clock menu more-arrow no yes

Filed under:

Normalized Strikeouts vs. Normalized Total Bases

New, 6 comments

So, Aubrey Huff is the man for the productive out.

Having already accomplished the walkoff groundout, our somewhat disappointing midseason acquisition did it again on Friday night.

With one out, the score tied and the amazingly clutch OP on second, Huff took an anemic hack, hit the ball all of 75 feet, and was out by six steps.

Not too good, except that the weak grounder he hit was in the direction of the first baseman, and enabled OP to take third. It was in fact, pretty darned good overall, and one of my keys to a game that had many.

I don't think you can overstate the importance of hitting the ball to the right side in that situation, and I don't think you can understate the damage a K there would have done.

Now, fairness compels me to state that the WPA contraption does not see a big difference between a man on second and a man on third with two outs there. And it says that the play cost us a little over 70 points of win probability.

But I say that's bullshit.

Look how many new ways for the Astros to win were opened up after Aubrey got Palmeiro over: A wild pitch. A passed ball. An error in the infield. A short single to left. A balk.

Sure Huff could have struck out and then Biggio could have doubled. Or Biggio could have gone 3 - 2 and THEN singled. Or OP could have taken third on a throw that escaped Molina, and then scored on Biggio's single. Hell, Biggio could have singled to the opposite field.

Or Junction Jack could have homered off the Conoco Pump.

But most likely, a Huff strikeout forces the Astros to get an extra base hit, or back-to-back singles. And that would have materially reduced the Astros' chances, in a big way, whatever the software says.

Alright. It wasn't my intention to argue with the Win Probability numbers when I started writing this. I assumed that they'd show what I thought to be intuitive. So I had to take a detour, and pardons.

But, yeah, my thinking after that whole sequence reinforced the bias I usually bring to the table: that strikeouts DO suck and that many times, being able to put the ball in play brings a distinct advantage. No matter what your basic Adam Dunn enthusiast might say.

In fact, I thought, probably going overboad in the full flush of a thrilling victory, perhaps the whole game can be reduced to the goal of reducing strikeouts while maximizing total bases.

It kinda made sense. Total bases and strikeouts are perhaps inversely related; freaks like Adam Dunn and Wade Boggs, instead of being outliers, are instead part of a main sequence, if you will, that expresses that relationship.

Well it sounded good. Having done the work, I see that at a certain point strikeouts become so frequent that they limit additional extra bases. There probably is no "main sequence," just a clump of hitters in the middle, and then some guys who rack up the whiffs or the extra bases at the edges.

Still, the graph I made is kind of pretty. First thing I knew is that I'd have to normalize the data, so the tendencies of someone like Biggio--with 10,000 at bats--could be fairly compared to someone like Jason Lane, with 1,000. So I didn't compare strikeouts and total bases, I compared strikeouts per 1000 at bats with total bases per 1000 at bats.

At first, I set a lower limit of 500 at bats, but that only provided me with a little over a 100 players, so I dropped it to 300 at bats, and that gave me 150 hitters that fit the bill, more or less, all time.

I got them all to fit, barely. I went ahead and colored the current players in red, and got this:

Click image to open largest version in new window
Something of a mess, yes, but when examined, there's a quick thumbnail sketch in every data point.

For example, knowing as we do that Moises Alou holds the team record for highest career batting average, doesn't his data point in the lower right hand corner of the graph mean he's the greatest hitter in team history, regardless of tenure?

And you may not see an easier way to compare the relative merits of Jeff Kent, Jeff Bagwell and Lance Berkman this year.

Immediately upon finishing this, it occurred to me that other scatter graphs of this sort might be illuminating. RBI chances converted vs. strikeouts is the one I wanna do now, maybe you can think of others. . . . .