THE BOOK--Playing The Percentages In Baseball

Friday, September 17, 2010

K’s

A writer asked me about K’s and Mark Reynolds:

He acknowledged that in certain situations - for example, a runner on third and less than two outs or a man on second and nobody out - a strikeout can be costly. But when looking at the 24 possible situations based on outs and bases occupied and weighing them according to how often those situations occur, Tango said a strikeout costs an extra 0.01 runs compared with a ground out or fly out.

“So if you have Mark Reynolds with 200 strikeouts, and given that the average hitter would strike out 100 times, Reynolds’ strikeouts cost one run,” he wrote in an e-mail. “I know it doesn’t ‘feel’ like it, and I know it’s incredibly frustrating to see a hitter strike out in clutch situations, since fans predominantly simply want the batter to put the damn bat on the damn ball.

“But anyone who sits down and works it out has always come to a very similar conclusion.”

Walt Davis says:

Batting average is the single most important component of offensive production and Ks hurt BA. Walks and power can make up for that damage (and then some) but let’s not pretend that Ks don’t hurt.

Strikeouts is not something that you can just reduce in a vacuum.  If you look at non-K performance, and you reduce your K, you will find that your non-K performance will also go down.  It would be something like this:
200 PA, strikeouts, , -0.30 runs per PA, -60 runs total
500 PA, nonK, +0.15 runs per PA, +75 runs total

Total?  700 PA, +15 runs.

Let’s say now that you reduce your K by 100.  This is what you will probably get:
100 PA, strikeouts, , -0.30 runs per PA, -30 runs total
600 PA, nonK, +0.05runs per PA, +30 runs total

Total?  700 PA, +0 runs.

Basically, in order to be really good when you don’t strikeout, you have to have a swing and approach that let’s you strikeout alot as a byproduct.  If you shorten your swing or change your approach such that you can reduce your K, you will also reduce your effectiveness when you don’t K.

Is this really what happens?  Well, for the majority of hitters, this is what would happen.  I have to believe that hitters are smart and have reached the optimal style for their talents.

But, for others, likely a minority, this is not what happens, as evidenced by this study I did a while ago:

And how about overall? The wOBA of the 5250 went from .345 to .344. The Kleskoyan increased from .340 to .354, while the Jedmonds dropped from .354 to .347. So, even though the Kleskoyan actually hit less HR, their lowering of strikeouts actually increased their overall production. And even though Jedmonds hit more HR, their increase in strikeouts lowered their overall production.

The Kleskoyan group of hitters are those hitters that did something different that allowed their K rates to plummet.  And their overall production increased.  The Jedmonds group of hitters are those that allowed their K rates to jump, and their overall production dropped.

When you do these studies, you are always at the mercy of selection bias: are the players in these groups representative or not?  Are the players that didn’t qualify for the study biased or not?  Suppose that you have players that actively changed their approach such that they reduced their strikeouts, but their overall production dropped substantially.  What would happen?  Well, they wouldn’t last long enough to be in the study, and therefore, we won’t see their effects.  You can possibly make the same argument for the guys that decide to become longballers and increase their K as well.  It’s hard to determine exactly where the bias is.

So, Walt is definitely right that it’s very possible that a reduction in K can improve your overall game, as evidenced by my study.  As I concluded in that article:

While I wouldn’t propose a star hitter like Dunn or Soriano change his approach, there are plenty of guys struggling to make it who would find it worthwhile to change their approach to hitting.

The testing of theories to changing approaches should be done on the hundreds of high K hitters struggling to make an MLB team.  I would not test those theories on Soriano or Reynolds or Dunn, not until I see results on the lesser players first.