THE BOOK--Playing The Percentages In Baseball

Saturday, February 04, 2012

Bill James Baseball IQ app

I downloaded Bill James Baseball IQ onto my iphone (I don’t think it is available on droid phones, but I’m not sure). Here is the web site for the app on Acta Sports:

http://www.actasports.com/titles/bill_james_baseball_iq_app/

It is pretty cool. You can read a description and see some screen captures on the above site, but basically it allows you to see heat maps and color maps of batters and pitchers (in all combinations, counts, situations, etc.) for K zone, batted balls, pitch type, etc.

Best of all, the app is free! Seems to me that they could have charged for this one, but I know nothing about the best way to make money from apps. It also seems like they could use these graphics more often on TV broadcasts.

Anyway, give it a try and see what you think…

Thursday, January 26, 2012

AL v. NL in 2011

It is generally accepted in the sabermetric community that the AL is a better league than the NL, at least for the last several years.  This is evidenced by the fact that the AL has a large advantage in IL games, although at least some of that edge could be something other than overall “talent”, although this is not likely and several people, including myself, have found little or no inherent advantage to the AL in IL games (e.g., the NL teams do not have any DH’s, so they have to juggle their lineup in AL parks, on the other hand, in NL parks, AL teams have to sit their DH’s or juggle their lineup, perhaps putting a bad defender - their DH - in the field, the AL pitchers typically are poorer hitters than the NL pitchers, etc.).

Read More

Thursday, October 27, 2011

Batter-pitcher matchups

Colin says:

Once we have that expected value, we can also look at the TAv from that batter-pitcher matchup from all previous seasons. We can run this data from 1951 through 2011, giving us sixty years of data and over 16,000 data points to look at.

Using a technique known as ordinary least squares regression, we can see how well our expected TAv and our prior batter-pitcher matchup TAv predict future batter-pitcher matchup TAv. After controlling for whether the batter has the platoon advantage, what we find is that our log5 estimate of the outcome of a batter-pitcher matchup is 67 times more predictive than the batter’s past performance against that pitcher. Now, that’s slightly better for the batter-pitcher matchup data than we might have expected; there were on average 78 times as many PA for the log5 expectation as there were for the batter-pitcher matchup. (Since there are both batter PA and pitcher PA against used to generate the log5 expectation, I used what’s known as a harmonic mean to come up with the PA totals for the log5 expectation.)

We can conclude that one plate appearance against a specific pitcher is slightly more predictive than a plate appearance against any pitcher at all. But that effect is dwarfed by the number of plate appearances a batter makes against all pitchers

...

But what about cases where a batter has really owned a pitcher in the past—just utterly demolished him? Let’s restrict ourselves to cases with a prior TAv of .520 against a pitcher, or twice the average TAv. (By happy coincidence, that’s just about two standard deviations above the average, for those of you who care about such things.)

Historically, these have been more predictive of batter success than ordinary batter-pitcher matchups. But they are still dwarfed by the predictive power of our log5 expectation, by a factor of about 24 times. A manager is likely doing himself a favor if he puts a guy with that kind of extreme success in the lineup in place of a batter who’s otherwise reasonably close in ability. However, such cases are extremely rare, and even in these extreme cases, the whole of a batter’s historic performance (combined with knowledge of the platoon advantage) is still a much better gauge of how a batter will perform against a pitcher going forward.

...

The data isn’t telling us that batters can’t pick up certain cues about a pitcher, or that a pitcher’s repertoire is equally suited to all batters. However, 10, 50, or even 100 plate appearances aren’t enough to tell us whether what we’re seeing is one player with a special edge against another, or simply a small-sample-size fluke, and there’s too much at stake for La Russa and Washington to let themselves be overly swayed by such statistics to the detriment of their teams.

Thank you Colin for doing the work!

Moral of the story: take your noses out of your spreadsheets and index cards, and watch the baseball game instead.

Colin: I’d like to know the regression equation, of how much to weight the batter-pitcher matchup and how much to weight the log5 expectation.

Monday, October 24, 2011

Small sample sizitis

Someone asked me about small sample size.  I answered as follows:

==========================

It’s a fair question to ask.  Basically, the choice is presented as follows:

1. Octavio Dotel has faced Ian Kinsler 8 times in his career (and got him out 100% of the time).

2. Octavio Dotel has faced 3800 MLB batters in his career (and got them out 70% of the time).

Therefore, how much weight do we place on the 8 Kinsler PA, compared to the 3800 non-Kinsler PA?  Ian Kinsler does have something in common with everyone else: he’s a MLB player.  That is a huge commonality we have.  In that group of players are guys who are better hitters than Kinsler, but also quite worse than Kinsler.

So, we can limit the 3800 batters if you want down to the say 1000 batters faced that are about as good as Kinsler.

Now our choices become:
1. 8 PA against Kinsler
2. 1000 PA against guys as good as Kinsler is a hitter

The choice however is not either/or.  You can overweight the Dotel-Kinsler matchups, and I have NO PROBLEM with doing that.  How much do we want to overweight that?  Two times?  Five times?  Ten times?  Give me a number.

So, let’s say that Dotel-Kinsler tells us 10 times as much as Dotel-GoodHitter does in terms of giving us an estimate.  The 8 actual PA becomes 80 weighted PA.  You still have 80 weighted PA to add to the 1000 other PA in the pool.  That 8 actual PA is still only 7% of the conversation when weighted 10 times.

Not to mention the reality is that if you study it, as we have in The Book, the matchups are simply not predictive.  This is not a matter of opinion.  It’s a matter of fact. 

If someone ignores fact because they believe in their gut they are right, Colbert coined the word “truthiness” for that. I have no argument against truthiness.  By definition, those who argue based on truthiness can never be wrong.

Tom

Friday, October 21, 2011

Performance in tied games

Note: You may feel you are walking in the middle of a conversation.  You can take a step back and go here.

***

Ok, given that we accept that there are pecularities with the 9th inning, bottom half, tied game scenario (unintentional-intentional walks; fielders de-aligned from norm to either defend against the bunt, to guard the line, or to play the OF shallow), I’m now going to focus only on the tied-game scenarios for all innings.

So, here’s the A_empty group (as a reminder, all bases empty situations, regardless of outs), for tied games, broken down by inning (1st, 2nd thru 8th, 9th, extra innings):

First 4 lines is top of inning, and next 4 is bottom. 

Our “control” group is basically innings 2 thru 8.  As we can see, at home, you get a 12 point advantage in OBP, 10 point advantage in BABIP, and since wBABIP=BABIP that means the rate of extra base hits per in-play hit remains constant (i.e., excludes HR). 

wBABIP is BABIP but with extra weight placed on extra base hits and less weight placed on singles.

As we saw in the other thread, the massive difference in OBP in the 9th inning is explained by the walk rate and singles rate.  The extra innings carries this issue as well.  So, any analysis of tied games in the 9th inning has to be able to handle these parameters, otherwise you are going to introduce a bias.

The interesting one though is the 1st inning.  We ALSO see a massive gap in the 1st inning, with a 27 OBP point gap at home than on the road.  As we said, 12 points of that is the home field advantage.  The K and BB rates are simply hugely different, as it plummets to 14.5% K rate in the bottom half (compared to 17.3% in the top half) and jumps to 10.0% in BB+HB rate in the bottom (compared to 8.7% in the top).  It might be clearer to see it as a K to BB+HB ratio: 2.00 in the top half and only 1.45 in the bottom half.

My original speculation was that the road pitcher, being the second pitcher to take the mound, simply hasn’t found his mound groove.  I wish I would have included one other parameter in my dataset (which I can do on the weekend): handedness of pitcher.  If both pitchers are right-handed, they may “compete” for the mound groove.  If they were opposite-handed, then it might make it easier for each to take their own side.  Just a thought, one that I will also leave with the PITCHf/x-ers.  Anyway, someone asked about day/night splits maybe.  So, let’s see.

Here are the splits by day:

The control group (innings 2 thru 8) shows a 14 point home advantage for OBP and a 8 point advantage in BABIP.  There’s no extra base hit advantage.

Innings 9 and extra innings shows the same massive gap, as explained earlier. 

Inning 1 shows a 29 point OBP advantage at home, and 16 point advantage on BABIP, numbers that are a bit bigger than overall.

Here it is at night:

The gap is somewhat smaller at night in terms of home advantage.  The control group shows a 11 point OBP advantage and 8 point BABIP advantage.  The first inning shows a 26 point OBP advantage and 13 point BABIP advantage.

Across the board, the day time advantage is an extra 3 points in OBP.  I would not be surprised that if you look at historical home/away splits, that the day splits show a slightly higher home advantage.

Anyway, let me now focus on the B_2out grouping (all PA with runners on base and 2 outs).  Maybe by this point, the mound familiarity starts to come into focus (and maybe not using the windup helps too in combating the mound effect).

Control group shows a 14 point OBP advantage and 3 point BABIP advantage at home.  The first inning effect shows only an 18 point OBP advantage at home, but 13 point BABIP advantage.  The OBP gap between 1st inning and control group is no longer very wide.

I think the next thing to look at is simply by 1st, 2nd, 3rd batter in the game.  I’d bet we’ll see a huge gap in OBP when comparing the 1st batter’s first plate appearance in games at home and on the road.  And then a progressively smaller gap for the 2nd batter faced in the game, and 3rd batter.  And that by some point, maybe the 2nd inning, the gap remains static.

If it’s not due to the mound, then perhaps this effect is that the road pitcher has to sit around waiting to pitch in the 1st inning.  The home pitcher gets to pitch right away after his mound session.  Maybe the solution is that the road pitcher should warm up during the top of the 1st inning even?

Wednesday, October 19, 2011

Kelly Johnson’s approach to hitting

An interesting observation, that when the chips are down, the pitcher will go with what he’s good at:

DL: When you’re studying video, do you pay particular attention to how a pitcher attacks hitters who are similar to you?

KJ: Oh, yeah, although I look at all lefties. And I always check with guys in scoring position, because a lot of times that will tell you what a pitcher wants to throw, and what he feels comfortable throwing.

There are of course reasons that the pitcher will do that beyond that this is what he’s most comfortable with (runner on 3B, less than 2 outs, pitcher will pitch for a K, and batter will swing to avoid K).  In any case, good idea to see a pitcher’s repertoire by the 24 base-out states.

Friday, September 09, 2011

Curveball as first pitch to start the game

This rarely happens, and when it does, no one swings at it:

Going back to 2009, PITCHfx has identified 83 first pitches as curveballs, or 0.6%. Of those 83 first-pitch curveballs, batters have swung at just three of them. Of those three swings, two have resulted in fouls, and one has resulted in a home run - Ichiro’s swing, last night.
...
That’s a first-pitch curveball, a little out of the zone. Not only did Ichiro swing at it - Ichiro made contact, and obliterated the pitch into the right field stands. Said Hochevar:

“The first-pitch curveball to Ichiro, I don’t even know what to say about that one,” Hochevar said. “The [first pitch] of the game—it was a good pitch and he hit it eight rows deep. I’ve never seen it, I’ve never had it happen. I normally don’t throw first-pitch curveballs in a game, but what do you do?”

Wednesday, September 07, 2011

Ken Singleton and plate discipline

Love the interview.

DL: Are you a believer in protection?

KS: Yes I am. Most definitely. I could sense what was happening as soon as they started batting Eddie behind me. He was younger and a little more aggressive than I was, and he was trying to make a name for himself. And he was doing so. He was in the process of building a Hall-of-Fame career. Pitchers had a choice. We were both switch hitters and they could either pitch around me and go after him, or they could try to get me and be careful with him. It was sort of a “pick your poison” sort of thing.

I believe him.  And we’ve shown as much that protection exists.  The question is if the change in pitching to a hitter because of protection makes the batter, OVERALL, more effective.  There’s a shift.  But, is that a shift that overall has an impact?  And the current research says: no.

Tuesday, August 02, 2011

Do Umpires Give Preferential Treatment To Some Players?

SABR presentation by Patrick Kilgo, Hillary Superak, Lisa Elon, Mark Katz, Paul Weiss, Jeff Switchenko, Brian Schmotzer and Lance Waller.

Here are their conclusions.  I’ll post their file on my site tonight.  You can also read from Phil.

Umpires absolutely favor veterans with respect to false strikes

Umpires most likely favor veteran pitchers with respect to false balls. No evidence of benefit to veteran hitters

“Good” pitchers see the most preferential benefit

No evidence of rookie hazing

SABR_Umpires_Talk.pdf (pdf)

Monday, July 11, 2011

Home field advantage on pitch calls, by count

Phil continues his look.

Conclusion: from 2000 to 2009, home teams were somewhat more likely to get a strike call in higher-leverage situations than in lower-leverage situations. This is significant only at approximately p=0.1.

Wednesday, June 22, 2011

Ryne Sandberg’s hitting approach

Fantastic stuff.  I’d quote the whole thing really.  This part I particularly liked:

DL: Why did you have so much success against Bruce Sutter [8 for 20, 4 home runs]?

RS: His split-finger fastball came into my hot spot. I was able to recognize that pitch, and also anticipate it, coming down and hard into me. With him, I would swing where the ball would end up, which is very unnatural. That’s what made him so tough; if you swung at the pitch where it was, by the time your bat head got there it was too late. The ball would have disappeared. I was able to anticipate where the pitch went, which was in my hot spot.

Mike Bossy would say something similar.  While the goalie was angled to one side, Bossy was in the slot in front of the goalie.  When he’d get the pass, and he’d have to one-time it, he would shoot the puck directly at the goalie, because he knew the goalie would have to move from his angle to face him straight on.  So, he anticipated that the goalie would get into the right position (for everyone else except Bossy), and thereby leaving the position he was previously in.

So, this part also captured me:

RS: I faced them both a lot, and Darling mixed his pitches up. He was effective against an aggressive hitter, an aggressive fastball hitter, because he’d throw his split-finger fastball — a forkball is what he had — which to me appeared to be a fastball. I considered myself a fastball hitter, and when I was ahead in the count I was aggressive on the fastball. I tried to put those balls in play hard somewhere. He knew me as a hitter and would pitch me backwards a bit. He was a guy I had to battle, so I’d take a single up the middle if I could, or a single to right field.

That’s game theory right there.  While Sandberg would get success with other pitchers by anticipating, Darling knew that’s what Sandberg would do, so he had to “pitch backwards” against him.

Cue the Youtube video of Princess Bride.

Friday, June 10, 2011

How specific can we get in determining the true mean of a particular matchup?

Every matchup has a specific and true mean.  God herself would establish that specific and true mean at that specific point in time-space with zero level of uncertainty.  Pujols at Busch on July 3, 2011 against Doc and God knows that he can’t handle an outside cutter well, and the next pitch is going to be telegraphed by Doc as an outside cutter?  God says that Pujols will contact that pitch 23% of the time (if allowed to replay in that time-space an infinite number of times) with 0 level of uncertainty.

But what about humans?  If Pujols v Doc has an expected contact rate of 70% any time Pujols swings (with a certain level of uncertainty, say 10%), then how much a better mean estimate can we get in more specific situations (we find more data about Pujols and or Doc and or Busch and or the weather), and how much more can we reduce the uncertainty level?

Friday, June 03, 2011

Hits Per Strike

I appreciate that the authors of this paper were able to figure out that “at counts” are silly almost all the time (pdf).  They realize that the performance “at counts” requires us to know that the plate appearance has ended.  It seems an obvious concept if you think about it, but most people won’t think about it.  So, they get lulled into thinking that “at counts” mean something that they really don’t.  Plenty an analyst has fallen into this trap.

Now, the solution is to instead look at “through counts”, which means, given that you are at an 0-1 count, what happens by the time the plate appearance has ended?  Virtually all the time you need to talk about counts, you want through-counts.  The linear weights (or wOBA) by count is based on through-counts.

The authors of this paper however instead talk about “hits per strike” (and hits per contacted ball).  These are partial views to the performance.  Without talking about how often they take a called ball, the discussion will always be incomplete.

Therefore, whenever you see one of these papers or articles, and the central point is not about through-counts, then a large hole in the analysis will exist.

Glove-slap: Peter.

Friday, May 27, 2011

Pitch recognition

Fascinating stuff, from our favorite saber-interviewer:

DL: What do you see when the ball comes out of the pitcher’s hand?

AG: I see rotation. I can pick up on what the pitch is as soon as the pitcher lets go of it. Most of what you see is innate. If you ask some of the great hitters, they won’t all say the same thing. Some just see balls. Some guys see speed out of the hand. I can’t recognize speed, but I can recognize rotation. Some guys can recognize speed but not rotation and some guys just see a ball and swing. They just let their abilities take over and that’s not something you can teach.

If he can’t recognize speed, then does that mean that Gonzalez has a harder time against a fastball-changeup pitcher, as opposed to a fastball-curveball pitcher?

Friday, May 20, 2011

Trading location for velocity and movement

Chris Perez:

SB: And when things get a little bumpy? Then [where] does that extra gas come from?

CP: usually adrenaline. When things get sticky, I sacrifice location for velocity and movement

Wednesday, May 04, 2011

Platoon splits recap

Steve gives us a recap of various platoon splits.  I’m going to link you to his references in the comments, but you should read from the top of the page.

Tuesday, May 03, 2011

Through-Counts

Here’s from Peter:

UPDATE: At-counts here.

At-Counts

Peter’s chart for 2002-2010.

1. The “through count” is 100 times more useful than the “at count”.  The at-count is saying “if you know the at bat ends on this pitch, what just happened”.  It is so prone to misuse, and so limited in its good use, that you have to be extremely careful.

The through-count is saying “now that I’m in this count, what happens by the time my turn at bat ends”.

I’m glad Peter said he’ll have the better chart next week.

2. Remove all IBB.  They are nothing but noise, and they cause problems.  If you insist on keeping the IBB, then treat them like regular BB in wOBA.

3. Last point: it’s easier to follow if you sort based on hitter’s count to pitcher’s count.  That means, in order:

UPDATE: Through-counts here.

Tuesday, April 26, 2011

Non-randomness in pitching approach

I was just thinking about this topic driving in.  And here’s RJ doing some good stuff.  He looks at the rate of a pitcher’s first pitch being offspeed or not.  And then he looks at the rate of offspeed on all of a pitcher’s pitches.

He then compares the ratio of those two rates to a pitcher’s component runs allowed (SIERA in this case).  And, basically, the conclusion is simply:

it seems that different strokes work for different folks

That is, guys who pitch in a random way aren’t necessarily better off.  Cliff Lee is the one that has the most visible pattern based on this metric, and, well, it’s not hurting him.  It may be simply like it is with Vlad et al on the hitting side, that the optimal approach is one borne of experience: if you are not successful, you try something different on occasion, and if it feels good, you adapt that to your repertoire.  It seems almost tautological.  If a guy is performing well, you shut up, because even if it shouldn’t work, it is working (or at least seems to be working).  If a guy is not performing well, then you can try different things, though chances are, nothing is really going to change (though occasionally, you get a real change).

Saturday, April 23, 2011

Why has no one done this research?

Pitch f/x guys, this is directed toward you… Unless I missed something…

In The Book, we show that batter/pitcher match ups (at least in less than enormous samples of data) have virtually no predictive value.  We even show that batter results against certain classes of pitchers have little or no predictive value.

However, we seem to reserve the notion that certain types of batters may perform “better or worse than expected” against certain types of pitchers, beyond what we looked at in The Book.  For example, we often give managers credit for “knowing” that a high ball hitter might do well against a high ball pitcher or a good curve ball hitter may do well against a pitcher who throws lots of curves.  Etc.

So how about one of our pitch f/x guys doing this:

Establish two groups of pitchers - those that throw a lot of high pitches and those that throw a lot of low pitches.  You don’t need to put all pitchers into one or the other group - just the extreme ones - maybe the top and bottom 10%.  Then do the same for batters, only 4 groups - those that perform well against the high pitch and those that don’t.  And those that perform well against the low pitch and those that don’t.

Now look at expected versus actual performance of each group of batters against both groups of pitchers.  Expected would be their overall wOBA matched up (using odds ratio) with the overall wOBA of the pitcher, adjusted for handedness using platoon ratios.

Of course, you have to make sure that the data used for the groupings and the data used for the expected and actual performance are different.

Do the same thing for different locations (say outside and inside pitches) and for different types of pitches.

Wouldn’t that pretty much settle the long-standing debate over whether a manager, scout, or anyone can recognize and utilize to their advantage a favorable or unfavorable match up?