Second Thoughts Game #12: Indians 12, White Sox 6
For the Cleveland Indians and the fandom that follows them, Saturday's 12-6 victory was an oasis in what has been an offensive desert of a start.
The team's offense had long been doing things within its control well, drawing walks at the third-best rate in the league, and it incurred strikeouts at the second-lowest rate in the league, but it had not actually hit the ball particularly well - to continue to torture a metaphor, the offense had diligently scouted the desert's topography for the oasis, but it had not yet encountered it. But after Saturday's game - with its twelve hits and eight walks, the team could figuratively bathe in runs and wash themselves both of the season's hitlessness as well as this horrible metaphor.
However, as the author has previously espoused, if one side of the ball determined the course of the game, one should focus on the other side, in order to counterbalance the prevailing narratives.
Mind the Gap
For those well-oriented in the ways of FIP, feel free to skip the next two paragraphs. However, for those new to the world of improved metrics, welcome! Fielding-Independent Pitching might appear a daunting concept, but the premises are rather simple: run prevention (i.e.: ERA) is entirely faithful, on one hand, in describing how effective a pitching staff has been, but is extremely poor, on the other hand, at predicting future runs allowed by a pitching staff. While this premise is extremely odd at first glance, certain factors that impact run prevention - the rate at which batted balls go for hits, and whether those hits are collected in a single inning or spread over several - oscillate wildly for even the best pitchers and pitching staffs.
It's because of ERA's predictive failures that Fielding-Independent Pitching - FIP - was developed. FIP takes into account only the Three True Outcomes - Strikeouts, Walks, and Home Runs: the only three outcomes that involves only the Pitcher, Catcher, and Batter - so as to isolate the actual skill of the pitcher, rather than extrinsic factors, such as the quality of defenders behind the pitcher. Using a formula that combines these three outcomes (strikeouts are slightly good, walks are bad, and home runs are very bad, is the thought process, mathematically represented by linear weights), one procures FIP.
If you're already oriented with FIP and have skipped the last two paragraphs, welcome back! It's provably false, of course, to say that the Indians' rotation, with their 5.18 ERA through Friday's game, has been effective at run prevention. Less objectively false, but likely just as contentious, is the assertion that the rotation's performance hitherto is predictive of positive results moving forward.
The basis for this assertion, as one might gather from the aforementioned exposition, is the team's Fielding-Independent Pitching - at 3.61, the Indians' rotation FIP is better than league average. In fact, the gap between the Indians' starter FIP and starter ERA is the largest in the majors thus far in 2014.
Still, given the Indians' unflattering company in the top ten, a logical mind might inquire just how predictive FIP is - for a staff, one might ask, could it not be the case that a large gap between FIP and ERA is in fact indicative of a flaw in one's pitching staff?
This concern is reasonable, of course, given how sharply FIP theory diverts from intuitive notions of what it means to succeed as a pitcher. Yet last year's top ten through the end of April is an informative guide.
The average E-F gap, as one might expect, is substantially smaller - because E-F becomes increasingly stable as additional innings are thrown, it makes sense that 2013's through-May-1st E-F leaderboard would have a smaller gap than 2014's through-April-11th leaderboard. It's difficult to call any early-season statistics genuinely predictive, but on the Athletics, Astros, Blue Jays, Padres, Angels, and Mariners (6/10 - as stressed, small sample size), end-of-season ERA was better predicted by one-month FIP than it was by one-month ERA. Not that this is in any way a rigorous proof, merely a possibly useful demonstration that the counter-intuitive FIP is indeed useful as a predictive mechanism.
In the case of the 2014 Indians, the E-F gap has two major causes, one minor, one major. The lesser of the two causes, the starting staff's low strand rate, is harmful, having cost the staff about two runs. The major cause, the high rate of hits on balls in play - BABIP - is substantially more harmful, costing the starting staff, per the aforementioned FanGraphs link, approximately eleven runs. The rotation's BABIP, as a statistic that is - for a pitcher - almost entirely skill-independent, is almost certainly going to normalize from its current absurd .380 before Saturday's game to a number nearer the league average of .294 (for reference, the league's highest rotation BABIP in 2013 was .324). That said, the defense behind Masterson on Saturday did the rotation's BABIP no favors, with Masterson putting up a BABIP of .462 for the game.
FIP instructs us to step away from the ERA-related cliff. The rotation, as reflected through its FIP, is doing those things that lead to effective run prevention. Cause, unsurprisingly, does precede effect, and so long as the Indians starters continue this above-average Fielding-Independent production, above-average run prevention will follow.
Masterson: Ace's High xFIP
While FIP uses Strikeout Rate, Walk Rate, and Home Run rate as its predictive mechanisms, there's a logic (to which the author subscribes) that suggests that, because Home Run-per-Fly Ball ratio is extremely volatile and typically skill-independent, one can better predict future performance by replacing actual home runs surrendered with a 'luck-independent' Home Run Surrendered, multiplying league average HR/FB ratio by total fly balls allowed.
What follows is the leaderboard of all Indians starters by xFIP, including Saturday's game.
In terms xFIP, the most effective predictive mechanism of the three, Masterson is the lowest ranked of the Indians' starting pitchers. This becomes particularly worrisome, given the fact that Masterson is a groundball pitcher, and xFIP, being explicitly reliant on fly ball rates, in turn explicitly favors groundball pitchers. If, despite his batted ball profile aligning with xFIP's ideal batted ball profile, his xFIP is still substantially worse than league average, one can conclude that his strikeout/walk rates have been - on the aggregate - very poor.
Masterson's strikeout rate, at 8.8 K/9, is quite above league average of 8.2 K/9; his BB rate, however, at 5.28 BB/9, is jarringly higher than the league average of 3.23. Justin Masterson's walk rate has never been his genuine calling card - he was never going to be Greg Maddux, nor did he in any way need to be a command pitcher. Masterson's ability to induce swings and misses is good enough to compensate for poor walk rates; Masterson is not, however, so good at inducing whiffs as to make up for a 5.28 BB/9 - a walk rate that places him among the ranks not of poor major-league pitchers, but of prospects with control issues.
Sample Size must always be taken into account, of course: Justin Masterson has had only three starts, and these three starts have totaled only 15.1 Innings; it's extremely likely, therefore, that Masterson regresses to (i.e.: approaches, if not inherently replicates) his career walk rate of 3.57 BB/9; however, if one regresses the walk rate, one must also regresses his presently above-average strikeout rate. And if one regresses everything in equal measure, one has in Justin Masterson a very run-of-the-mill pitcher, rather than the ace of a staff, or even the number two, that his extension suggested.
In short, one can say for much of this staff that regression to the mean is likely to lead to a rather good pitching staff; for Justin Masterson, however, his own cause statistics must be improved before improved result statistics follow; he alone among starters shall be excluded from the aforementioned oasis of positive regression.
John can be reached on Twitter at @JHGrimm. He can also be reached by e-mail at firstname.lastname@example.org.
Two, the Indians defense sucks, which I believe is the bigger factor. Asdrubal has no range whatsover at short. Santana is still learning on the job at 3rd. Raburn and Murphy are plodders in right field.
The BABIP will come down, but I doubt it ever gets below .310 or so as long as AC is playing shortstop.
The biggest problem is that the starters have given up 13 first inning runs in 12 games. They run their pitch count up in the first inning which limits their ability to pitch beyond the 5th even under the best circumstances.
It should be an interesting year for White Sox fans. Their team should score a lot of runs and also give up a lot.
The big question for them is whether their new Cuban cleanup hitter (Abreu) is the next Frank Thomas or just a guy that started hot until the pitchers found the holes in his swing.
During his bad years has given more runs, homers, more fly ball out than ground ball outs.