Introducing K% – BB% – ISO

I often read articles that say strikeouts are bad if you don’t have power. Inspired by the success of K%-BB% with pitchers I tried to do something similar with hitters to generate a stat that gets predictable with a smaller sample size than wRC+ due to the elimination of BABIP. This could be useful for prospect analysis or also early-season stats.

The rationale basically was to take something bad (Ks) and subtract good things (ISO, BB). To do this, I first scaled BB and K from percent to decimal dividing it by 100. So instead of 22% Ks I would use .22 to get it to the same scale as ISO. You could also scale ISO to percent, but it does not really matter.

Doing that, I found out that most good hitters were below zero. I looked at players that had at least 1000 PA from 2014 to April 2017.

Here is the complete document.

The worst K-BB-ISO was a positive .133 by Chris Johnson (75 wRC+) while the best was a negative .256 by David Ortiz. The average was a negative .05. The 25th percentile was negative .012 (Rajai Davis, 95 wRC+) while the 75th percentile was a negative .09 (Charlie Blackmon, 110 wRC+). Based on this, I conclude that good values are something like negative .1 or better, while values that approach zero are bad and positive values are atrocious.

Overall, the Pearson Coefficient between wRC+ and K-BB-ISO was a negative .75. A negative correlation is expected because the good values are below zero, and the correlation is significant.

The top 20 in K-BB-ISO all have a wRC+ above or equal to 120 and are ranked in the top 50 in wRC+. In the bottom 20 there are three hitters with a wRC+ slightly above 100 but most are near the bottom of the leaderboard.

Now BABIP is not random and there is a skill that is related to contact quality, but then again ISO is also related to contact quality — the guys who hit the ball hard and at decent angles usually also have good ISOs, while the put-everything-in-play-weakly guys usually have bad ISOs (and often bad BB%).

So here is what I look at in a prospect:

excellent: <-0.15 (expect 120 or better wRC+)

good: -0.08 to -0.14 (expect 105 to 120 wRC+)

OK: -0.03 to -0.07 (expect 90 to 105 wRC+)

red flag: above -0.03

Now there is a disclaimer to this: The K-BB-ISO might underrate ground-ball-heavy hard hitters who have lower ISOs but generally solid contact quality. For example, Christian Yelich is just 146th out of 246 in K-BB-ISO over that time frame but 45th in wRC+. It might also overrate fly-ball-heavy pull hitters with high pop-up rates. Examples of this are Brian Dozier (68th wRC+, 29th K-BB-ISO) or Jose Bautista (17th wRC+, 3rd K-BB-ISO).

Also you have to consider park and league factors as there are some very hitter-friendly leagues and parks in the minors (for example the PCL) and HR/FB luck also needs to be considered.

But overall, the leaderboards look quite similar and K-BB-ISO might be a good indicator for success if you want to eliminate BABIP from the equation. Basically this is pretty simple — if you don’t walk or slug a lot, you better not strike out. And if you strike out a lot, you better have something to make up for it.

My analytical background is not the best, though, and maybe somebody who has a little more skill in that field could look at the data and see if I’m onto something.

Newest Most Voted
Inline Feedbacks
View all comments
6 years ago

You could potentially add pop out rate to bring the Yeliches and Bautistas back to expected levels.

6 years ago
Reply to  Dominikk85

Hey Dominik, spielst du selber auch? Ich spiel selber in der 1.Bundesliga und wollte mal fragen, wer du bist.

Brad Oremland
6 years ago

This is really interesting – I like it. But the correlation between negative values and positive wRC+ is confusing.

It might be more intuitive to reconfigure the formula as: ISO + BB% – K%

Same idea, but then higher values would be better.