Platoon Differential

[DrungoHazewood]

From a current article on Bill James Online entitled " Xavier University Business School Speech":

One of the things that we actually do in my field is to try to straighten out the misconceptions that people have, based on misunderstanding the statistics.    One of the largest of these, actually, has to do with the platoon differential.   The platoon differential, for those of you who are not baseball fans, has to do with the fact that left-handed hitters hit better against right-handed pitchers, and right-handed hitters hit better against left-handed pitchers.

            But what almost nobody understands, inside of baseball or outside of it, is that this is not an individual variable.   It’s a condition of the game.   It applies to everybody.

            But we MEASURE what hitters hit against left-handed and right-handed hitters based on individual hitters, year by year.   Because we are dealing with a relatively small advantage and relatively small data groups, the data for individuals is all over the map.   A right-handed hitter may hit .260 against right-handed pitchers, but .420 against left-handers.  That’s called the platoon differential.   This SEEMS to be tremendously important, and, because we measure it on an individual basis, people ASSUME that it is an individual trait.

            But it actually isn’t; it’s a condition of the game that applies to everybody.   There isn’t actually any such thing as a hitter who hits .420 against left-handed pitchers, or a hitter who has a 160-point platoon differential.    That’s just a statistical fluke, just like the guy who hits .420 on Tuesdays.   The platoon differential is basically the same for everybody.  

            But baseball people don’t understand this; they didn’t understand it 30 years ago, and they don’t understand it now, so they talk literally every day and many times every day about platoon differentials, always assuming that it is an individual variable.    Because they grow up with that understanding, and because that understanding is re-enforced every day by selective observations, it takes a long, long time to get people to understand that there is no such ability.    But we will eventually succeed.     A hundred years from now, people will generally understand that there isn’t any such ability, and never was.    

            I should be careful about speaking in absolutes.   There actually is at least one major league player who has an abnormal platoon differential.    Adam Jones of the Baltimore Orioles is a right-handed hitter who actually does hit better against right-handed pitchers than left-handers.   But he is probably the only one. 

[sportsfan8703]

I think this is true because Jones has trouble with the low and away pitch.  If lefties can spot that fastball near the outside corner than he is in big trouble.  

Another thing is that when an AL East team gets a LH SP they are usually beasts.  Like Sale, CC, Price, Pomeranz, Happ.  These guys are #1-2 SP's.   It's not like we're in the AL West and seeing random LH bum SP's.  

[Frobby]

Instinctively this makes no sense to me.    It seems natural to me that some LH batters would have more trouble hanging in against LH pitchers than others, depending on their stance and pitch recognition skills.    

[DrungoHazewood]

19 minutes ago, RZNJ said:

Agreed. It defies logic.

Doesn't it really defy preconceived notions that help you define what you consider logical?

[DrungoHazewood]

100 years ago logic dictated that a good hitter didn't care which hand a pitcher threw with. He was just as good either way.

[Phantom]

41 minutes ago, Frobby said:

Instinctively this makes no sense to me.    It seems natural to me that some LH batters would have more trouble hanging in against LH pitchers than others, depending on their stance and pitch recognition skills.    

I'm not 100% sure, because this is such a small out of context quote, but knowing James I would guess he doesn't mean that no one has an abnormal ability to hit same handed pitchers at a better clip than others simply that we lack the ability to measure and conclude it accurately. Hell, AJ himself had a season where he hit almost 100 points better against lefties than righties.

 

Think about his Tuesday example. It makes perfect sense that some players might play better on Tuesdays for a variety of reasons, right? But how many years of data would you need to be sure that a player is actually better on Tuesdays and its not just noise?

 

We fall into this trap with guys who start the season either fast or slow. A guy will start 4 or 5 seasons off fast or slow and we will use that as proof his is a fast or slow starter which is roughly equivalent to flipping tails 5 times in a row and declaring the coin weighted. We may be right, but the vast majority of the time, it's just a fluke.

[Moshagge3]

1 hour ago, sportsfan8703 said:

I think this is true because Jones has trouble with the low and away pitch.  If lefties can spot that fastball near the outside corner than he is in big trouble.  

Another thing is that when an AL East team gets a LH SP they are usually beasts.  Like Sale, CC, Price, Pomeranz, Happ.  These guys are #1-2 SP's.   It's not like we're in the AL West and seeing random LH bum SP's.  

Random LH bum SP's are the ones who are murder on the Orioles! I still have nightmares about Jason Vargas!

[Phantom]

12 minutes ago, RZNJ said:

Seth Smith and every LH hitter have the same platoon differential?

Where's the evidence to back up the theory?  They are just telling us to ignore the evidence and accept the theory.

What you call evidence they call small noisy sample.

 

I don't think he is saying that ever hitter has the same platoon differential. I think he is saying that as far as we know nearly ever player has the same platoon differential.

[DrungoHazewood]

24 minutes ago, RZNJ said:

Seth Smith and every LH hitter have the same platoon differential?

Where's the evidence to back up the theory?  They are just telling us to ignore the evidence and accept the theory.

You need to interpret the data correctly. You have a bunch of 50, 100, 200 at bat trials. You need to heavily regress them to get to an estimate of true talent. 

You need thousands of PAs to get a significant sample, and you'll never get there in most careers. So we've just been assuming this large individual skill at hitting lefties or righties, mostly based on SSS information.

[DrungoHazewood]

16 minutes ago, Phantom said:

What you call evidence they call small noisy sample.

 

I don't think he is saying that ever hitter has the same platoon differential. I think he is saying that as far as we know nearly ever player has the same platoon differential.

And interpret "same" as "close enough to call same."

[DrungoHazewood]

This is a case where simulation would help. Program a sim where everyone has a fixed platoon differential around their overall talent. The result probably looks a lot like real life just based on small sample randomness.

[NCRaven]

My only problem is that Bill James did not explain how is understanding of platoon differential can be used to change the way we play the game.  Okay, give James theory, how should managers change the way that they manage the game?  Most managers try to optimize their lineups by playing lefty hitters against righty pitchers as often as possible.  And vice versa.  Should that change?  I guess I'm asking, what's his point?

[DrungoHazewood]

15 minutes ago, RZNJ said:

I get it.  I think.  They say there isn't enough evidence.  However, what is the evidence to say that every hitter has the same platoon differential?

Quick example. Ran some RH v RH Data. Big Sample, careers of 7000+ PAs since WWII. 141 players. 115 of the 141 had a RH v RH OPS difference from their total between .010 and .050.   More than half between .015 and .035.

No one over .075 no one under -0.010.

The bigger the sample the more platoon splits converge on a set amount.

What Bill is really saying is there is a much smaller individual variation than is commonly thought.

[DrungoHazewood]

11 minutes ago, NCRaven said:

My only problem is that Bill James did not explain how is understanding of platoon differential can be used to change the way we play the game.  Okay, give James theory, how should managers change the way that they manage the game?  Most managers try to optimize their lineups by playing lefty hitters against righty pitchers as often as possible.  And vice versa.  Should that change?  I guess I'm asking, what's his point?

Don't assume John Lowenstein would OPS 1.000 vs righties and .500 against lefties regularly.

[Frobby]

45 minutes ago, Phantom said:

What you call evidence they call small noisy sample.

 

I don't think he is saying that ever hitter has the same platoon differential. I think he is saying that as far as we know nearly ever player has the same platoon differential.

I think this is a situation where you have to both look at numbers and observe the player.    There's no question that players' splits vary a lot from year to year, but sometimes you can just see that a particular batter is super uncomfortable.