THE BOOK--Playing The Percentages In Baseball

Yes, good stuff.

Some people are surprised at the results because they assumed that good hitters get fewer fastballs because they are more likely to crush those fastballs!  That is not true! It could be that better hitters are better fastball hitters than poorer hitters, but that is probably not true.  In fact, it is probably just as likely (maybe even more likely) for weaker hitters to be better fastball hitters than better hitters.  If THAT is true, then we should see fewer fastballs to weaker hitters, right?  Wrong!  Here is why:

The number of fastballs seen by batters is really a proxy for the percent of pitches intended to be thrown in the strike zone (or to a particular location actually - and that location might not be in the strike zone).  IOW, an offs-peed pitch, especially a curve ball, is hard to throw to where you want.  As well, when you throw an off speed pitch, it is usually more important to keep it in a certain “hard to hit” location, usually low. For example, if you want to throw a change-up or a splitter, you better throw it low, or else it is a horrible pitch.  Not so with a good fastball.  If you miss your location with your fastball, you have more leeway.

That is the reason why pitchers throw more off-speed pitches in pitcher’s counts even though the batter knows they are more likely to be thrown. Because when you try and throw an off-speed pitch to a corner (say a curve ball down and away), you will completely miss the strike zone a lot, but since it is a pitcher’s count, it doesn’t hurt you that much.  That is the whole idea of a waste pitch, right?

So, given this, why do poorer hitters see more fastballs, but not in all counts?  Because you don’t want to walk the poor hitters!  Because if you throw more hittable pitches, you don’t get hurt as often or as much.  That is the ONLY reason why poorer hitters see more fastballs overall and especially at hitter’s counts.  At hitter’s counts, you are in danger of walking the batter if you don’t throw a pitch in the strike zone.  With a poor hitter at the plate, the last thing in the world you want to do is to walk him.  So in hitter’s counts, you throw him lots more fastballs in order not to walk him.

Additionally, the batter is often ready to “tee off” on a fastball in the middle of the strike zone, and might lay off other pitches.  If the poor hitter tees off on your fastball down the middle, it is not that big a deal.  What is the worst thing that Juan Pierre, Jason Kendall, or David Eckstein can do with a fastball down the middle?

Now, what about in pitcher’ counts?  Why do we see better hitters getting MORE fastballs?  Well, the control thing is out the window now.  In a pitcher’s count, the pitcher is not much concerned with walking the batter and the difference between a ball and strike is not all that great.  He is most concerned with making a good pitch on the corner or out of the zone.  Obviously he throws more off-speed pitches in pitcher’s counts to all hitters.

But why more off-speeds to the poorer hitters? Because one reason that they are poor hitters is that they have poor strike recognition skills.  So they are more likely to swing at that curve ball or change-up in the dirt!

So everything makes sense if we remember two things:  One, you don’t want to walk the poor hitters, and two, poor hitters will swing at more off-speed pitches out of the zone, especially in pitcher’s counts.

Now, rather than using wOBA to break the players down into two groups, Dave should have used SLG or something like that.  He should have eliminated the walks.  What he will find will be a larger difference in the number of fastballs seen and the locations of pitches between the two groups once the walks are eliminated from the grouping criteria.

For example, players who have little power but get a lot of walks and are therefore in the “good hitter” category will NOT see fewer fastballs. Those players will be pitched like a poor hitter, with more fastballs and more pitches in the strike zone.  In fact those players will see more fastballs and more strikes in the zone then good hitters who are power hitters.

How many fastballs a batter sees (ignoring whether they are good or bad fastball hitters in general) is a function NOT of how good a hitter is overall as measured by wOBA or OPS, but how good a slugger they are - IOW, when a pitch is in zone, what do they do with it?

Again, the percentage of fastballs groups of players get is a proxy for whether you want to pitch them in the zone or out of the zone in general - IOW, how dangerous they are when they swing at a pitch in the zone (essentially SLG) and their pitch recognition skills.  It has nothing to do with whether good or bad hitters are good at hitting fastballs or curveballs in general.

Is anybody else intrigued that bad hitters only see 55% of pitches in the strike zone on 3-0 count, considering that they take in those situations like 92 or 94% of the time?

Why the heck would they see more pitches in the strike zone on 3-1?

If you look at the other groupings, they all make sense in this respect.

The 3-0 count makes no sense, as does the 2-0 count to good hitters.

Dave: did you remove IBB?

Unless he didn’t remove IBB’s, and assuming that his numbers are accurate, I would assume that it is due to the intentional unintentional BB’s.  Even bad hitters get those sometimes.

MGL: cool.  So, if he were to show the breakdown based on bases being empty and/or a runner on 1B, then we should see drastically different numbers?

I would imagine there is also a potentially big selective sampling issue in the pool of pitchers who go 3-0 to a bad hitter.  That scenario (3-0 to a bad hitter) probably happens disproportionately when pitchers with poor control are on the mound, and possibly also when the pitcher is tired or hurting or struggling with his mechanics and less likely to throw strikes.

Really cool article, by the way.  Along with MGL’s post, that’s some very enlightening info.

Ok, so we really would like to split it up by good/bad pitchers, by count and with bases empty.

Actually, since the run value of the walk changes so much, let’s look for those situations where the run value is so high, relative to the other events, that the idea of an unintentional intentional walk is not really feasible.

I published the boLI numbers for each event a while back (base-out Leverage Index).

Pitchers AND batters.

Dave: did you remove IBB?

Cool call Tango, I forgot to do that.  Here are the correct zone numbers with IBBs removed.

        top   bottom
0-0   0.514    0.550
0-1   0.428    0.473
0-2   0.325    0.325

1-0   0.522    0.580
1-1   0.478    0.526
1-2   0.376    0.424

2-0   0.548    0.607
2-1   0.545    0.580
2-2   0.443    0.489

3-0   0.578    0.598
3-1   0.607    0.646
3-2   0.553    0.598

The numbers increase, but not drastically, and 3-0 is still lower than 3-1.  It must be MGL’s suggested intentional unintentional BBs.  The n for that, bad hitters in 3-0 counts, is 487.

Good stuff. 

Ok, we now have all the numbers looking alot more normal, and the 3-0 count is just a little bit away from being correct.  So, yeah, the unintentional-intentional walk might be the culprit.

If we go to post 1 here:
http://www.insidethebook.com/ee/index.php/site/comments/leverage_index_by_base_out_states/

The least likely base-out state to get that kind of walk is any 0 out situations.

On top of that, is men on 1B, 1 out.

And bases loaded, any outs.

And 1B+2B, 2 outs.

So, Dave, if you were to separate the situations based on that, we might have an answer that makes more sense.

Or, if you want to look at the opposite, any 2-out situation with 0 or 1 runner on base, or 1B open is where you should find almost all of these UIBB.