Spec's use of ISO for minor league pitchers
Wending our way through the minors corn maze

.

Q.  Which of Spec's pets is being ISOlated on today?

A.  He's got a piece up on Kevin Hunter.  Check it out instanter.

.

Q.  He just ain't gonna leave well enough alone on this Pitcher-ISO thang.  Where does Dr. D stand on that?

A.  Well ... the last 12-18 months, it has been a hotly-debated topic around the 'net.  It has its detractors, and their basic protest is that low ISO is --- > driven by low HR/F, is --- > driven by luck.

This article, for example, pegs the correlation between ISO and HR/F as 0.69.  Meaning that seventy percent of a good pitcher ISO is being caused by fly balls caught on the warning track in the short term.

That's the case against.  It's got its issues, though.

.

Q.  And what is the case FOR using pitcher ISO?

A.  Well, BABIP in the short term *is* very often driven by luck ... and more to the point, by spotty defense.  Especially in the minors!

BABIP is driven mostly by singles, so ISO gives you the top shell of the Oreo cookie.  It's measuring "quality of hit."  In theory, that should work better in the minor leagues, right? 

"Frequency of hit" is highly suspect in the minors.  See Paxton, James.  The defense just isn't the same - the positioning is terrible, for instance.  So how about walking away from "Frequency of hit" and asking "Quality of Hit" out for a date?

.

Q.  CAN you measure "Quality of Hit"?

A.  If you were SABRMatt, you could saunter over and grab the Yankees' files on pitcher batted-ball MPH...

.

Q.  But if ISO is driven so much by HR-per-Flyball, don't you have to throw it out?

A.  Well ... suppose for the sake of argument that ISO is 70% luck.  You've still got a stat that is 30% pitcher ability to invoke weak contact, right?

Look, whatever the theoretical arguments, it is obviously possible for a major league pitcher to achieve a good ISO through skill.  Check Randy Johnson:

  Hitters faced ISO Remark
Unit, vs RH 15,000 138 Includes early and late years
Unit, vs LH 2,100 095  
MLB average   130  

Johnson allowed a very low ISO to left-hand hitters over the course of a 22-year big league career.  Even assuming that lefties put the ball in play, the ball wasn't going anywhere.

In fact, you can split out any pitcher, and he's going to have a platoon split on ISO.  So don't underrate ISO as a predictive stat.  You might emerge on the other side of a Kyle Hunter cornfield.

Another quick chart:

  Hitters faced ISO Remark
Felix, career 7,500 112  
Joe Saunders, career 5,700 162  
MLB average   130  

You could isolate these charts to compare apples-to-apples, GB% vs GB%, and you'd still find that Antony Vasquez gives up a whale of a high ISO.

.

Q.  But you like pitcher ISO even better, when used in the minors?

A.  The high-falutin' Hardball Times studies -- the ones that protest ISO is unstable -- are measuring the very best pitchers, MLB pitchers, in very stable conditions.  The differences they are trying to detect are subtle!

When Spec finds a pitcher in the lower minors, giving up a super-low ISO, he is working in an environment in which the differences are not so subtle.  Kyle Hunter may very well be overmatching the hitters.  If he is, ISO might be just the right place to look.

Comparing one ML pitcher's 120 ISO, to another one's 135 ISO, that's one thing.  But you get some minors pitcher with an 037 ISO, yeah, we wanna know about it.  It's like finding some switch-hitter at Cheney who had a .456 OBP last year, in 54 games (which one did).  Why wouldn't we want to know who it was?

Or finding some Cheney pitcher who gave up only 22 hits and 1 home run in 35 innings, leading the org in WHIP by a country mile.  Hope he gets his shoulder fixed right.

Golfclap,

Dr D

Comments

1

Someday I'll launch my crusade against trying to evaluate baseball by putting all of everything into one number.
I always evaluate ISO-against for pitchers in combination with things like K/9, BB/9, WHIP, HR%, XBH%, etc.
The more colored boxes the better!
Sure, the difference between .120 and .140 probably is without substance.  The difference between .120 and .050 probably is.
In other words, I tend to zero in on it when it is really low.
 

2

Like the idea of finding a "potentially" useful stat to leverage for gain.
Nice of Spec to clarify TWO critical points. It ain't the only stat he uses -- AND - he pays the most attention to the extreme results.
For a "high luck" stat, concentrating on the extremes suggests the at least "some" of the skew is based on ability. How much is debateable. But, it's at least a starting point.
But, no matter what stats you key in on - you shouldn't look at them in ... isolation. (ba dum bum)
For me, the second thing about viewing a high luck stat is - sustainability. When dealing with any high luck stat - (BABIP - or pitcher ISO) - I think one of the "proper" adjustments to make is to pay attention to performance OVER TIME.
Recognizing a stat is high-luck says one should not except EVERY abberation at tracking back to ability (good or bad). But ... extreme results suggest movement outside of the luck box - and "sustained" skew also suggests movement outside of the luck box.

Add comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd><p><br>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

shout_filter

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.
  • Web page addresses and e-mail addresses turn into links automatically.