THE BOOK--Playing The Percentages In Baseball

I have a headache now..

In game theory situations, most people have predictable patterns of behavior.  Some people favor “go with what worked before” and some favor “change it up” but there are definite preferences.  The trick is figuring out who is who, which I suppose is just an engineering problem.

Truly, you have a dizzying intellect.

MLB : Princess Bride :: Greg Maddux : Westley

Something tells me the following exchange never took place:

Molina: Fetch me an inside changeup.
Maddux: As you wish.

The correct approach would be some sort of randomized strategy in all situations, regardless of the perceived knowledge of the opponent.

Before someone gets on Andy’s case, he means “randomized strategy around a likely true mean”.

OK, I just read the beginning of Eric’s article, and I have not read the ESPN article he sites (I don’t think), but I take umbrage with statements like these:

If an equivalent to Kasparov exists in the world of baseball it would have to be Greg Maddux.

Teammates remarked to Keown that Maddux can watch a seemingly run of the mill swing and deduce, with great accuracy, what will occur over the course of the next few pitches.  This specifically comes in handy with regards to foul balls headed towards their dugout.

Maddux was a great pitcher at least because he had an unbelievable ability to locate pitches and because he had good stuff (a good tailing or sinking fastball mostly).

And BTW, I have closely watched him pitch for his entire career, and I have NEVER noticed anything in particular about the way he pitches, as far as the “mental” side or game theory side of the game goes.  I think that if you have virtually any pitcher Maddux’ physical abilities to throw an locate a baseball, they would be quite successful.  That does NOT mean, though, that he does not ass SOMETHING to his overall talent, but virtue of some kind of above-average pitching “smarts.”

However, I have NEVER, EVER, EVER seen any evidence that he was or is a “smarter” pitcher than anyone else.

Sometimes someone gets a reputation that he deserves and sometimes a person gets a reputation that he does not deserve.

We already know that 99% of the stuff people say about baseball and its players are crap.

Why should we believe that Maddux has any more or less ability for “smart pitching” than anyone else?  because he “looks” like he is smart?

Again, it certainly could be true, all or some of the things that people say about Maddux, but I have NEVER seen ANY evidence to back it up.

So let’s stop saying things about people just because that is what we have heard from OTHER people.  That ain’t science and that ain’t the truth.  And it is not what I am about.

Now that I said that, I will read this article and any links it has, and see if indeed there IS evidence that Maddux is a pitching “genius,” whatever that even means.

I don’t know that Maddux is some sort of pitching “genius” or not, but it appears that his teammates and pitching coaches often think he is. I remember a conversation an announcer had with Leo Mazzone back when he was the Braves’ pitching coach and Maddux was still with the team. Mazzone said that the team’s advance scouts would write up reports about how an opponent’s batters should be approached. If Maddux disageed with the reports, Mazzone said that they always went with Maddux’s suggestions because they felt that he knew better.

Then there’s this item from ESPN the Magazine that I read on the Baseball Toaster site.

When Brad Penny and Maddux were teammates on the Dodgers, during the last two months of 2006, they had a conversation one day that led Penny to reach a stunning conclusion: This guy knows my stuff better than I do. It was eerie, really, how easily Maddux dissected Penny’s repertoire and suggested ways to maximize it. Penny, figuring he’d take advantage of the situation, asked Maddux to call a game for him against the Cubs. And so, on the night of Sept. 13, Penny glanced into the dugout before every delivery and found Maddux, who signaled the next pitch by looking toward different parts of the ballpark. Penny threw seven scoreless innings with no walks and beat the Cubs 6-0. “Maddux probably won’t tell you that story,” Penny says. He’s right.

Of course, such anecdotes relate to reputation, not to any kind of “genius”. However, it indicates that there is something about Maddux’s approach to pitching that impresses his peers.

I’m sorry, I got nothing - absolutely nothing - from this article.  There is simply no way to garner much from a few PA’s between any one pitcher and one opponent.  There are too many variables for why a pitcher would do something and why the batter would do something.  Not to mention, you NEVER know the pitcher’s intent as far as the location of the pitch, especially if you are not watching the game (if you are watching, you can at least see the catcher’s target).

There are a few things that I think are factually problematic in the article:

Maddux threw an 0-2 fastball towards the inside corner. It came in a bit too high and closer to the plate than he would have liked and Molina lined a double.

Unless he saw where the catcher was set up (maybe he did), how did he know that the pitch came in higher than he would have liked?  I’ll grant him that if the pitch was near the center, that it probably caught more of the plate than Maddux would have liked, but how did Eric know that he was not trying to throw one up and in?

Molina adjusted his approach as he is generally not one to swing at the first pitch.

I always thought that Molina was a free-swinger?  What is his first pitch swing % versus the average batter?  And of course, if a batter is a free swinger, then the pitcher will know that and throw more pitches outside of the strike zone, such that the swing percentage of a free swinger and one who is more patient may actually be around the same!

If you want to know if someone is a free swinger or not, you have to compare him to the average batter, for pitches in the same location.

The common creed with Maddux is that, since he is deadly accurate and despises wasting pitches, the first pitch you see may be the best one. He abides by the “best pitch is strike one” motto and so by taking the first pitch a batter is usually seceding the ability to be ahead in the count.

That is true that Maddux does not like to throw waste pitches when ahead in the count (I don’t know why, but I’m sure he has an answer).  Of course, if you have more control than other pitchers, it will LOOK LIKE you are throwing fewer waste pitches, even though you may NOT be (if you and other pitchers are throwing to the EXACT same spot on an 0-2 count, the other pitchers will miss that spot away from the strike zone more often and to a greater degree than you will, and it will LOOK like they are throwing more waste pitches).

Anyway, wouldn’t that mean that Maddux first pitch is NOT necessarily the best pitch to hit?  Isn’t it the exact opposite of what Eric is saying?  If you have a pitcher who has great stuff, you want to swing early before he has a chance to get ahead and then put you away.  With Maddux, you are much more likely to get a hittable pitch when he is ahead in the count.  Yes, Maddux probably throws more first strikes (I assume), but that is just because he throws more strikes on ALL pitches and counts.  But that does NOT mean that you should swing earlier with Maddux than with other pitchers.  You are going to swing more on ALL counts because he throws more strikes.  And since he presumably does throw fewer waste pitches, you are also going to swing particularly often when you are behind in the count. (With some pitchers, you can almost take on an 0-2 count.)

So I don’t get what Eric is saying and I don’t think it is true.

Eric throws out the term “adjusts” rather gratuitously.  I don’t even know what that means.  All pitchers randomize their pitches. They MUST do that.  When a batter hits a pitch hard one time and the pitcher throws him a different pitch the next time in the same situation, that is NOT necessarily an “adjustment”.  It is mostly the pitcher randomly mixing up his pitches.  And as far as batters “adjusting,” well I definitely don’t even know what that means.

I hate when commentators use that word, and I did not like the use of that word in this article either.

I couldn’t even get through this article to tell you the truth.

I mean it has a lot of nice informational stuff, but I just don’t see how it helps us at all to understand how Maddux pitches (or Molina bats).  At all. With all due respect to Mr. Seidman, who is s good writer and researcher.

Eric also says somewhere (I can’t find it right now) that “Molina is patient on balls outside of the strike zone.” Earlier, he said that he swings at those pitches 30-33% of the time, or something like that. I have no idea if that is patient or not, without knowing the average swing rate for all batters, on those same pitches (again, it may be that Molina’s pitches out of the zone are MORE out of the zone, BECAUSE is IS a free swinger, and that is why he does not swing at that many of them, if that makes any sense).

And I DID read that ESPN article about Maddux.  It was a fluff piece, disguised as analysis.

So, I still see NO evidence that Maddux is any different than any other pitcher in terms of his mental game.  Let’s please start with the premise that he is no different than any other pitcher, mentally, and if anyone finds evidence indicating otherwise, I would be happy to see it.

MGL, that’s a very fair analysis and I appreciate you taking the time to (try to, at least ) read it. 

My goal wasn’t necessarily to garner anything concrete or to be able to say without a doubt that Pitcher or Batter was better off due to knowing where said Pitcher or said Batter succeeded in the past.

Rather, my goal was to explore one area that Pitch F/X data may be able to help shed light on in the future.

Something I mentioned in the comments section, that I realized after writing the article, is that it would be MUCH better if I could provide videos of each at-bat.  That way we could watch each AB of, say, Maddux v. Molina and compare the PFX results with what we actually see.

That’s going to my goal for the next piece like this I attempt to do.  With the video we can see, as you mentioned as did I in the comments, if someone misses his target, or perhaps if there is a noticeable difference in body language.  Ryan Howard, for instance, clearly looks lost at the plate and you don’t need to be an analyst to deduce he has no idea what’s coming most of the time.

I’ll be the first to admit when I make mistakes or blunders and in retrospect I definitely would have liked to provide comparisons between Molina and the average.  That’s something I’m going to get on now to somewhat amend in its current form.

To see how Maddux pitches or gain any insight I think we would run into small sample size issues across the board.  We would have to control for runners on base, what specific base, what specific count, what type of batter he’s facing, said batter’s pension for hitting certain types of pitches and/or locations, as well as Maddux’s sequencing… and then compare that to the average results of not ALL pitcher but rather pitchers with similar “stuff.”

I might be looking too far into it but I think if there was a way to control for all of that we could gain some insight.  Then again, it would take a lot of pitches and we might not be at that point until after this season, when he will likely retire, and at that point it will be of no use to any team trying to decipher what he throws.

I applaud your effort to look into the situation, and I have nothing but admiration and respect for you and ALL the guys who tackle pitch f/x data!

I think that you and anyone else who wants to look at “pitching style,” “pitching smarts” or whatever you want to call it, needs to take MUCH smaller bites.  Begin by asking very simply questions, and using large sample of data, if possible, by aggregating players.

The fact that pitchers should and do randomize their pitches, trying to figure out “how” and “why” they pitch certain batters, and what “adjustments” they may make and the batters may make, even if you appropriately define the word “adjustment”, is a difficult if not impossible task, at least without taking very small steps at first.

I know I am not making much sense or being specific, but that is because it is an enormous task to try and analyze pitchers’ and batters’ “mindsets” and the like from the pitch f/x data.  I do not disagree with Bill James when he says that it will take 10 (or 20, I forgot what he said exactly) years to figure out what to do with the pitch f/x data.

I definitely understand what you mean.  For next time, I’m going to break it up into smaller questions with the larger samples.  I could go with something as simple as lefty pitchers vs. righty hitters.  Then break it in lefty power pitchers vs. righty power hitters, lefty power pitchers vs. righty contact hitters.

To do all of the possible situations, like a righty finesse pitcher against a lefty power hitter with runners on and less than two outs we will definitely need that bulk of data… which is annoying because this data is so much fun to look at and it’s easy to want to jump the gun so to speak.

I don’t have enough info to have an oppinion if Maddux has any exceptional “pithing smarts” or if his “mental game” is stronger than that of the typical pitcher.

However I would think that the TYPE of pitcher that Maddux is (control pitcher that relies on accuracy) has a lot more use for the kind of “game theory” type thinking that is discussed here than what a power pitcher or a knuckleballer would have.

So perhaps by his success and longevity it is likely that this part of his “tools” is at least stronger than average, by “reverse” selection bias if nothing else…

One thing about Maddux is that he messes up my pitch count estimator.  He simply throws fewer pitches per batter (when I looked a couple of years ago), compared to other pitchers with the same K, BB profile.  I’ve always ascribed to something “special” to him because of that.

#15, I think that simply has to do with the fact that he REALLY does not like to throw waste pitches like MOST pitchers when ahead in the count.  Normally, I would say that is NOT a good idea, but as good as he has been throughout his career, I would hesitate to say that ANYTHING he does is not a good idea.

#14, while “game theory” applies to all pitchers, I see your point when thinking about the extremes (which is always a good idea to do).  If a pitcher has such a good fastball that it is unhittable even if the batter knows it is coming, all he has to to do is throw the fastball.  If a pitcher has several pitches that are equally good, he needs to randomize those pitches.  Not that that is all that hard to do.

Actually a pitcher needs to randomize his pitch selection around a certain mean which is determined by the game situation and the batters’ weaknesses at the various counts and game situations (and the effectiveness of those pitches in general of course).

Maybe Maddux is particularly good at that. Maybe he is not.

I think to assume that because he has the “stuff” he has (not overpowering) and because he has been so successful that he must be a super-intelligent pitcher, is a mistake.

Call me crazy or what have you, but I like to see evidence of something before I reach any conclusions, especially something which is not so obvious to me.

#16, Actually what i meant is that to a certain extent a power pitcher (but with “poor” control/accuracy) has a lot of this “game theory” built in to his pitches.

The batter can look for “tells” all he wants (either actual physical tellls or historical tendesies) but if the pitcher has very little control over the location the ball ends up at it likely won’t help him.

Also I would like to add that randomizing isn’t neccesarily the end-all as far as pitch selection strategy goes. The concept of tells and tendencies applies to batters as well as to pitchers. If you have more information you can eliminate some of the bad choices entirely.

Also, I agree that there needs to be more information to say that he is a “super-inteligent pitcher”, it’s just that I feel that it is unlikely that he is a “dumb pitcher”.