Monday, January 20, 2014

Thoughts on K's and BB's

There was a recent article on Fangraphs by the Whistleblower From Seattle, Tony Blengino who now writes for the blog, entitled Quantifying the Effect of K and BB Rates on Offensive Production.  In it, I noticed something interesting that I believe probably had nothing at all to do with the point he was making.  The article itself had a lot of very busy tables that I found confusing and the point he was making was extremely difficult to follow, at least for my feeble mind.

Just to summarize, he arbitrarily took all hitters who changed teams so far this offseason and divided them into 4 groups depending on whether their K and/or BB rates were above or below league average.  The 4 groups were as follows:

1.  Low K, High BB.  We'll call this group the Sabermetric Dream Team.

2.  Low K, Low BB.  We'll call this group the Hackers

3.  High K, High BB.  This group is affectionately known in sabermetric circles as 3 True Outcomes.

4.  High K, Low BB.  We'll call this group Da Bums as they are widely hated by almost all in the sabermetric community.

Blengino then compared the average offensive performance of all 4 groups counting only balls in play + HR's vs all PA's.  Not surprisingly, and I'm not sure why he spent so much time on laborious numbers crunching to prove such a trivial point, the group that suffered the most from adding K's into their BIP performance were Da Bums!  OK, I think we already knew that based on both intuition and fairly simple mathematical estimating.

There was something else that Blengino should have noted from just standing back and looking at the names of the hitters in each grouping.  Here they are for your own perusal:

Sabermetric Dream Team:  Aoki, Cano, Choo, Doumit, Fielder, Kinsler, McCann, McLouth, Morneau, Logan Morrison, Schumaker.

Hackers:  Beltran, Rajai Davis, Mark Ellis, Ellsbury, Infante, David Lough, David Murphy, AJ Pierzynski.

3 True Outcomes:  Fowler, Freese, Ibanez, Kelly Johnson, Ruggiano, Salty, Seth Smith, Drew Stubbs, Mark Trumbo, Chris B. Young.

Da Bums:  Arencibia, Brandon Barnes, Marlon Byrd, Garrett Jones, Jhonny Peralta.

OK, I'll concede that there are a lot of good hitters on the Dream Team and not quite as many ratiowise on Da Bums, but the most striking thing to my eye is that there are significant numbers of good hitters and not-so-good hitters in all 4 groups!

What this tells me is that while K and BB ratios may help identify trends, there are multiple plate approaches that can be successful and what determines whether a hitter is good or not lies in some other measurement or equation than simply looking at K and BB numbers.


  1. Replies
    1. Wasn't listed, but I'm pretty sure he is a Bum, low BB's and high K's.

  2. Isn't it a closed loop, in that, if a good hitter (for sabers) is one with good sabermetrics, then, there are only good hitters in the Dream Team group?

    1. Well, maybe, but I don't think even the most ardent Saber would argue that Logan Morrison is a good hitter at this point.

    2. I think there is some natural amplification in the BB rates for good hitters in that good hitters, especially those with power, are going to get pitched around more and pitchers may be more careful about challenging them.

  3. I think you are correct, DrB, that there are good hitters in all the buckets.

    That's because, in my mind, each population is made up of a normal curve, and while sabermetrics have found that certain characteristics are good, there are those who defy those characteristics and perform well and others who have the good characteristics but perform poorly. So there should be good hitters in each group, but each group with have differing percentages of good players, aligned with the general rubric of less K's are better and more BB's are better. And thus the mean of their normal curve is better or worse, depending on the K% and BB%

    The difference, I believe, is that sabermetrics has been working on the DIPS assumption that hits are random luck that just happens, whereas I (and I believe I've seen you say this too) feel that not enough emphasis is placed on the ability to hit the ball well, that BA has been unfairly denigrated and thrown to the wayside when there is good value in being able to hit the ball for a basehit. Personally, I think that instead of inputting OBP and SLG into an offensive lineup calculator, it would be better to use BA, ISO-o, and ISO-p (or OBP and SLG), as I think the ability to hit is more valuable than those who can take a lot of walks but can't hit worth a lick.

    As I've seen in a number of sabermetric books, sometimes it is good to confirm rules that many people take as a given, just to verify that they are the rules we expect. Particularly as more and different types of data comes into the statistician's playbook.

    1. I pretty much agree with all of what you said there, ogc. Good summary.

  4. How does this make you feel about White Adrianza and his K/BB ratio?

    1. Sorry, I meant to say Ehire, not white. #autocorrect

    2. As I have always said, I think Adrianza has been underrated as a hitter and can become a good one in the majors with time and a little experience. On the other hand, the names in these categories remind us that there is more than K's and BB's in evaluating hitters.