Rubbing Mud: Rubbing N... (05/27)
the archives are now free.
1 comment has been left for this article. BP Comment Quick Links
| |
|
|
May 29, 2015 Rubbing MudContext is King
Danny Salazar had a rough night in Texas on Tuesday. In 5 2/3 innings, he gave up three runs (all on a Prince Fielder home run), struck out six, walked three, and hit a batter. In eight starts, he has a 3.65 ERA, a 3.50 FIP, and a 3.56 DRA—the stats of a perfectly good starting pitcher, but not those of an ace. As we always say, though, ERA is a flawed statistic. It does nothing to remove the vagaries of batted-ball luck, nor to adjust for park factors. It also makes arbitrary, often inscrutable judgments about which runs count against the pitcher, and which ones don’t. Some of those things are corrected for by FIP, which takes as its inputs only strikeouts, walks, hit batters and home runs, and thereby mostly strips defense and batted-ball luck out of the equation. Sometimes, though, batted-ball luck isn’t luck, and a pitcher is exercising some influence over the efficacy of his defense. Unfortunately, FIP doesn’t capture that. It also doesn’t correct for the parks in which those home runs are hit. Deserved Run Average does. We introduced it as the new king of descriptive pitching statistics last month, and I, for one, stand by that. DRA accounts for the defense behind a pitcher in every plate appearance. It accounts for the umpire behind the plate, and for the framing skills (or lack thereof) of the catcher. It accounts for the ballpark in which every plate appearance takes place. Crucially, it accounts for the batter’s role in determining the outcome of each plate appearance, so giving up a hit to Miguel Cabrera is (as is proper) treated differently than giving up the same hit to Miguel Olivo. And it accounts for a pitcher’s contribution to slowing the opponent’s running game. It would be virtually impossible to better estimate the pitcher’s real role in run prevention with a simple statistic. DRA does have one failing, though: it’s not a predictive stat. That much is spelled out in the conclusion section of the introductory piece you (surely) clicked over and scanned through, above. DRA removes the effect of other actors and major factors like park and league offensive levels, but it doesn’t necessarily quiet the rest of the noise that makes the pitcher’s true, predictive talent level so hard to pinpoint. We have another stat that does those things better, though. Its predictive power is unmatched by any other holistic pitching statistic. Jonathan Judge, who built the construct, calls it cFIP. It’s FIP, but with contextual adjustments. Pitchers are evaluated only on the basis of strikeouts, walks, hit batsmen, and homers, but all of those outcomes are weighted according to who was at bat (good hitter? Bad hitter? Same- or opposite-handed hitter?), where the at-bat took place, and who was behind the plate (both of them: umpire and catcher). It is to FIP as DRA is to RA, which is just a way of restating the first two sentences of this paragraph: cFIP is imperfect as a descriptive stat, but it captures present talent and predicts future performance better than any other pitching statistic. Yes, the ‘c’ in cFIP is for ‘context,’ but one could be forgiven for thinking that it stands for ‘Cleveland.’ That’s because, this season, the cFIP leaderboard looks like this: cFIP Leaders, 2015 (min. 40 IP)
(I should mention, just for ease of comprehension, that cFIP is set on a scale whereby 100 is league-average, and lower is better.) If you think that list is radical, you should have seen it at the beginning of this week, when the top five went: cFIP Leaders, Through Games of 5/25 (min. 40 IP)
I wrote an article back in March in which I examined the extremely wide error bars on the Indians’ rotation. Corey Kluber is the reigning Cy Young winner. Carrasco and Salazar have flashed some of the most electric pure stuff one can find anywhere in the majors. It’s not surprising to see them performing exceptionally well. It’s pretty jarring to see them rated more highly than Clayton Kershaw or Max Scherzer, though. A subscriber was perusing these stats a week or so ago, and was so vexed by what he saw that he sent us an email with the subject line: “cFIP stats are very favorable to Cleveland pitchers.” His message, in full: Danny Salazar, Corey Kluber, Carlos Carrasco, and Zach McAllister all have cFIP values of 76 or better. Salazar AND Kluber out rank Dellin Betances! Could there be a typo in the Cleveland contextual factors that is giving their pitchers a boost in value? Are the factors and algorithm published so that others can review the results? Well, in that introductory article to which I linked above, Judge laid out more or less the whole process that goes into constructing cFIP, so the answer to that last question is yes. As to the first question, Rob McQuown did his due diligence, but nothing was amiss. As far as cFIP is concerned, Kluber and Salazar are the two best pitch-for-pitch starters in baseball right now, and Salazar is more or less right where Chris Sale finished 2014. (Before Tuesday night, he was just a tick behind where Kershaw finished.) Kluber’s presence at the top of the pile should surprise no one, but Salazar’s number is pretty wild, and I don’t fault the reader who wanted it double-checked. Since it’s true, though, as far as it goes, let’s move on to a related query: Why? What, exactly, makes Danny Salazar a cFIP darling? And should we care? First of all, we should establish one thing: Danny Salazar is better than you think. He’s probably the best pitcher you’ve almost never heard about, and he’s almost certainly the most underrated young pitcher in baseball. In 38 career starts dating to mid-2013, he has struck out over 28 percent of opposing hitters, walked less than one-fourth as many, and posted a 3.44 (unadjusted) FIP. He struggled to get deep into starts during his first two (partial) seasons, and batters have a .327 career BABIP against him, but these are about the worst things you can say about him. Still, it’s hard to believe that a guy who never appeared on a top-100 prospect list, and who carries a 3.83 career ERA in this pitcher-dominated era, is secretly the best hurler in baseball. Of the 142 starters with at least 200 innings pitched since the start of 2013, Salazar’s FIP ranks 40th. Clearly, cFIP sees something FIP alone does not. Here are two things that might be separating Salazar from the field, in the eyes of cFIP. Adjustment for Quality of Opponents Here are the broad-stroke numbers: Danny Salazar’s Opposing Batters, 2015
This is how you impress a context-corrected framework like cFIP. Salazar has faced guys who rarely strike out, but has fanned them anyway. He’s faced guys accustomed to hitting for very high averages and racking up strong OBPs (the Royals and Tigers, especially), but has held them very much in check. And when he has allowed long balls, it’s been to players who can take advantage of even good pitchers. That’s not cheating the system. That’s a real reflection of dominance on the mound. Adjustment for Handedness Interactions Should it? I argue that it absolutely should. In fact, this is one of my favorite things cFIP does. See, a pitcher who can consistently outclass opposite-handed hitters is platoon-proof. He can more easily remain a starter, because he need not be hidden from anyone. When he does start, managers can’t simply stack the lineup card with platoon bats. Sure, there’s the John Danks/Mike Mussina Gambit, popularized by Joe Maddon during his time with the Rays, wherein a team counters a changeup-reliant pitcher with a bunch of same-handed batters, but that’s a deeply imperfect solution. Just because a pitcher might prefer to face opposite-handed hitters doesn’t mean hitters prefer to face that (same-handed) pitcher. Curtis Granderson, a left-handed hitter with notorious trouble hitting southpaws, never got any relief when facing the lefty-friendly Danks: He hit .080/.115/.080 against him in 27 plate appearances. There’s a practical limit on how hard same-handed batters can hit a decent pitcher. (Certainly, cFIP would also capture the red flag it would be if a guy really couldn’t get outs even with the platoon advantage.) If a hurler can make hay even against opposite-handed hitters, he’s well on his way to succeeding on a large scale. Salazar has cleared that hurdle, and cFIP takes notice. Conclusion With that lens, though, we can see Salazar as a Pedro Martinez prototype: a flame-throwing, undersized Dominican righty with a killer changeup; a true ace. He’s probably one more good pitch from really reaching elite status, but then, as Eno Sarris observed late in April, he seems to be adding one: a curveball with serious potential, particularly as it fits into Salazar’s existing arsenal. It might not ever come together as nicely for Salazar as it did for Pedro. Hell, the chances it would ever come together that nicely for Pedro were pretty thin. The point is, I don’t think cFIP is seeing things that aren’t there. I think, in the case of Salazar, it’s helping us see something we might otherwise have missed: one of the best pitchers in baseball.
Matthew Trueblood is an author of Baseball Prospectus. Follow @MATrueblood
|
Free Agent Watch: Week Nine
Danny Salazar had a bad outing against Texas, but it was in Cleveland.