THE BOOK--Playing The Percentages In Baseball

I don’t think you’re wrong.  I made a similar point a few years back when Oakland was getting all the sabermetric press and the Angels were winning the world series.

The Angels didn’t have a great rotation in 2002, though the pitchers generally had good years, they weren’t impressive with the stuff, getting kind of lucky on the way, and most of them fell apart in 2003.  But the bullpen was tremendous, with Shields, Donnelly, Percival, and K-Rod when the playoffs came around.  I wanted the pitch counts to go high, so the bullpen would come in sooner.  It’s only a bad thing when your bullpen is overworked, and the other team is getting to face the 11th or 12th guy on the staff.

As for getting the other team’s pitch counts up, the best way to do this is not try and get a tiny advantage in pitches thrown per PA, but stop making outs - send more people to the plate.  An optimal approach, of course.

Well, of course, the real reason to take pitches is to not swing at less than ideal ones, increasing the likelihood of getting the batter a pitch he can more readily drive somewhere.  Almost as important, if the batter doesn’t find a pitch to his liking, he increases the chance of the free pass.

I hear what you are saying, but the same exact process of ‘working the pitcher’ is a GREAT idea from the batters perspective, so this is really kind of a moot point.

MGL, click on my name for a study I did a few months ago on the subject of pitcher fatigue.  Times through the order, once you control for pitch count, does have an effect in that it reduces the chances of walks and strikeouts and increases the chances that an out will be made on a ball in play.  In this context, it’s a pretty good guess that if a pitcher has been made to throw a lot of pitches early by the opposing offense, he’ll probably change his approach to conserve his pitch count and not fool around so much with his pitches.  That leads to more balls in play.  Humans are, of course, not static creatures.

Also, there are cases in which it’s better to try to get your batters not to swing so much.  Let’s take the absurd example of a batter who swings at everything, and unlike Vlad Guerrero isn’t getting much accomplished with all the swinging.  Some of the stuff I’ve done on signal detection theory and the strike zone shows that some hitters actually swing too much and would be better off (in terms of minimizing the number of strikes) swinging less.  And some should swing more.

I disagree with the thought that all batters use the optimal approach at all times.  As a therapist, I’ve seen far too many instances where people don’t select the optimal approach to all sorts of things in life.  (Know anyone who smokes?) If batters always use the optimal approach, why have coaches?

It seems that the “take a lot of pitches” strategy could become an example of game theory quickly.  If the pitcher sees that the batters are taking pitches, the obvious thing would be to throw a lot of strikes.

The point of the strategy seems to be “let’s see if we can get to the ‘pen in the 5th or 6th inning when the other manager is forced to use the replacement level relievers, or alternately, the better relievers for longer stints.” (Does anyone know what the relative value is of a facing a lousy reliever for an inning vs. an average starter?) Spread the other team thin enough and they might break. 

The problem with the “take a lot of pitches” strategy is that it might wear down an opponent’s bullpen a bit, but chances are that the beneficiaries are going to be the next team that the opponent faces.

I disagree with the thought that all batters use the optimal approach at all times.

That was not what I said, although that is what I said. You did not read and/or think through my comments carefully enough.

I said that we “have to assume” that batters approach their PA optimally, even if they don’t (of course all batters don’t).  Because if some batters don’t and you want them to take more pitches in order to increase their offensive production, that has nothing to do with making the pitcher throw more pitches.  In addition, we are talking about an organizational philosophy.  It would be ridiculous to assume that all batters in baseball, as a whole or on the average, take too few pitches, in terms of their offensive production.  And again, even if that were true, then I would teach batters to take more pitches to increase their production, not to make the opposing pitcher throw more pitches, although I suppose I could give any reason I wanted to.

Times through the order, once you control for pitch count, does have an effect in that it reduces the chances of walks and strikeouts and increases the chances that an out will be made on a ball in play.

I just said that time through the order of course affects pitching performance (we showed that in the Book), but taking pitches does NOT change time through the order.  So why are you just repeating what I said and suggesting that refutes my argument?  I don’t understand that.

The point of the strategy seems to be “let’s see if we can get to the ‘pen in the 5th or 6th inning when the other manager is forced to use the replacement level relievers, or alternately, the better relievers for longer stints.”

That is ridiculous.  Forcing a pitcher to throw 3 extra pitches per outing (or whatever it is) is NOT going to force the other team, on the average, to bring in their worst reliever as opposed to an average reliever if you didn’t take so many pitches!  And a replacement level reliever is about the same as an average starter anyway (take the worst of the worst starters, who are 1 run worse than average, and there are a ton of these available, and turn them into relievers and all of a sudden they are average pitchers!).

Here is a study for someone.  Look at average relief performance for the rest of the game when a reliever comes in after 100 pitches versus when a reliever comes in at 105 pitches, or some such thing.  I think you will find a negligible difference.

I still find it funny that a lot of hitters don’t even try to hit the ball with the edges of the grain side of the bat. IMO, I think it is better to see more pitches before hitting the ball. However, taking a fat pitch is a mistake. That’s how I feel when I go up to bat. I want to get a good pitch to hit, but if it is not there, you have to take the walk.
However, three batters coming up hacking could make your half of the inning only 5-6 pitches. Your team’s pitcher should have a longer rest than that I bet. It absolutely sucks having to go out there after a long half inning and go back out in only a couple of minutes.

Well, I think there are a few points/reasons that you failed to take into account, mgl…

First off, you’re assuming that a team’s middle relief is actually better than the starters, which is not necessarily the case.

For instance, let’s say you’re playing the Mets and Johan Santana (2.85 ERA) is on the hill. If you can get him to 100 pitches through 6 innings, then you probably get to face Aaron Heilman (5.34), Pedro Feliciano (3.61), Joe Smith (4.02), Duaner Sanchez (3.75), or Scott Schoeneweis (3.13) int he 7th and 8th. Obviously, you would rather face any of those pitchers instead of Johan Santana.

That’s just one example...I’m sure there are plenty of other teams with weaker middle relief that a team would be thrilled to bat against. Or, perhaps you could be in a situation where a team has one good middle reliever, but he’s pitched three straight days, or 4 out of 5, and you’re fairly certain the opposing manager won’t bring him in, opting for a lesser pitcher instead.

Secondly, and in my opinion more importantly, is the fact that the more pitches you take, the less likely you will swing at a bad pitch. This point was already made by Lou in Comment #2. As a Yankee fan, watching Robinson Cano and Melky Cabrera pop-up neck-high fastballs and swing and miss at sliders two feet out of the strike zone frustrates me to no end. Of course, there are opposite examples too, as someone like Jack Cust probably takes too many pitches in the strike zone.

I think I once heard a quote by Manny Ramirez about his hitting approach and he said “If I think it’s a strike, I swing. If I think it’s a ball, I don’t”. Brilliant in its simplicity

Lastly, walks are really good for an offense. If a pitcher throws you four balls out of the zone and you don’t swing at any of them, you’ll reach base 100% of the time (assuming of course that the ump calls them correctly). Whereas, if you swing and make contact, you’ll probably make an out roughly 70% of the time. If we all agree that having a high OBP is better than having a low OBP, you need to draw a lot of walks to really be among the league leaders in OBP.

Take Juan Pierre as an example...his career .300 BA is 25 points above the NL average in that time, but his low walk rate gives him a roughly league-average OBP (it’s actually 1-point below). So, despite the fact that when he puts the ball in play he gets a lot of hits, his low walk-rate causes him to only be average at getting on-base.

Cano is another example...career .303 BA, .336 OBP. His impatience prevents him from having a higher OBP, which actually hurts the Yankees offense.

One last thing that I’m not sure has been mentioned is that although it might be the case that on a game-by-game basis, there is little advantage in “getting to the bullpen”, in a 7 game series there clearly seems to be an advantage.

Bullpens do not last forever, and a pen that had to pitch 4 innings per game (pitcher leaves in 5th) vs. 2 innings per game (pitcher leaves in 7th), over a 7 game series, would have to pitch 14 more innings. In a 7 game series, that’s 28 innings of bullpen.

That’s a lot of innings.

Will some of these innings be eaten up by possibly the best pitchers on their team? Sure. The closer and the primary set-up man will get many of these innings. But that’s still a lot of innings left over to be pitched by bad pitchers.

Dan and everyone else, if you preach more patience to your hitters in order to get the starter to throw more pitches, and you are pretty sure you are not changing their optimum approach - in fact you might even be making it more optimum accidentally, how many extra pitches do you think you are adding and how many innings per game do you think you are adding to the bullpen.

You can’t support an argument with evidence that is ridiculously exaggerated.  For example, you can’t say, “Yes, it is a good idea to preach patience as an organizational philosophy because if you can get the starter out 2 inning earlier, you create an advantage.” Well that might be the case, but you are not getting the starter out 2 innings earlier.  You are getting him out, what maybe, 1/4 inning our earlier?

I think the reality of the situation is that lets say the average hitter sees 3.7 pitches per at-bat and the manager tells the team, you need to all go up there taking more pitches.  In reality the “best” hitters take 4.1 pitches per plate appearance.  Using 100 pitches as a metric that means if everyone is average the pitcher will face 27 batters.  If all are exceptional, 24 batters.  If all three of those extra batters are exceptional that amounts to 12 extra pitches.  Keep in mind this is the ideal case from the stand point of “we need to take more pitches”. 

The reality of the situation is that on average you would be lucky to get half of your batters to perform at that level (4.1P/PA), so we are talking 6 extra pitches per game.  At the end of the day that is a mere fraction of an inning and I don’t see that fraction being a problem.

Remember this is from the pitchers and mostly managers perspective.

Right, that’s MGL’s basic point. 

***

Let’s look at a recent example where the hitters were told to be more patient exactly to get the pitcher out: Joba went into the rotation, everyone knew that he was on a limited pitch count (it was even announced).

His first start, against the Jays, he threw 62 pitches to 12 batters (5.2 pitches per batter!).  His next start, 78 pitches to 19 batters (presumably taking alot as well, for an average of 4.1 pitches per batter).  Then, 89 out of 26 in his 3rd start for an average of 3.4 (shouldn’t the Astros have been taking too?).

After that, he never went below 90 pitches: 903 pitches thrown to 217 batters (4.16 pitches per batter).

It’s more likely that his first start was a combination of the Jays taking, and him not hitting the strike zone (32 strikes, 30 balls).  The first three starts taken together is 229 pitches to 57 batters (average of 4.0 pitches per batter).

So, even if you are telling them “take!”, we really don’t know if that makes any sense.  A pitcher, sensing that you are taking more than you should, will simply throw more strikes.

I don’t think it’s a feasible approach, except for these rare circumstances, like a pitcher who hasn’t started in a long time or coming off an injury.  A pitcher in his normal command mode should be able to exploit hitters who are not approaching the PA in a way that suggests they will take for the sake of taking.  Especially if you are taking strikes.  That will DECREASE pitch counts more than anything.

That is a very good point - that if batters start taking more pitchers, pitchers will simply throw more strikes. All getting back to game theory.  But one of my main points still holds, I think.  That is, if batters change their approach for any reason, one of two things has to happen.  One, they increase their production, or two, they decrease their production, since there is only one “optimal” approach to an AB.  If taking more pitches in order to pump up pitch counts increases production, then the discussion is over.  They should be doing that anyway (regardless of whether it increases pitch counts or not).  If it decreases production then even if you gain a little by pumping up pitch counts, you lose back some or all (and perhaps then some) of that advantage in a sub-optimal approach.

I think no matter how you look at it, it is NOT a good approach to encourage batters to change their batting style in order to try and bump up pitch counts.  Now whether certain batters should do that anyway is another story.

Sometimes “sabermetrics” screws things up, especially when in the wrong hands (e.g., Billy Beane).  If you also notice, the Red Sox eschew the sac bunt attempt to a fault.  I assume that is because James thinks that the sac bunt is almost always wrong - conventional sabermetric wisdom. He is not aware of some of the cutting edge research that indicates that the sac bunt attempt, if used properly, is a viable weapon.