Yankee fans love to complain, and in 2010 there was no shortage of complaining about one Joba Chamberlain. Fans bemoaned his inconsistency, his bad attitude, his inability to handle pressure, and his overall terribleness. A few of us. though, had faith in Joba, and I’m here to explain why. Please enjoy, in four acts, the tale of why Joba Chamberlain is a good pitcher. Settle in and get comfortable, because it’s a long story:
Wins, Saves and Holds: For old school fans, pitching stats revolve around two key numbers: ERA and Wins/Saves (for starters or relievers, respectively). For relievers, wins are not a great measure of performance, because they can only receive a win if they have the good fortune to be the relief pitcher of record when their team retakes a lead. Even then, a win could be a good thing, because they kept the game close while the team came back to win, but relief wins are also common when a reliever blows the a lead given to him and gets bailed out by his team retaking the lead in the next inning. Saves are similarly misleading, because a pitcher can perform well, but not receive a save for his efforts, since most teams have a set closer who regularly finishes games when there is a save situation. However, relievers can get achieve a hold, if they do their job in a save situation but do not finish the game. This happens often in the later innings of close games: For example, on Oct. 2, 2010, Joba Chamberlain pitched the seventh inning of a game against Boston at Fenway Park. While he did allow one inherited runner to score, when he exited at the end of the inning, the Yankees still held the lead. Had he performed the same way in the final inning of a game, it would have counted as a save, but instead he received a hold. The rest of the game, which went through some twists and turns to say the least, is irrelevant. The fact of the matter is, a pitcher that earns a hold performed the exact same duties as a pitcher who earns a save, except he does it before the final inning of a game. In 2010, Joba earned 25 holds, and 3 saves. On the other hand, he recorded 4 blown saves (which a pitcher can receive regardless of finishing the game) and 4 losses over six games (two games resulted in a loss and a blown save).
This is a simplistic way of looking at things, but if you condense Joba’s holds and saves on one side and his blown saves and losses on the other, the result is a record of 28-6. In other words, in thirty four situations where the Joba entered a close game, he succeeded in 82% of them, and that’s before you count his additional three wins. This ratio is, of course, a very rough estimate of Joba’s effectiveness in 2010, but it does provide some evidence that his performance in close games was at least reasonably good. To put those numbers in perspective, in 2010 closers as a group (remember, closers are regarded as the elite of relief pitchers, and thus make for a good baseline against which to judge a reliever) picked up on average about 28.1 saves, with roughly 4.7 blown saves, only slightly better than Joba’s theoretical 28-6 line.
What that means for Joba is that if you consider his holds to be of similar import as saves to a closer, his 2010 performance actually compares favorably to a group that includes many of the games best relievers. Of course, correlation does not imply causation: That is to say, Joba’s similar numbers may not necessarily be a result of similar performance, but, especially considering other factors, it’s not an unreasonable conclusion to draw from the data. Joba also happens to play on a team with a legendary closer, which means that, at the latest, he will regularly pitch in the eighth inning, if not earlier. If Joba’s 2010 season (which was widely considered a down year) came close to approximating an average closer, it only stands to reason that he could do even better in a good year.
ERA: The other “classic” stat for pitchers is ERA. As you know, ERA measures every run scored by a runner that the pitcher in question put on base. There are several components to the Joba Chamberlain ERA discussion, the most complicated of which deals with that last clause, regarding the assignment of blame when calculating ERA. Let’s start simple though, by comparing Joba’s ERA to other pitchers in the the league.
There is a stat called ERA+ that I have previously mentioned on this blog, and ERA+ is, in effect, a simple way to compare different pitchers’ ERA’s in a given year. What ERA+ does is take a pitcher’s ERA and compare it to the average ERA of the league. The formula used so be extremely simple (League ERA divided by the pitcher’s ERA, times 100), but it has been updated slightly, in order to be more accurate. In either form though, ERA+ rates a perfectly league average pitcher with a score of 100 (for: a 4.50 ERA, in a year where the league average was also 4.50). In fact, with the update to the calculation, the score is linear, meaning that a score of say 110, is 110% better than average. At this point you may be wondering “How did ERA+ rate Joba Chamberlain in 2010?”
98. Barely below the league average. But that’s not all. ERA, as I emphasized before, only measures a pitcher’s performance by accounting for runners that he put on base. It fails to account for him allowing inherited runners to score, as well as for poor performance by another reliever, who allows runners to score who were put on base by the pitcher in question. There are two ways to combat this potential inaccuracy: using FIP, or Fielding Independent Pitching, or by analyzing the inherited runner situations specific to the pitcher.
FIP and BAPIP: Because it’s the simpler discussion, I’ll tackle FIP first. What it is, put simply, is what a pitcher’s ERA would be, if you only considered situations that did not involved fielders: Strikeouts, Base on Balls, Hit by Pitch’s and Home Runs (I won’t use xFIP here: it’s slightly more accurate, but far more complicated). Because it takes most hits out of the equation, one advantage to using FIP is that it eliminates much of the effect of excessively bad (or good) luck for a pitcher. Generally, a pitcher (or team) will have a FIP within half a run or so of their ERA; large discrepancies often indicate that good or bad luck had a big effect on that pitcher or teams performance. BABIP stands for Batting Average on Balls In Play, and it measures is batters average on any ball that they hit into play, which means it ignores walks, strikeouts, etc (A pitcher’s BABIP is calculated from all batters he faced in the season). An average BAPIP will be around .300, and like FIP, an unusually high or low BAPIP can be indicative of bad or good luck, respectively. For example, a high BABIP means that a larger percent of balls hit into play become hits, usually is a result of good luck at the plate: line drives that sink just in front of an outfielder, jam shots that drop in for a single, bad bounces that turn into an infield hit While a large discrepancy in one stat or the other could always be a result of a statistical oddity*, corresponding swings in both BABIP and FIP can help confirm a streak of luck that affected a pitchers ERA. 
; Take the 2010 Yankees pitching staff for example (min 50 IP) for example:
A negative number in the “Difference” column indicates that that pitcher’s FIP was was worse than his ERA, suggesting that he may have benefited from good luck. A positive number is just the opposite: The pitcher’s actual performance may have suffered as a result of bad luck. Remember, an FIP within about a half run of the ERA is generally a result of normal statistical noise, and thus is not significant in looking at that pitcher’s performance. Four of these pitchers, though, have more than a run difference between their FIP and ERA: Mariano Rivera (-1.01), Dustin Moseley (-1.03), Sergio Mitre (-1.36) and Joba Chamberlain (+1.42). The first three of these pitchers have a significant negative change, indicating that they likely benefited from some good luck.** Their BABIP’s offer some confirmation: all below average, Mitre’s and Mo’s strikingly so. Joba, however, has an extreme positive change. Not only is it almost a run more than the next biggest change (AJ Burnett, at +.43), it’s also the largest change in either direction on the team. Such an extreme difference implies that not only did Joba suffer from bad luck, but that he and Sergio Mitre likely have inaccurate ERA’s as a result of luck. Joba’s BABIP also backs up this contention: at .342, only David Robertson’s was higher. This fits with the evidence provided by the data on Mitre: his ERA was much lower than his FIP, consistent with his extremely low BABIP.
*David Robertson is a good example of this: His high BABIP is not supported by a matching change between his FIP and ERA, suggesting that his BABIP may not be as significant as in other cases, such as Mitre or Chamberlain, where multiple stats corroborate each other.
**Obviously, no offense intended to Mariano Rivera – a 1.80 ERA and a 2.81 FIP are both highly impressive.
Bequeathed/Inherited Runners: In addition FIP, which can highlight cases where ERA doesn’t tell the full story, we can also explore situations where a pitcher’s ERA was affected by something out of his control: the performance of a following reliever. In 2010, the average strand rate was 69.2%, meaning in situations where a reliever inherited runners upon entering the game, about 30% of them came around to score. There are two ways that this can affect a relief pitcher: He can suffer (or benefit) from the performance of the pitcher who follows him in the game. In addition, his performance with regards to inherited runners, which is not reflected in his own ERA, should also be taken into account. When it came to inherited runners, Joba stranded 28 out of 37 runners, or 75.7% of them, which indicates that he was at least average, and possibly above average, at stranding runners he inherited when entering. Just like a pitcher who stranded very few runners is not punished in their own ERA, what this says to me is that Joba’s ERA does not reflect his above average performance is such situations. On the flip side, Joba bequeathed 14 runners to pitchers that followed him in the game, and 7, or 50% of them scored, much more than the 30% that would be expected. Compare that to David Robertson: in about ten less innings than Joba, Robertson handed over 31 runners to a pitcher relieving him. In his case though, only 8 runners scored. While Robertson left behind more than twice as many runners as Joba (in less innings, to boot), but the high quality performance by following pitchers helped keep his ERA down. Robertson also inherited 32 runners, while stranding 22 or 68.7% of them. While Joba actually slightly outperformed Robertson when entering with runners on base, due to the differing performance of relievers entering after them, Joba’s ERA suffers and, Robertson’s benefits greatly. However, Joba’s ERA contains no indication of his superior performance when entering with runners on.
Conclusion: So, to recap, I’ve looked at several different ways to judge the performance of relief pitchers, and each test vindicated Joba’s 2010 performance. In terms of game decisions, he racked up 28 holds and saves, with only 6 blown saves or losses, for a completely respectable, if not stellar, record. When looked at through the lens of ERA+, his 2010 ERA rates right around average, but other methods indicate that, if anything, his ERA should have been lower than the 4.40 that went into the record books. His excellent 2.98 FIP was second on the team only to Mariano Rivera, while his above average .342 BABIP corroborates FIP’s evidence that his ERA was, in fact, inflated due to poor luck. In addition, examining Joba’s performance with respect to bequeathed and inherited runners provides even more evidence of the same: Joba’s performance was above average when entering with runners on base, and while he is not rewarded in his ERA for succeeding in these situations, he is punished with a higher ERA due to the poor performance of pitchers entering after him. While many claim that Joba’s 201
0 season was a disaster, or insist that, after watching him, they would not trust him entering a game, the numbers don’t lie. Four separate methods of analysis all vindicate Joba’s performance, showing that he was at worst an average pitcher, and at best something considerably more.
Obviously statistics don’t tell the entire story with every player, and the value of watching a player in action will always be extremely important. In Joba Chamberlain’s case, though, many fans are insisting on remembering a few disastrous moment, while ignoring the more numerous occasions where Joba entered and did his job well, often spectacularly so, and the statistics offer irrefutable proof that those occasions were, in fact, much more numerous. Like I said above, the numbers prove that Joba’s 2010 was, at worst, average, and even if it was truly merely average, Joba Chamberlain deserves far more credit than he has received. Now that the Yankees have signed Rafael Soriano, there will be discussion ,at least among fans, about moving Joba back to the rotation. I won’t address that issue, at least in this post, but, regardless of Joba’s role, I have plenty of faith in his abilities as he and the team move forward.
Logical Fan Blog 