The inimitable Josh Tomlin: old dogs, new tricks, and revisited parades

I try to avoid declaring any one person unique, as such declarations imply there are non-unique people—boring and generic caricatures, leading rote and empty lives. This feels like a nasty thing to say, even only as an implication.

But Josh Tomlin is unique and you are leading a rote and empty life.

Let’s begin with some TOMLINY attributes.

• He gives up considerably more home runs than the average pitcher;
• He walks fewer batters than the average pitcher;
• He had Tommy John surgery to replace his UCL, and missed the 2013 season;
• He doesn’t strike many people out; perhaps because…
• He doesn’t really induce many swinging strikes; and yet…
• He’s maintained a lower than league-average batting average on balls in play;
• Because he walks so few, generally his K/BB ratio is well above average.

Now, none of those qualities is particularly interesting; there are lots of pitchers who fit several of these criteria.  But taken in concert, you’ve got a fairly interesting mix.  You also have a major league pitcher who can have a decent career in the back-end of a respectable rotation.  Start tweaking any of these characteristics even a little though, and the result becomes oddly dominant or entirely unpitchable.

For instance, Tomlin has allowed relatively few batted balls to become hits for his career. League average is about 31%, whereas Tomlin’s rate is around 27%.  If you allow something like 500 (fair, non-HR) batted balls over the course of a season, that 4% difference ends up accounting for 20 fewer hits—twenty times he magically turns a hit into an out.

Let’s do some back of the napkin calculations and try to convert that into runs.  We’ll be conservative and assume that those 20 batted balls, rather than the outs he actually recorded, would’ve become singles (rather than doubles or triples, we’re leaving out home runs, because they’re not balls in play and therefore get bucketed separately).  The difference in run expectancy between an out and a single is pretty easy to calculate; in fact, it’s been done using linear weights several different times. We know from that work that while an out reduces run expectancy by something like 0.29 runs, a single increases run expectancy by 0.46 runs.  So that’s a difference of 0.75 runs.  Multiply that by 20 (since he turned a single into an out 20 times over the course of the season) and all of a sudden Josh Tomlin lets up at least 15 fewer runs (because we assumed all singles) over the course of a season simply by maintaining a better than average BABiP.

Or look at his stinginess with walks.  For his career, Tomlin has issued walks at roughly 55% of the league average rate.  Over the course of a 200 IP season (700 batters faced, for round-ish numbers), that means he’s going to walk 27 fewer batters than an average pitcher.  Again, using linear weights to convert that to runs and we’re talking about nine additional runs Tomlin prevents simply by throwing more strikes than the average bear.¹

Now, 24 runs over the course of the season may not sound like a lot to you, but let’s work it out.  Again, let’s assume 200IP which is what I’ve been doing in this exercise.  Twenty-four additional (earned) runs would inflate a pitcher’s ERA by more than a full run (1.08).  That’s effectively the difference between great and good (2.50 and 3.50), or good and average (3.50 and 4.50), or average and David Huff (4.50 and ∞).  Josh Tomlin is basically cutting more than a run off his ERA simply by being so JOSH TOMLINY.

Which is really important, because if he didn’t do all that good TOMLINY stuff, he’d be basically unpitchable because of all the other bad TOMLINY stuff he does.  Josh Tomlin’s career ERA—even with all that fancy command/control/weak-contact-sauce—is still a fairly mediocre 4.73.  Add more than a full run to that, and we’re getting into DFA-land, where the Brett Myerses and Tomo Ohkas of the world go to die.

To put it another way, Josh Tomlin makes up for being really bad at some things (like strike outs and allowing home runs) by being outrageously good at other things (like balls on play and walks).  It’s weird.  It’s unique.  But then again, maybe you are too!

Up to this point, this entire post could’ve been written last year or the year before or the year before that.  None of this is new, exactly.  Josh Tomlin has always survived his faults by way of his virtues—there’s a lesson in there about life, but let’s not philosophize today.

What’s new is that in 2014 Tomlin has turned one of his glaring faults into a virtue, and has consequently pitched not like his normal back-end self, but like a legitimately good pitcher.  That one thing he’s done?  That fault he’s transformed?  He’s striking batters out at a slightly above league-average rate.  That’s it.  That change, and Josh Tomlin goes from being fringe-average to really good.

Before this season, Tomlin had faced 1,424 batters and struck out 188 of them—good for a K-rate of 13.2%.  League average hovers right around 20%; this was just one of those TOMLINY things we learned to live with.² You take the good with the bad, and Josh Tomlin has plenty of both.

But this season Tomlin has jumped his K-rate right past league average all the way up to 21.4%.  More than one out every five batters Tomlin has faced this year, he’s struck out.  That’s crazy.  He’s done this while still limiting his walks at an absurd level (3.1% for Tomlin vs 8.1% league average). He consequently has an ERA of 3.12 and a K/BB ratio that would rank fourth in the American League among starters were he qualified for such leaderboards.

He doesn’t qualify, of course, because he’s thrown only 40 innings, and anything can happen in 40 innings. This is the part where I tell you it’s all unsustainable, right?  Well, maybe.

When Tomlin was coming back from Tommy John surgery, what I worried most about was his command and the consequences even a slight drop-off there would wreak on his walk-rate and HR-rate.  After all, so much of the good TOMLINY stuff is directly correlated to his pinpoint control—inducing weak contact and limiting free passes.  If his ability to command the strikezone goes, he’s basically me with a funny haircut and a mouth full of chaw.  And I’m not very good.

But that didn’t happen at all.  Tomlin’s walk rate is actually a tick better than it is for his career, and his home run rate is still high, but not outlandishly so (1.34 HR/9 in 2014; 1.36 for his career).  The only glaring change is the strikeout rate—the one thing I expected to stay flat coming back from surgery.

Is Tomlin throwing harder?  Nope.  His fastball is averaging 89 miles-per-hour, basically identical to his career average.  Is he inducing more swinging strikes?  Perhaps, but he’s still below league average (7.9% career; 8.5% this season; league average 9.3%).  His peripheral stats look largely unchanged from his career averages, with the potential exception of opposing batters’ ability to make contact when swinging at pitches out of the strikezone (“O-Contact%”; 71% career, only 64% this season).  Is this a product of more perfectly placed curveballs in the dirt or the simple myopia consequent to a 40-inning sample?

His past performance coupled with my tendency toward cynicism would lead me to believe that Tomlin is simply not built to strike batters out at a league average rate—that before this sample gets too much bigger, he’s bound to come crashing back to earth with a seven-inning, one-strikeout game that realigns reality with our expectations.

The thing is though, that Josh Tomlin has been weird from the start.  He’s more TOMLINY than most, and for that reason alone I simply do not know what to do with him.  He breaks xFIP because of his home runs.  He breaks DIPS because of his batted balls. Even K/BB, my go-to stat for starting pitching, doesn’t quite do him justice because NO ONE IS MEANT TO WALK SO FEW BATTERS.

The first piece I ever wrote about him was called “Josh Tomlin and Rainy Day Parades” wherein I claimed that his great debut in Yankee stadium against a lineup of All-Stars was more or less a mirage, fueled by luck and grit.

That was almost four years ago.  He’s still lucky.  He’s still gritty.  And he’s still a reasonable starter in a reasonable rotation, whether he regresses or not.  I guess I’m just tired of betting against him.

1. There’s a little hand-waving going on here, to be sure.  It’s not like Tomlin is recording an out every time he doesn’t walk someone; sometimes he’s giving up hits.  Which is why I didn’t subtract the out value from the walk value; I simply penalized the average pitcher 27 times for the 0.33 run value of the walk—tada, nine runs! [↩]
2. This, and the home runs, of course. [↩]