Gavin Floyd and Extreme Outcome Pitchers

Last week I was witness to a very impassioned internet debate that very shockingly drifted from a civil and reasoned exchange of opinions to a bizarre and vulgar bout of name-calling rather quickly. Before spiraling out of control, however, one particular grenade tossed in the fray caught my eye. It was alleged that both Gavin Floyd and Max Scherzer seem to have some shared brand of feast or famine quality to their performances-- that on any given day it seems plausible that either pitcher is just as likely to throw a complete game shutout as he is likely to get hammered for 6 or 7 runs before the 5th inning.

This intrigued me because, as a native Chicagoan, I've certainly heard plenty accounts of Floyd's all-or-nothing tendencies, and as a big fan of saber-sympathizer Max Scherzer, I've watched him throw an equal amount of gems as he has thrown duds. So I set out to create some type of "Dud to Gem Ratio" by which we could test the validity of these impressions. (I apologize if this has been attempted before, but the concept is not particularly google-friendly.)

I pulled up my database and ran a query using retrosheet files from 2007-2011 (since Floyd has been a White Sox). I defined a "gem" as any game where the pitcher throws at least 7 IP and allows no more than 1 run and a "dud" as any game in which the pitcher surrendered at least 5 runs in no more than 6 innings.

As it turns out, both Scherzer and Floyd have exhibited this dichotomy, although Floyd to a much higher degree. In fact, of the 42 pitchers who threw fairly equal amounts* of duds and gems within that timeframe, only three pitchers threw either one or the other more often than Floyd:

DUD/GEM RATIO 2007-2011

*where D/G > .8 and < 1.2. ExtOut% is DUD%+GEM%, min 50 GS. Here is the the full list of 42.

Many of the names here certainly have that same type of reputation as Floyd. They seem potentially just as dominant as they can be awful. Francisco Liriano at #2 really ought to be the poster-boy for this brand of pitcher and his profile is one that I suspect may facilitate these types of polarized, extreme outcomes. It's possible that when a pitcher is both a high-K and high-BB producer, he may skew strongly toward either the walk or strikeout tendency from game to game, polarizing the outcomes. Other pitchers with similar K/BB profiles like Daisuke Matsuzaka, Edinson Volquez, and A.J. Burnett all seem to fit this idea as well.

Interestingly, the exact opposite type of profile may have the same effect. Aaron Cook, well known for his perpetually low strikeout and walk rates, is the perfect example of this at #3. Cook's strikeouts, walks, and HBP's added up to just 18% of his plate appearances over those 5 seasons. That allows the very temperamental, untamed forces of BAbip to control the outcome for the remaining 82% of Cook's batters faced. His game-to-game fate, then, might best be described by that old quip from DIPS theorist Sam Elliot, "sometimes you eat the bar and sometimes the bar, well, he eats you."

Jason Vargas at #1 certainly seems to fit this mold as well-- his BIP% (including HR) over that time was a full 78.6%. And the same may be true for any of the other pitchers with a BIP% well above league-average-- the sinker-ballers, crafty-lefties, and 'pitch-to-contact' types:

But what is it about pitchers like Gavin Floyd, then? His BIP%, unlike most of the pitchers on this list, is regularly at or near the league-average. Is it just coincidence? Naturally, good pitchers will throw more gems, bad pitchers will throw more duds, and average pitchers will throw a bit of both. So if Floyd is in the middle of the pack on both duds and gems, maybe we shouldn't be so surprised. But perhaps that's not the whole story.

Floyd profiles as a slightly above average pitcher, both by ERA and FIP standards, with an ERA- of 92 and a FIP- of just 94 in that span. This makes him one of the better pitchers to achieve extreme outcomes so regularly. In fact, of the 30 pitchers in our first table, only Scherzer and Vazquez had posted better FIP- and ERA- during those seasons. Yet the percentage of Floyd's starts that have ended as either a gem or a dud is almost a full 10% more than both Scherzer and Vazquez-- so what gives?

Among those most similar to Floyd according to FIP- in that span, only one other pitcher similarly posted both duds and gems over 20% of the time-- Francisco Liriano:

The lazy analysis that gets tossed around here in the Chicago media is that Floyd has issues "between the ears". Of course, I can not say this is categorically not the case, but I'm more concerned with quantifiable explanations than pure WAGs. I seem to recall Ervin Santana, Edwin Jackson, and (certainly) Javier Vazquez all receiving the same criticism at one point or another, as well, and I'm curious as to whether there might be some measurable common quality among these pitchers.

Whether this is just randomness or not is a study for another time. But with both Floyd and Liriano now in the same rotation in a very close pennant race heading into the final days of August, Chicagoans may want to brace themselves for quite a few extreme summer evenings.

Follow @JDGentile on twitter for riveting statistical anecdotes and vain, self-obsessed assertions of his ravishing good-looks.