MASTER NOTES: Finding the Common Denominator

Like you, at this time of year, I start firing up my spreadsheets looking for some leading indicators of potential value that is somehow going untapped. I feel like Billy Beane in Moneyball, only nobody is making a movie about me. And unlike past years, I came to a conclusion very quickly when I got rolling this season. The conclusion is that I don’t get why our pitcher metrics have so many different denominators.

Take a look. Dom and Ctl use per nine innings. GB-FB-LD percentages use per ball-in-play, counting home runs, but BABIP uses per ball-in-play, not including home runs. Infield fly balls are presented as a percentage of fly balls. And K% and BB% use Total Batters Faced (TBF).

So as I started my 2017 pitcher research, I wondered if there might be some way to set all our pitching metrics to one denominator, and if doing so might open the way to better and more transparent cumulative pitcher metrics. I thought about per-inning, but then I thought that an “inning” is really just a collection of outs, so what about eliminating the extra calculation and use per-out?

The problem with that was it it’s still just basically per-inning (or per nine), just with a different constant. After some other ideas, I settled on TBF, which is wayyyy out there in column “CC” in the BIS Excel files.

TBF has a few in-built advantages. First, it captures the basic transaction in a game: pitcher-versus-batter. Second, it is relatively easy to convert all the other ratios. Third, this adaptability might let us add up TBF-based stats to derive more revelatory combinations. Fourth, it has a direct batter analog in Plate Appearances, which eventually might allow us to have a common ratio denominator for pitchers and hitters. And fifth, “TBF” is really easy to spell.

Before I get to some of the results of using TBF for ratios, some level-setting. A top starter will usually amass 850+ TBF in a season (and keep that number in mind the next time someone tells you that pitchers don’t do as much as hitters). The leaders in total TBF 2014-16 are David Price at 2,848, Johnny Cueto at 2,708, Max Scherzer at 2,705 and Corey Kluber with 2,697.

Ace relievers are around 250-300 TBF. In 2016, prized relievers like Dellin Betances, Roberto Osuna, Ken Giles, Sam Dyson, Kelvin Herrera, AJ Ramos, Andrew Miller, Mark Melancon, David Robertson and Cody Allen were all in that 250-300 range.

So what happens when we start applying various established ratios, replacing their current denominators with TBF? I first tried K% (of TBF, as all these examples will be) versus K/9. After ranking all 140 400+ TBF starters in both categories, I found general overlap up and down the rankings, but...

> In the Top-10s, Michael Pineda, Danny Salazar and Juan Nicasio fell out of the Top-10, replaced by Clayton Kershaw (#3, behind Jose Fernandez and Yu Darvish), Justin Verlander and Vince Velasquez. In a trade, I’d take the Kershaw group.

> In the 11-20 cohort, Corey Kluber, Danny Duffy and Chris Sale moved up, at the expense of Aaron Nola and Nathan Karns. And Chris Young, #20 by K/9, falls to #32 by K%.

> In the 21-30 rankings cohort, Carlos Rodon, Francisco Liriano, Scott Kazmir and Archie Bradley fall out, replaced by Jon Lester, John Lackey, Jake Arrieta, Jacob deGrom, and Steven Matz.

And so on. The cream seems to rise, sometimes just a little, but to rise all the same toward the top.

Since there were no obvious shortcomings to the TBF-based approach, I next moved into some uncharted waters. For a while now, I’ve been mulling over pitcher outcomes, and how to possibly combine them into a broad metric for pitcher effectiveness.

So I amassed all the stats for all pitchers with 200+ TBF from 2014-16 to see what kinds of contact hardness/trajectory combinations had the most favorable outcomes, as measured by the lowest Hit Rates (H%, known outside of HQ as BABIP):

        Ground Balls   |    Fly Balls   |    Line Drives
       Sft   Med   Hrd |  S    M     H  |   S     M     H
=======================|================|=================
2016   14%   23%   51% |  8%   7%   49% |  66%   73%   64%
2015   15%   23%   52% |  8%   7%   48% |  64%   71%   65%
2014   14%   23%   53% |  7%   8%   45% |  64%   72%   66%

Two columns jumped off the screen at me: The Soft and Medium FB have H% under 10%. I had suspected that there would be a discrepancy between Hard FBs and the others, but not to this extent. But these metrics are all based on balls-in-play (minus HRs), rather than on TBF. So I calculated the percentages of soft- and medium-hit fly balls for each pitcher as a percentage of his TBF and in the usual way as a percentage of balls-in-play (not including HR).

The results were again pretty expected. Most pitchers were within 10 ranking slots in one method compared with the other. I did notice that most of the big gainers were premium pitchers: Scherzer, Bumgarner, Stephen Strasburg, Darvish, while the bigger droppers were lesser lights like Jordan Zimmermann, Derek Holland, Hisashi Iwakuma and Bartolo Colon.

Even if the results had been closer, I still think this would be progress. Besides moving better pitchers up the rankings, the common denominator of TBF will let us combine the previously un-combine-able, K rate and what we might call Weak-Hit Rate, to find pitchers who don’t give up a lot of balls-in-play, and limit the damage when they do by inducing more of those relatively harmless soft- and medium fly balls, which turn into outs 90+ per cent of the time (and which probably don’t often advance baserunners, either).

In the combined metric, the top pitchers are... Oh, gosh, I’ve hit the 1,000-word mark for this week. Tune in again next week, when we’ll dig into this idea a little more.


Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.