THE BOOK--Playing The Percentages In Baseball

Thursday, August 26, 2010

3 strikeouts, or one HR?

Beltre was ejected by a rookie umpire who, RULES ARE RULES, are within their rights to eject any player for practically any reason.  We are supposed to sit there and take it, not boo about it, yada yada yada.  So, ignore my golf shot, and focus on this:

“I was talking to Felix, we had a little bet where he told me he was going to strike me out three times and I told him I was going to take him deep. He struck me out in the first at-bat and I went back to my position and he was talking smack,” Beltre explained. “I was facing him, talking in spanish. I told him the next at-bat I was going to take him deep over the Monster. That’s all I said to him. I heard the umpire talking to me. I told him I was talking to [Hernandez], and he threw me out. I had no idea why he threw me out. I wasn’t even talking to him.”

In his career, Felix has struck out 1002 of 4615 batters.  So, 3 K’s would happen 1% of the time.  He’s given up HR at a rate of 2%.  Beltre has struck out in 16% of his PA, just a bit better than league average.  That doesn’t help us there.  And his HR rate is 3.7%, meaning that Felix-Beltre should give us a, say 2.5% chance of a HR.

Bad, bet, Felix.  Bad bet.

Monday, August 23, 2010

Good hitting v good pitching

Go to Amazon’s Look Inside of The Book, and do a search for
Profile #5
and that will bring you to page 94.  The study is available in its entirety at pages 94-96, with the conclusion being:

The Book Says:

There is no platoon effect with respect to the quality of the pitcher or hitter. Good pitching beats good hitting as much as good hitting beats good pitching.

Related threads here and here.

Thursday, August 19, 2010

Knuckleball hangover, part 2

Ben takes a look again, and sees nothing there.  This is really the great part of sabermetrics that you spend all this time to look for things, you do your best, and you don’t find it.  That is called a success.

Friday, August 13, 2010

In a universe filled wih Paul Malholm, it’s Bill Hall, not Prince Fielder, that is the hitting star

Jeremy makes the case.

Wednesday, July 21, 2010

Change in number of pitches thrown

Joe Kerrigan is reported to have said:

Kerrigan commented that the average game in 2010 required about 30 more pitches to complete than an average game did 20 years ago.

We actually have pitch count numbers for a bit over the last 20 years. I don’t have my numbers handy (but I will run them tonight, unless someone beats me to it), but the number of pitches per batter was around 3.6 twenty years ago and around 3.8 today.  If you figure around 38 batters back then and 39 batters today, per 9 IP, and we’re at 274 pitches per 9IP for two teams twenty years ago and 296 today.  That’s a difference of 22 pitches.  Thirty sounds like a stretch.

The reason is batters are taking more pitches, and the question is why are they taking more pitches.  Is it because of hard throwers, and so, they are just waiting it out?  Is it that they have become more patient?  Is it that teams are finding more patient hitters?

Monday, July 05, 2010

Knuckleballer hangover effect

Just beautiful stuff from Ben:

Read More

Saturday, June 26, 2010

Should batters automatically take with a 3-0 count, close game, and RISP?

I am always frustrated when batters take 3-0 pitches (at least it looks like they are taking, often not swinging at fastballs right down the middle) with runners in scoring position in a close game, especially with 1 or 2 outs and a base open, when the walk is not worth that much.

Obviously it depends on the batter, his eye, whether he is comfortable swinging at 3-0, etc., but is seems to me that batters simply take too often is those situations, similar to the way I think they swing too often on 3-2 counts, especially when the walk IS important.

Let’s take a middle of the road example.  Score is tied or batting team is down by a run, with 1 out, and a runner on second base.  We’ll assume an average batter with an average batter on deck for simplicity’s sake.  Count is 3-0.

Should the batter automatically take?

Anyone up to the analysis?  Obviously we can never know how batters would do when they are taking if they were not to take, especially since those are the ones who presumably don’t like to swing at 3-0, but I suppose we can get some clues from the average-type batters who do swing on 3-0 counts in various situations.

Thursday, June 10, 2010

Switch-hitters batting same-handed against changeup pitchers

Rays theory:

Ever since Dallas Braden and his nasty change threw a perfect game against the Rays, Maddon has stacked his lineups with players who bat with the same hand as the starting pitcher in order to neutralize that pitch. The change-up is a pitch that is typically used to neutralize opposite-handed hitters, and so Maddon is attempting to take away this advantage from pitchers with great change-ups by reducing the number of opposite-handed hitters in the lineup. ... However, the Rays sent up switch-hitters Ben Zobrist and Dioner Navarro to bat right handed against Marcum, and even more telling was that they not only used right-handed catcher Kelly Shoppach as the DH, but they hit him clean-up.

So, it’s a battle of
a. the same-handed hitters neutralizing the changeup as a weapon
vs.
b. the same-handed hitters being taken advantage by all the other pitches as a weapon

It would seem to me that it would have to be one mother of a changeup for a hitter to do that, but, I haven’t worked it out.

Fascinating thought though, and exactly the kind of thing that can get the field manager and front office working on a solution.

Tuesday, June 08, 2010

Game theory on 3-2 counts

David in a fascinating piece:

If you do the math, you find that hitters want to swing 74.5 percent of the time, which makes sense since pitchers will obviously just about always want to throw a strike given the high cost of throwing a ball when the batter is taking. In fact, hitters swing 73.2 percent of the time in 3-2 counts, which is just about what game theory would predict.
...
Do the math and you find that pitchers want to throw a strike 89.8 percent of the time. In actuality, however, they throw strikes on only 59.2 percent of their 3-2 pitches. Again, the hitters confirm our model’s predictions while the pitchers do not! What is going on?

My issue here is with his chart:

If the pitcher threw a strike and the batter went after it, for example, the hitter ended up with a .330 wOBA, on average. If the hitter didn’t offer, his wOBA was just .273. If the pitcher threw a ball and the hitter nonetheless took a swing, his wOBA was really pitiful—.188—but if he kept his bat on his shoulder, it was a hefty .688.

First, I don’t know how the wOBA can be .273 on strikes that the batter takes.  It should be .000 (a strikeout).  The wOBA on balls that that the batter takes should be close to .700, and it is (a walk). 

When he’s working out his probabilities to get to the equilibrium point, he’s using these numbers as fixed.  Well, the .188 wOBA on balls out of the strike zone that the hitter swings at is based on his expectation that the ball might be close to the strike zone.  And similarly, his .330 wOBA on swings at pitches in the strike zone is based on his expectation of a decent pitch.

However, what happens if, as David’s model suggests, that the batter will take every pitch?  The pitcher will throw easy strikes, and on the rare occasions that he doesn’t take, his wOBA on swings at pitches in the strike zone will be say .500.

So, unless I’m not following along correctly, we need a model that will float the values of the swings based on the ball/strike frequency.

Love this stuff.

Wednesday, May 05, 2010

Best 2-strike hitters

Someone asked on BJOL’s site about who is the best 2-strike hitter, and I replied:

For 2-strike hitters, it’s Pujols, Bonds, and Gwynn. At 0-2, I have them Pujols #1, Bonds 5th, Gwynn 6th. At 1-2, I have Gwynn 1st, Pujols 2nd, Bonds 4th. At 2-2, Gwynn, Bonds, and Pujols are top 3. Interestingly at 3-2, those three guys are between 20th and 30th best and a bunch of other non-star are in the mix. It’s an interesting count to see who the leaders are. Since Bonds and Pujols are likely the best hitters at ANY count, I’ll give the nod to Gwynn, relatively speaking. This is based on Weighted On Base Average, and on “through counts” meaning we care what happens to the end of the PA, once the 2-strike count is entered (and not just what happens if the PA ends at a 2-strike count… taking a ball on a 2-strike count is good and should count in the evaluation).

Monday, April 12, 2010

Time of Game

Poz the researcher gives us the data. 

The Science of Hitting, with Saberist Morgan Ensberg

Very good Q&A:

Matt Lentzner: There seems to be two basic approaches to hitting. There are the “see the ball, hit the ball” types that go on pure reactions and those that “guess” and try to anticipate what the pitcher will throw. Which type of hitter were you?

Morgan Ensberg: I was a see it and hit it. But when I was going really badly, I changed into a guess hitter and that was not the correct move. It would have been better to just attack. Of course that is hindsight.

...

ML: Did you ever see a “dot” on a pitch and if so, which pitches did you see it on? What other, if any, visual clues were you able to pick up from a pitched baseball?

ME: I saw dots on sliders all the time. I could see seams tumble on changes. We see everything.

Interesting. 

Friday, April 02, 2010

The Perfact Game Theory decision matrix model

Excellent job by Justin to give us something concrete to discuss.

Thursday, April 01, 2010

Strasburg-type of analysis

Pitchers-as-batters likely tells us nothing about the pitchers-as-pitchers they are facing.  That is, who cares how well Carpenter and Pineiro do when facing pitchers-as-batters?

Indeed, I strip away all pitchers-as-batters and batters-as-pitchers stats when I compile my reports.  Strasburg in college facing guys with no prospect of MLB action is the same thing: junk data.

That’s my theory.

Monday, March 29, 2010

A long PA has an adverse effect on the next batter?

This doesn’t look right to me at all:

Especially when he says there’s no bias:

I thought that maybe there was some selection bias involved, but in each grouping both the batters’ OSP and the pitchers’ OPS against were similar across the board (around .750). It wasn’t a matter of good pitchers dominating poor hitters after throwing a bunch of pitches to the previous guy.

Unless someone wants to do it themselves, I’ll see if I can repeat this later in the week.

Sunday, March 14, 2010

Change in performance as the game goes on

In The Book, we showed how the performance of the batter v pitcher improves each team he sees the pitcher.  It was based on data from 1999-2002.  One of my hopes is that aspiring saberists would extend alot of the work we did in The Book to look at historical data.  Matt did just that:

Friday, March 12, 2010

Changing the plane of a hitter’s swing

Jeff is dead right:

Let’s look at the breakdown statistically. Gomez’s career wOBA on grounders is .286, on flies it is .230, and on liners it is .689. Converting into runs, every fly ball he turns into a ground ball is worth about 0.05 runs. Every line drive he trades for a ground ball costs his team about 0.25 runs—a sacrifice of five times as much as the benefit gained by each fly-turned-grounder.

(Actually, .35, not .25.) But, yeah, just a fantastic point to make.  The key plane of the swing is to get line drives.  This is true for any hitter.  For some hitters, you alter the plane slightly to get more FB and others to get more GB.  But, you can’t live and die on the groundball.  You’ll die.

Juan Pierre is .221 on GB and .159 on FB.  He can’t be such a GB hitter that he throws away all his line drives and the .658 that comes with them.  So, I’m very interested in these hitter experiments, but these I presume are going to be slight adjustments, not wholesale changes.

Sunday, March 07, 2010

Changing batting approach in new parks

Love the research here.

Friday, March 05, 2010

Batting eye on 2-strike counts

I get super-excited on anything to do with counts and approaches, because that is the heart of baseball. 

I need a 101 version of this.  It looks like Craig knows what he’s talking about, but I need to catch up to him.  His site is frankly overwhelming for me.

Monday, February 08, 2010

How do you spot a lefty masher?

Pat asks:

However, lefty/right analysis has advanced since my adolescence, and I think this post from MGL is a must read for what I’m talking about. In it he says:

IOW, how a batter does against RH pitchers informs us on how he will likely do against LH pitchers and vice versa. Why? Because there is not much of a spread in true platoon splits among ML baseball players yet there is a large spread in overall true hitting talent among ML baseball players. So if we see a large platoon split, like for a player like [Ryan] Howard, it is likely a fluke. If a player does really well versus RH pitchers but terrible against LH pitchers, both the “really well” and the “terrible” numbers are likely fluky and the “truth” is somewhere in between.

Howard has a .719 OPS in the last 4 years versus LHP. How would we estimate his “true” OPS versus LHP? You might be tempted to just use the .719, which is not too good or you might be tempted to use the .719 and then regress that toward the league average for a LH batter of Howard’s physical characteristics, which might be around the same or a little higher – I don’t know. Both of these methods would be wrong. You cannot ignore the fact that he also hit 1.052 in OPS versus RHP over the same time period (last 4 years) and in many more PA. This suggests that he is a very good hitter overall (which he is) and that the .719 is somewhat of a fluke.

And then Pat asks the interesting question: do splits change by age?

I’d also like to see how players do split wise over the course of a career. Obviously the skill of hitters diminishes over time, but it’d be interesting to see if the splits are larger. Ryan Howard certainly is not “old” at 30 years old, but his skill set and physical size certainly have shifted over the years. While he was never slender, Howard is certainly a “bigger” guy than he used to be, and probably a good amount slower. Besides that, as players get older they tend to lose some hand-eye skill, an effect that may be magnified when facing a pitcher of the same handedness. Here are Howard’s wRC+ from 2006-2009, going from overall to versus lefties and then versus righties.

2006: 166, 133, 182
2007: 140, 110, 159
2008: 123, 91, 143
2009: 141, 71, 178