THE BOOK--Playing The Percentages In Baseball

RHB - 2200 PA
LHB - 1000
RHP - 700
LHP - 450

It sounds like Nate’s post was lifted substantially from this chapter, especially the part about pitchers.  Their platoon ideas seem to be lagging - remember the “record vs. LHP/RHP” adjustment in the playoff odds?

Nate’s a standup guy, so I don’t think it’s fair to say “lifted”. 

***

I agree about the disaster of the playoff prediction based on the LH/RH performance.

***

Thanks for info!  It seems that Nate regresses toward the “generic” mean way too much.

Sorry, I didn’t mean “lifting” in a dishonest sense, just rehashing stuff that’s already been covered.

Thanks for clearing it up.  It’s easy to write something real quick in a blog, with an honest intent, and have it read completely differently.

Dan Fox calls to our attention (and reprints) a BP analysis of his own from 2006 on platoon splits.

http://danagonistes.blogspot.com/2007/11/of-crowds-and-splits.html

Gee, I am not usually one to complain about authors not citing other people’s work, but come on…

I have been talking about this and publishing numbers for years (Neyer’s article specifically referenced my work), and as was mentioned above, we devoted an ENTIRE section to it in The Book, whereby Andy gives the exact (well, as exact as you can get given the limitations of sample data) amount of regression for RH and LH pitchers and batters.

Nate strongly implies that he is breaking new ground here.  Nothing could be further from the truth.  Well, some things can I guess, like there were WMD’s in Iraq or waterboarding is not torture.

And I don’t think that the data tells you that “players are born with a certain platoon ratio which does not change over their lives.” It just tells you that it doesn’t jump randomly all over the place throughout their career. It is entirely possible that there is a distinct aging curve or learning curve associated with platoon splits, just as their is with batting and pitching (and defense, baserunning, etc.).  I am not sure that anyone has looked into that.  I haven’t.

And yes, he is right (and it should be obvious, but is often overlooked) that when doing projections, a player’s past and expected future percentage of same and opp handed opponents should be considered.  As Tango says, it probably doesn’t matter much what division you are in (although it could I suppose), it probably matters a little more what team you are on (but still not that much), but definitely matters if a player changes roles (such as going from a platoon player to full-time or vice versa, or a pitcher going from closer to generic reliever or vice versa, or even starter to reliever or reliever to starter). Speaking of that, I don’t know whether Tango or anyone else controlled for handedness of opponents when looking at how pitchers do in relief roles versus starting roles, but it could be that part of the difference is that (that relievers have the platoon advantage more than starters do, on the average - and I don’t know whether that is even true or not).

I am surprised that Nate would say something (like one division has a lot more lefties than another) without checking it out.  Also, this is being a little picky, but he says “good” left-handed pitching in reference to the AL Central.  Just to be clear, whether the left-handed pitching is good or bad has nothing to do with anything.  Now, whether the left-handed pitching is “tough” (as in a “tough lefty") or not is another story, tough refering to the true platoon ratio of the pitcher, the larger that true ratio, the “tougher” that pitcher is, like a sidearmer or pitcher with a great slider (pitchers with great sliders that they use a lot tend to have large platoon splits).

Historically, the average LHP = RHP.  This is what we would expect, if everything was lined up.  For example, when Blacks first came into MLB, the average Black was better than the average White.  This is easy to do if your Blacks consisted of potential MVPs, while your Whites consisted of all types of players (MVP and scrubs).  Equilibrium is reached when one subset equals another.

Now, I published more recent numbers for LHP and RHP, and there is a tiny bit of gap (which could be due to selective sampling).  So, on a league level, it still wouldn’t matter.

On a player level, if he faces predominantly RJ, Santana, et al, then he will end up showing a wider gap in his platoon splits.  So, you not only want to control for handedness in your forecast, but also the quality of opponent. 

***

Nate’s entry is a blog entry, so I let those slide in terms of “research”.  It would have been better if he didn’t bother to cite Rob Neyer, since that opens up the door to criticizing why he wouldn’t cite MGL, Andy, or Dan Fox (and whoever else did some research on the matter).

At least Nate is one to take his lumps (justified or not) and be fair about it in future posts.  Plenty of other bloggers would take a Karl Rove position that the best defense is a good offense.

I agree that in a blog context (and othre ones), citations and the like are definitely not necessary.  But, as I said, this entry had the tone of, “Here is something I just thought of...”

When I am writing something informal or even formal, as in an online article, I usually AT LEAST state something like, “I am not aware of any research in this area,” or “There has been some or considerable (etc.) research in this area.” I try to at least put that in.

Let’s be fair though. Unless one is writing an academic paper, citations are not madatory.  If someone tries to take credit for something as original (intentionally or not), or implies it, that is another story.  As I said, personally I don’t care much if someone cites specific prior work or not.  But this particular blog entry was bad.