THE BOOK--Playing The Percentages In Baseball

Was there someone else to turn to out of the pen?

Well, Edwin Jackson had 3 days rest.  Verlander had one (after throwing 120 pitches), but it was his normal day to throw.  We don’t know how that would’ve worked out, but obviously Rodney in his 4th inning of work had run out of gas.

Whatever.  El Borracho and Full Pack Leyland losing was a sight to behold.

Knowing that Rodney could pitch for 3 innings, why not just bring in Lyon in the 6th or 7th and then Rodney in the 8th or 9th?

Boy, you’re not kidding about Chip.  His call of the DP that ended the 10th ("Line Drive! Base Hit!") was absolutely brutal.  Of course, the rest of the game was brutal as well, but that was the low point.

Re: #7, my thinking was that he wasn’t really looking for a DP but just felt that he’d rather face Casilla, Tolbert, and Punto than Young.  I’m not saying he’s right, but that was my impression of his thought process.

Casilla did make a baserunning error by not going back to the bag.

If you are going to tag it means you think the ball is deep enough score tagging up.

What good does it do to be off the bag?  If the ball is caught, you lose time getting back to the bag as happened to Casilla.

If you immediately go back to the bag, you get a better push off from the bag and time your take-off to the ball dropping or touching the fielder.

Runners are taught that on a ball to the outfield you immediately go back to the bag unless you do not intend to tag up.

In this case Darling was right.

Well, the bonus is the possible GDP.  But the price is you now don’t have a base open, so that if you walk the next batter, you are in big trouble.

Let’s call that a wash (and I don’t think it is, given the low likelihood of doubling up Casilla) Given that Casilla is lefty and Young is righty, why would you want to face the lefty?  Since Young is overall the better hitter, maybe that is a wash given the platoon situation.  Probably no big deal I guess.  Even though in The Book, Andy has shown that pitching around someone yields little or no advantage, if it is me, I am pitching to Young and pitching carefully.  Pitching to Casilla with runners on first and second in that situation is too nerve racking.  You are bound to be worried about walking him if you get behind in the count.

"Runners are taught that on a ball to the outfield you immediately go back to the bag unless you do not intend to tag up.

In this case Darling was right.”

Yes, I have coached and played baseball for many, many years. I am acutely aware of how all of that works.  You did not digest my post.  If the OF’er is playing shallow enough (which they were obviously) and a line drive is hit and you are not sure whether it is going to drop or not, you may HAVE to get a lead in order to be able to score if it drops.  That is not true when the OF is playing normally.  If it drops and you already tagged, then you simply trot home because the throw it too far to get you out at home on the base hit and the OF’er is not likely to throw home anyway.

However, if the OF is playing in and you are the winning run on third, you CANNOT tag on all line drives that might drop in for a hit.  The reason is that if you do, you can get thrown out at home on a hit!

What if the batter had hit a line drive that was going to drop 10 feet in front of the already drawn in left fielder?  Guess what happens if you tag?  You actually get thrown out at home (assuming a good throw).  So at what point do you tag?  It depends on how much you think the ball is going to drop in and where you think it is going to drop!

Again, that is only true with the OF playing in and the winning run on third (and the OF’er throwing home even if the ball drops in).  Otherwise you are taught to tag up on any ball that may or may not be caught and that is reasonably deep.

Leyland should have used lefty Bobby Seay, who projects better than even Rodney, when Span (L), Cabrera (R) and Mauer (L) were due up. Even better, he should’ve used Seay instead of Miner when they were due up in the seventh.

As for the IBB, Leyland probably saw Young hitting .280 and the fact that he was “hot” and made his decision based on that.

I don’t think Seay was available.  He’s been hurt.  Lots of people on the game thread on BBTF complaining about the use of Miner.  Not too many people think that Leyland is a very good in-game manager either.

I watched the game on mute.  That will be all I say on quality of the broadcast team.

I’d agree with everything you’ve said but would like to note three, ok 4, things.

1 - Casilla may or may not have made a mistake, depending on what he could see and whether it looked like the catch would be made, etc.  However, those of us that watch ~140 Twins games a year have often noted that Scotty Ulger (3rd base couch whose name was mispronounced by the broadcasters) is abysmal.  He should have been throwing up the biggest stop sign in the world there.  The throw beat Casilla by at least 20 feet.  Only a pretty fantastic slide and the fact that the throw was all the way on the other side of the plate made it close.

2 - Gardy pinch hit Harris (wOBA .294, RH) for Morales (wOBA .339, switch hitter) against a righty?  Oh, and replaced Kubel instead of Young with Gomez, which he always does with defensive substitutions.

3 - The “shoestring catch” was actually a sliding attempt.  It was a makeable play, but obviously a bad decision due to the game situation.  Granderson also took a pretty bad path to the ball backing it up, which is what turned it from a double into a triple.

4 - Check out the fangraphs graph of the game.  It’s pretty good.

I guess I’m glad I was watching the game at the gym with no sound.

The worst play was definitely Cabrera not scoring on that high chopper. When I saw the second basemen throwing home I first thought it was an act of desperation.

watercott/10, you really think it was a mistake for the Twins to send Casilla on that fly ball? With Casilla’s speed and the fact that Raburn might be a little rattled after butchering that flyball I think the odds were high enough to go.

http://mlb.mlb.com/media/video.jsp?content_id=7016483
the triple

The Cabrera force out took about 3.3 seconds on the video from contact.

"The throw beat Casilla by at least 20 feet.  Only a pretty fantastic slide and the fact that the throw was all the way on the other side of the plate made it close.”

No way was that a mistake.  I would think that you only need to be safe around 20% of the time for that to be a correct send.  Clearly the chances of being safe were more than 20% since as it is it was almost a tie (it is not even clear on the video that he was out).  Actually, after checking the WE charts, you only need to be safe greater than 14% of the time for the runner to attempt to score.  That means you pretty much go on anything.

OK I watched the video of the triple.  It is hard to tell exactly what happened, but it looked like the ball was eminently catchable or at least trapable and that Rayburn horribly misplayed it or lost sight of it at the last second.  I can’t tell what Granderson was doing.  Since it wasn’t deflected, it would seem as if he should have backed it up better than he did.  There is no way he “should have played it as a single.” It was definitely catchable and even when you don’t make that catch, there is at least a 70% chance you trap it or keep it in front of you as long as you don’t “ole” it like Rayburn did.

Re: #10, MGL, have you ever experienced that type of pressure with a full count, bases loaded and 2 Outs and the season on the line. Of course, Laird was swinging at ball 4, but I think it is lame to say any major leaguer should be able to recognize that as a ball. Those guys are not machines…

After the game, Cabrera said that it was the most exciting game he’s ever played in.  WTF? Cabrera played in the 3-0 comeback with the Sox in 2004. Maybe it was because this one went into extra innings? No, Game 4 AND 5 of the 2004 ALCS went into extra innings. Maybe its because Cabrera was instrumental in this win? No, he had a .379/.424/.448 line in the 04 ALCS. Do we have to assume he just got caught up in the moment and let it go?

That game was crazy.  Other odd things that jumped out at me:

Why wasn’t Nathan brought in to start the 8th inning?  The Tigers best hitters were coming up.  In the end, Guerrier messed it up and Nathan had to come in anyway, with runners on.

The Twins made some defensive substitutions in that same inning that ultimately led to Gomez hitting behind Mauer.  Whoops.

Inge got “hit” on his jersey, but not awarded a base.  Big non-call there.

The grounder up the middle that scored the Twins 5th (?) run.  I thought Polanco was going to get it and turn it into a DP. 

I think you covered the rest of the crazy stuff I saw.

"#10.  You know how the pitch f/x research indicates that batters swing at WAY too many pitches out of the zone on a 3-2 count?”

Do we actually know that’s true?  I’m sure hitters swing at many pitches out of the zone, but that doesn’t mean it’s too many.  Swinging more often also turns some called strike 3s into hits (of course hitters are better at taking these pitches at 3-1—the cost of taking a strike is much lower).  As we learned from that nice Baseball Analysts article, hitters begin to swing before they even know they are doing it, so it’s not really a “decision.” At each count, the hitter can bias his response toward “probably swing” or “only swing at good pitch”—but that’s really all he can do.  After that it’s reflexes and the unconscious mind doing the work. 

So, do hitters bias themselves toward being too aggressive at 3-2?  To answer that, we’d have to know how many additional called strike 3s they would get, in addition to extra BBs, by being more selective.  We obviously can’t just say they should take more of the pitches out of the strike zone.  Being less aggressive might also result in lower BA/SLG on balls put into play.  I don’t know if/how we could sort that all out.  But I don’t think the pitch/fx data alone can tell us the answer.

An interesting game to watch.  I think that’s what we want from a game like that.

On the Zach Miner observation: Living in southeast Michigan, I’ve had the chance to watch Miner pitch more than anyone should.  I’d be willing to argue he’s one of the worst pitchers in the major leaugues.  I can’t understand why the guy still has a job, let alone gets to pitch so often.

The 3rd base coach was yelling at Casilla to get back to the bag.  He knew it was going to be caught.  It wasn’t a particularly difficult catch to make.  You have to be able to judge that better than Casilla did.  Casilla must have thought the outfielder was playing at normal depth.  That was a very easy play for Raburn.  He took a few steps to his left and actually had time to set his body for the throw home.  He caught it face high.  There’s no chance that’s a base hit so you go back to the bag and that’s why the 3rd base coach was yelling at Casilla to do so.  It was definitely a mistake. 

Besides, there were runners on 1st and 3rd so even if the ball did drop in for a hit and you realize you can’t score, you’ve got bases loaded and 1 out.  You’re essentially in the same situation you were. 

Watch the replay.  That was an easy out and the only reason it didn’t sound like it was because Chip Caray is awful at his job.  A good announcer recognizes how easy an out that line drive is.  The 3rd base coach did.  Casilla should have.  The Twins won so it doesn’t matter, but that’s an awful mistake on Casilla’s part.  It could have cost the Twins the season.  It’s Casilla’s job to pay attention to the outfield positioning and if he had done that, the Twins win that inning.  The only way that ball was going to fall in safely was if Raburn dropped it.  Casilla should have known that.  All he had to do was listen to his 3rd base coach.

I’m torn on this.  I think that if he felt it was better to be down the line if the ball dropped in, then that’s absolutely fine.  But if you’re not going to score from standing on the bag on a base hit, you’re definitely not going to do score if the ball is caught in the air (probably fielded in the almost exact place that it would have been fielded otherwise).  On top of it, feeling you need a headstart to get home on a hit AND attempting to tag up was probably not the best situation.  If the hit was that you feel needed a head start, then what’s to think you can run back to the bag and then score, fielded without a bounce?

The problem is whether or not we think the ball would be fielded around the same place if it were caught vs. dropped.  But I think with a ball like that, it could be generally assumed that it’s a push, no?

"I would think that you only need to be safe around 20% of the time for that to be a correct send.”

This cannot possibly be true.  If you hold you have the bases loaded and 2 out.  Your chance of winning on that PA is equal to the next hitter’s OBP, plus chance of reaching on error.

Rally:  if he holds it’s 1st and 3rd, 2 outs.

Yeah, but with 1st & 3rd, 2 Outs and Span up next, it has to be higher. Even when not accounting for the possibility of 2 BB in a row I guess Span’s true talent level BA is around .300 ...

I think Gardenhire’s usage of Nathan was downright criminal.  Nobody’s talking about it because it was a by-the-book, risk-averse (vis a vis talking to the press after the game) move.  Eighth inning, you go to your eighth inning guy, right?

Joe Nathan is easily their best reliever, one of the best in the league.  If you’re not going to use him to protect a late one-run lead with the season on the line against the opposition’s best hitters, why use him at all?

This happens all the time in the playoffs, and it drives me batty.  I wrote about this in last year’s THT annual, which I don’t have with me right now.  I believe this happened with Papelbon (or was it K-Rod?  Can’t remember).

Keith Law went off on Leyland this morning for the decision to lift Porcello after Kubel’s home run in the sixth and only 92 pitches, considering that moments earlier he had made Mauer look foolish. Does that decision bother anyone here?

Mariano Rivera has a career 438 batters faced, 117 IP in 76 G, in the post-season.  That’s 5.8 batters per game.

In the regular season, he has a career 4.5 batters per game.

That’s a huge shift, and from my memory, it seems that Torre was not afraid to bring him to pitch 2 innings in the playoffs.

back to Casilla on 3d: has anyone here ever seen a runner starting at 3d thrown out at home on a single to the OF in any situation?

His biggest mistake, though, as noted by David and Millsy, was thinking that it could possibly be a single when the OF was playing in.  TBS had shown several helpful angles of the Tigers OF that inning (not commented on by Chip and Ron) and the OF was playing in, as would be expected with the winning run on 3d, after playing deep after Cuddyer led off with a triple (which I was thought was interesting).

The error in judgement by Casilla was not knowing where the outfielders were playing.

He should’ve known they were in, he should’ve immediately gone back to the bag, and then judged if the ball is going to be a base hit or not.

He did the opposite. Judged, then went back, which cost him a good start off the bag, and got him thrown out.

Actually, you are right Rally, anyway (whether it is bases loaded or 1 and 3).  I screwed that up.  The difference is not the BE point.  The correct BE point for an average team in an average situation is around 29%.  Thanks for catching that.

Clearly it was correct for him to go, even if he needs a 30% chance of being safe.

I don’t disagree with anyone on the tag up thing.  My only beef is with Darling (and anyone else) who claims that he is supposed to tag up on any ball hit to the OF, whether he thinks it is going to drop or not.  It is true that you almost never see anyone get thrown out at home on a “single” (I’ll guarantee it has happened before in the history of baseball) but if the ball drops well in front of the OF in that situation, he is in fact likely to get thrown out at home, so there is a point at which he is NOT supposed to tag.  That is my only point.  Now, if he made a bad judgment about where the ball was likely to end up, that’s another story.  I have no problem with that assessment.  I didn’t watch the replay closely.

Guy, maybe you didn’t follow the research, but according to the pitch f/x data, batters swing at WAY (way, way, way, way) too many pitches out the strike zone on 3-2 counts.  They swing at more bad pitches at 3-2 than at 2-2.

And yes, #16, I realize that that is a pressure situation.  Why he swung at that pitch, I don’t know.  Some batters would and some batters wouldn’t.  It was obviously a mistake.  How “bad” of a mistake, who knows?

No way was that a mistake.  I would think that you only need to be safe around 20% of the time for that to be a correct send.  Clearly the chances of being safe were more than 20% since as it is it was almost a tie (it is not even clear on the video that he was out).  Actually, after checking the WE charts, you only need to be safe greater than 14% of the time for the runner to attempt to score.  That means you pretty much go on anything.

MGL - I don’t know how you figure that.  The Book’s WE charts say that a team tied in the bottom of the ninth or later with men on first and third and 2 outs will win 64.5% of the time.  Let’s figure wins for 100 attempts of trying to score on a catch.  If he is successful 29% of the time that gives his team 29 wins.  When he is out the other 71% of the time his team has a 50% chance of winning for another 35.5 wins.  Total wins are 64.5.  That is your break even point by “The Book”.

Mariano Rivera has a career 438 batters faced, 117 IP in 76 G, in the post-season.  That’s 5.8 batters per game.

That’s a huge shift, and from my memory, it seems that Torre was not afraid to bring him to pitch 2 innings in the playoffs.

If Rivera’s 1995-96 playoff appearances are removed — he was a middle/setup reliever in those years — the numbers change a bit, to 361 batter faced and 97.1 IP in 65 games.

That drops his batters faced per game to 5.55, though we can see simply from the games total and IP that he has thrown more than one inning in half his postseason appearances as a closer.

Peter, we must have cross-posted…

#27, Porcello seemed to be rattled after the homerun and started overthrowing. So, what do you do if you’re the manager? He’s young and inexperienced, so do you risk leaving him out there on the chance that he’ll pull himself together? I personally don’t think the situation was all that clear cut.

MGL:  I found Dave Allen’s articles on swinging at 2-2 and 3-2, but I don’t buy the conclusion.  Hitters SHOULD swing more at 3-2 than 2-2, even though a ball is worth more in the 3-2 count, because it’s much more likely the pitch will actually be a strike.  Swinging more often will result in more swings on balls the hitter ideally shouldn’t have swung at, but also more (and perhaps better) swings at pitches they should swing at. 

If hitters are in fact swinging too much at 3-2, then wouldn’t we expect worse outcomes when the batter swings?  But that’s not what happens.  At 2-2 batters are .189/.285 BA/SLG, but at 3-2 they are .233/.379.  The aggressiveness at 3-2 pays off, because the 3-2 pitches are easier to hit.

The pitch f/x data can tell us the benefit/cost of swinging at pitches in particular locations. At 3-2, batters are swinging at a lot of out-of-zone pitches.  But you still can’t say whether hitters are swinging with the right frequency without taking account of the proportion of strikes/balls from the pitchers, the outcomes on BIP, and—the toughest to figure—how a reduction of swing% on balls out of the zone affects the swing% on balls in the zone. If hitters were less agressive at 3-2, they would draw more BBs but would also end up swinging less at balls in the zone at a real cost (as we see at 2-2)—likely offsetting the gain in BBs.

Addendum:  Dave A. reports that hitters swing at 90% of in-zone balls 2-2, but 93% of in-zone pitches at 3-2. And there must be a lot more in-zone pitches at 3-2.  So the higher swing rate on bad pitches at 3-2 does also yield more swings on good pitches.

MGL - Yes, sorry.

Guy, I’d have to re-read the article/study.  Shouldn’t batters swing at fewer pitches out of the zone no matter what at a 3-2 count than a 2-2 count?

There are some things that we just know.  We know that batters swing too often at 3-2 counts because we know that they are taught since they were young not to get called out on strikes even on 3-2 counts.  I would bet my house and my life that they swing too often on 3-2 counts without seeing any data.  How much “too often” I don’t know. The pitch f/x data can tell us, assuming, as you say, that everything is analyzed.

"Shouldn’t batters swing at fewer pitches out of the zone no matter what at a 3-2 count than a 2-2 count?”

Well, hitters should never (almost) swing at pitches outside the zone.  But of course that’s easier said than done.  But no, that’s not necessarily the case.  If there are significantly fewer OOZ pitches at 3-2 (true), and if swinging more frequently at them also means swinging more at the much larger group of in-zone pitches (also true), then swinging more often at 3-2 could easily be correct.

Let’s turn it around:  at 3-2 (vs. 2-2), the next pitch is much more likely to be a fastball and much more likely to be in the strike zone.  Given that, don’t we want hitters swinging more frequently?  (and they do:  75% vs. 65%)

Or look at it this way:  Hitters are swinging at 57% OOZ pitches at 3-2, and 48% at 2-2.  So there only need to be 16% fewer OOZ pitches at 3-2 to yield an equal number of “bad” swings.  That seems very likely (I couldn’t find these #s).  And look, if 3-2 hitters really are swinging at more bad pitches, shouldn’t they have more Ks and fewer solidly-hit BIP?  But we see the opposite:  they have a higher BA and SLG when swinging. 

Dave’s article is here:  http://baseballanalysts.com/archives/2009/08/swinging_in_a_f.php

“There are some things that we just know.”

Well, I’ll just observe that this is the kind of comment you would crucify poor Joe Sheehan or Rob Neyer for making. :>) And I wouldn’t bet your house or life on this issue just yet.....

I went to mlb.com to watch online. They were advertising ‘Postseason TV’ online, and I was already an mlb.tv Premium subscriber, so I got in with my existing account. I got a new video player which let me select a camera, but I never found a selection for the live TV feed (such as nbc.com has for Sunday Night Football). So when I watched with the cf camera, I got to se the pitcher and home plate, even when the ball was hit elsewhere. No balls in the of, or plays at first. If this was supposed to be some premium service, either I didn’t figure it out, or it sucks.

That said, I was able to watch the pitchers when the ball was in play. Did Fernando Rodney evr back up a base? When the triple was hit (the potential winning run in extra innings) Rodney is shuffling around behind the mound watching the play and the throw to third. If the ball was overthrown he wasn’t covering and the winning run would have scored. On the much discussed throw to the plate on the flyball, he again never left the mound area.

"Well, I’ll just observe that this is the kind of comment you would crucify poor Joe Sheehan or Rob Neyer for making.”

That is absolutely, completely untrue.

I disagree 100% that batters are correct if they swing at more OOZ pitches at a 3-2 count.

For one thing, if they are more certain that they are getting a fastball, they are MUCH less likely to be fooled by the pitch. My guess is that 65% or more of all OOZ swings are on off-speed pitches. Anytime you are more certain that you are getting a fastball, you will swing less at OOZ pitches simply because it is so much easier to recognize the location of a fastball. If you are looking fastball and get an off-speed pitch, you can bet that you are swinging at pitches out of the zone.  If you are looking off-speed and get a fastball, you usually take pitches in the zone.

More importantly, when a pitch is on the way, at some point you make a decision whether it is in the zone, out of the zone, or you are not sure.  On a 2-2 count, if you are not sure, you have to make a conscious decision to swing more often because the payoff for the take if it is a ball is not that great.  At a 3-2 count, clearly on a marginal pitch, you should take more often because of the payoff.  I see no way that on a 3-2 count the correct strategy by the batter involves swinging at a higher percentage of OOZ pitches. No way.

So yes, while it is true that if you expect to get more pitches in the zone, you will swing more often, that is more than made up for (IMO) by the fact that you are getting more fastballs which are MUCH easier to recognize (location) AND you are simply going to swing less often at a marginal pitch - a pitch you are not sure of.

Aren’t there more OOZ swings on an 0-2 count than a 1-2 count and a 1-2 count than a 2-2 count?  Why would there be a sudden turnaround at a 3-2 count?  You expect fewer pitches in the zone at 0-2 than 1-2, fewer at 1-2 than 2-2, and fewer at 2-2 than 3-2.  Your logic makes no sense to me.

The Nathan point is a very good one, particularly when you consider Guerrier isn’t a very good pitcher for an “eighth-inning guy.” I did a quick projection and got him as a 4.20 ERA, which is actually below-average for a reliever.

I have Guerrier projected as the second best pitcher in the MIN pen, with an ERA .5 runs less than an average pitcher (starters and relievers combined), which is no great shakes for a short reliever, but OK.

Det, on the other hand, has no real good relievers.  Rodney is not very good for a closer at all.  Lyon is a little worse, and everyone else is average to replacement level.  Miner is near replacement level and should not have pitched in a high leverage situation in that game.  The way to handle must-win games like in the late innings is to try and get the platoon advantage as much as possible by constantly changing your pitchers.

As far as taking out Porcello which got a lot of criticism, I have no problem with that as he is not a very good starter in the first place.  Any reliever is going to better than him after a few times through the order.  It’s not like taking out Verlander too early.  Not even close.  You can take Porcello out in the 5th inning when he is pitching a shutout and you are not going to lose anything other than perhaps some flexibility in the later innings.

"Aren’t there more OOZ swings on an 0-2 count than a 1-2 count and a 1-2 count than a 2-2 count?  Why would there be a sudden turnaround at a 3-2 count?”

No, and that’s the point.  OOZ swings:
0-2 .414
1-2 0.478
2-2 0.484
3-2 0.568
Why don’t hitters get killed on 0-2 OOZ balls?  Because they know there’s a high probability of a non-strike, so they swing less often.  The more balls in the count, the more often the pitcher throws a strike, so the more often the batter swings.  You see the same pattern for in-zone pitches:
0-2 0.850
1-2 0.880
2-2 0.898
3-2 0.927
The two rates rise and fall together by count.  Compare any two counts, and if hitters swing more often at OOZ pitches they also swing more often at IZ pitches (and vice-versa).  What that tells us is hitters have a general tendency to swing more or less at each given count.  Which makes sense:  if hitters could suddenly discriminate better between balls and strikes, why wouldn’t they do it at other counts too? 

Let’s say that the IZ/OOZ ratio is 70/30 at 3-2, but just 60/40 at 2-2.  In that case, at 3-2 hitters are “correct” .7*.93 (strike/swing) + .3*.43 (ball/take) for a 78% rate.  At 2-2, it’s .6*.9+.4*.48 or just 73%.  I don’t know if 60/40 is right at 2-2, but the point is that the IZ/OOZ ratio is critical.  And again, if hitters are showing worse judgment swinging at 3-2, why are the outcomes better? 

Interestingly, hitters strike out looking at the same frequency, about 5%, at all 2-strike counts.

Would hitters be better off taking the 2-2 approach, taking 35% of pitches, at 3-2? That seems unlikely to me, though I suppose it’s possible.  With so many more pitches in the strike zone, hitters would draw a few more walks, but at the cost of fewer hits.  This is the question we need to answer.  The fact that hitters swing at more OOZ pitches at 3-2, by itself, tells us very little.

I guess I was wrong with my assumption (that there are more OOSZ swings the greater the pitcher’s count).  However…

That jump from .484 to .565 is WAY too much.  As I said, I KNOW beyond a shadow of a doubt that hitters swing at too many 3-2 pitches.  There is no way they can’t. We know two things that dictate that that is true.  One, all through their careers, almost all players are taught incorrectly that you should never take a called strike 3. Almost no one is taught that the better the count (for the hitter), the more you SHOULD take a marginal called strike 3.  Almost all young players get negative feedback when they take a called strike 3 even with a 3-2 count on a marginal pitch.  That is even true on a professional level.  There are few if any professional coaches and managers that actually tell their batters that it is OK to take a marginal pitch on a 3-2 count or that if they rarely or never get rung up on 3-2 counts, they are swinging too often (which is true of course).  In addition to that, when you get rung up on a 3-2 count with the bat on your shoulders, you feel bad and generally get no positive feedback on any level of baseball, especially the amateur level.  All of that I am afraid leads to one unequivocable conclusion - which is that batters as a whole will swing too often at 3-2 counts for obvious reasons.  Why wouldn’t they?  You can’t “feel” the gain of those extra walks you get, but you can surely feel the pain of those extra called strike 3’s!

Anyway in this case, the empirical evidence is somewhere in the data.  I have not looked close enough and I am not sure I could find it anyway.

Saying that you don’t know that hitters swing at too many 3-2 pitches without evidence is like saying that you don’t know that more than 50% of average (random) people will sacrifice monetary expectation for security.  You DON’T need any evidence for that.  You KNOW that it is true.  There are many tendencies that you KNOW are true without needing any empirical evidence.  (There actually is “evidence” - it is just not necessarily conventional, empirical, or even obvious.) That is why I said that if someone like Neyer or Sheehan made an assertion without any empirical evidence to support it is NOT a given that I would crucify them.  It would obviously depend on that assertion.

Anyway, let’s leave it for another time for someone to really analyze the pitch f/x data to determine how much batters over-swing at a 3-2 count!

MGL, would you agree that a smart hitter, you know, your Pujols, your Tony Gwynns, etc, would not swing excessively at 3-2 counts OOZ?  That is, given that there is a great inefficiency at 3-2, and given that there are some smart hitters who are simply efficient, then we expect to see a greater spread in OOZ swings at 3-2?

"That jump from .484 to .565 is WAY too much.”

You have no idea if that’s true.  The answer depends on 1) the relationship between that jump and the jump in swings on in-zone pitches (which you don’t seem to be acknowledging), and 2) the relative proportion of in-zone and OOZ pitches at each count.  The fact that more 3-2 pitches are in the zone is hugely important.  Also, the two pools of OOZ pitches are not the same:  the 3-2 OOZ pitches are almost certainly closer to the zone and therefore more likely to be called strikes when taken and easier to hit when the batter swings (and the 3-2 in-zone pitches are presumably more hittable, another reason to swing).

You’re making two (at least) mistakes here.  The first (also in Dave’s article) is looking at this strictly in terms of the relative value of strikes and balls at each count.  If you look at it that way, hitters should take MUCH less often at 0-2 than 3-1, right?  At 0-2 the value of the ball is trivial, and the cost of a strike huge, relative to 3-1.  But hitters take 53% of 0-2 FBs and just 45% of 3-1 FBs.  Does this mean they’re making a huge mistake in approach?  Probably not, because at 0-2 there are far more pitches thrown out of the zone. 

The second mistake is looking at the swing rate on OOZ pitches in isolation, when there is clearly a link between that and swing rate in-zone.  Once the hitter primes himself to swing a lot, some, or a little, it will affect BOTH the in-zone and OOZ swing rates.  If you teach the hitter to take more at 3-2, he’ll take a bunch of called strikes as well as draw walks.  (And of course pitchers will start throwing him more 3-2 strikes, compounding the problem.)

And one more time:  if hitters’ approach is better at 2-2 than 3-2 (your view), why are the outcomes on 2-2 swings worse? 

* *

“There are many tendencies that you KNOW are true without needing any empirical evidence.”

Well, there are things we know are true because past research has shown it—or something very similar—to be true.  When you said that doubleheader splits are no more common than any other 2-game sequence, I was fine with that—but that was based on lots of evidence that game outcomes are almost entirely independent, aside from strength of the 2 teams and HFA (and it turned out to be true).  If you want to say clutch hitting is at most a small skill, without providing documentation, fine—but again, that’s based on actual evidence.

This is entirely different.  You think we should accept as fact that trying hard to avoid called strike 3 is wrong, simply because it is baseball conventional wisdom.  (And btw, hitters seems to avoid called strike 3 equally on all 2-strike counts, nothing special about 3-2.) How is that any better than blindly accepting the conventional wisdom itself?  I guess you’re saying that this advice seems motivated by some kind of misplaced machismo, not baseball analysis.  But that hardly proves it wrong.  The fact that it keeps getting taught may mean it has passed the test of time.  The fact is, you have no idea whatsoever whether it’s wrong. 

If there is going to be a default assumption, it must be that hitters swing the correct amount.  That is what we usually find in these game theory exercises.  Not because hitters or coaches are brilliant and understand the probabilities, but just because “natural selection” in the evolution of the game (trial and error) tends to produce equilibrium outcomes.  The burden of proof rests entirely on anyone claiming that hitters swing too much (and as best I can tell, we have no actual evidence of this).  I think it’s entirely possible you’re right about this.  But in the absence of evidence, our best assumption is that you’re wrong.

Tango/47:
Looks to me like Pujols changes his approach at 3-2 (compared to 2-2) in the same way other hitters do.  Pujols:
Ball/Strike/SLGcontact
2-2: .39 / .15 / .660
3-2: .30 / .14 / .744
At 3-2, Pujuls is less likely to take a ball, slightly less likely to strike out, and hits the ball much better.

All hitters (2009):
Ball/Strike/SLGcontact
2-2: .40 / .24 / .466
3-2: .31 / .22 / .560
Exact same pattern.  It’s interesting that Pujols is no more likely to walk at 3-2 than the average hitter.

*

One way to think about this is to pretend that the 2-2 ball produces a walk, just like the 3-2 ball.  If it did, the 2-2 hitter would have about a 40-point OBP advantage over the 3-2 hitter (.510 vs. .470), but at a cost of about 90 points in SLG (.285 vs. .379).  That’s not a very good tradeoff.  So even if we neutralize the 3-2 hitter’s basic advantage—that an additional ball equals a BB—the 3-2 hitter still looks slightly better.  So I certainly don’t see any prima facie case that 3-2 hitters are making a mistake.

Tango, I don’t know.  I would think that perhaps that there is more of a spread than there should be, but I also think that virtually everyone makes the same mistake (maybe not to the same degree).  Pujols has also been taught from an early age not to get called out on strike 3, that “the pitch was too close to take” etc.  Heck, I would probably swing too often at 3-2.  It is human nature.  And what makes you think that Pujols is “smarter” than the average hitter.  He is an enormously talented hitter, but I don’t know that he is any smarter than the average player.  Unless you presume that to be that good, you have to have a minimum requisite “smartness” which may be true.

Another big factor: if I remember correctly discussions on umpires, they are more likely to expand the strike zone on full count, so borderline pitches are more likely to be called strikes. As for difference in SLG between 2-2 and 3-2, I think that standard training is to shorten the swing and try to foul off difficult strikes and borderline pitches in pitcher’s counts with two strikes. At 3-2 hitters become more aggressive. What are 2-2 and 3-2 splits for hitters like Howard and Fielder, who seem to want to hit HR at any count?