This is not normal: the David Robertson editionBy
Relievers are tricky little buggers. Their appearances are by nature short and frequent, and they accrue statistics in drips and drabs. As a result, even the most overused relievers typically have statistically insignificant samples of data by this point in the season, and within those samples of data sometimes we see a little bit of crazy. That’s what’s going on with David Robertson so far in 2011. Four things in particular stand out:
6.2 BB/9 rate, 87% LOB rate, 0.0 HR/FB%, .368 BABIP
Regression is a word that gets tossed around quite a bit, but it’s reasonable to expect Robertson to experience some regression in each of these areas. Some of this will work in Robertson’s favor, and some will not. Let’s start with the good.
6.2 BB/9 rate. Robertson’s control hasn’t always been the best at the major league level, but in the minors he averaged 3.6 BB/9. This is a good mark, although he was an advanced college arm when he came into the system, so perhaps the most relevant marks are his 4.4 BB/9 and 3.7 BB/9 in his two final years at Scranton. David has struggled with his command in the big leagues, averaging about 5 batters walked per nine innings throughout his career. For most relievers this would be intolerable, but David frequently makes up for it by preventing subsequent batters from putting the ball in play by striking them out. Regardless, Robertson probably isn’t a 6.2 BB/9 guy going forward unless something’s wrong with him. We can probably expect him to cut down on the walks just a bit, which is always a plus.
.368 BABIP. Most people are familiar with BIP theory so we won’t go through the primer. A .368 BABIP is not normal, and there’s no good reason to expect Robertson to sustain a batting average on balls in play this high. The Yankees defense isn’t horrific – it’s rather good in the outfield – and Robertson is clearly a major league pitcher capable of getting guys out. Robertson can’t be a pitcher good enough to sustain a very high strikeout rate, which he clearly is, and simultaneously be so hittable so as to render his BABIP of .368 normal.
Batted ball profile aside (and it checks out just fine), I ran a Play Index query seeking single season totals for pitchers with over 100 IP, a K/9 of over 9 and a BABIP of over .350 from 1919 to 2011 and came up with two pitchers: Darryl Kile in 1996 and Randy Johnson in 2003. If you set the parameters for just relievers and a lower the minimum IP require to 50, you net 50 pitchers with a strikeout rate of over 9 and a BABIP higher than .350. In other words, it’s very rare, the stuff of flukes, and likely to sort itself out over time if given a long enough runway. Robertson has always been a high BABIP guy, but .368 is a touch too high, even for him. Figure that fewer balls in play will be converting to hits, and figure that Robertson will get better results in this regard. Count this one as a plus.
Not enough home runs: 0% HR/FB, 0 HR/9, 0 HRs, however you’d like to put it. Robertson’s isn’t a particularly ground-ball heavy pitcher as it is, which means a fair amount of batted balls are going into the air when he’s pitching. Eventually, these fly balls are going to leave the park. From time to time pitchers have been known to go a long time without yielding home runs. In fact, since 1901 33 pitchers have thrown at least 50 innings in relief without yielding a single home run. This sounds like a decent number until you realize that in that time frame there have been 3,835 pitchers to throw at least 50 innings in relief. Those 33 pitchers are well into the 99th percentile of home run rates, and only 1 of those pitchers ever repeated his feat (Greg Minton, who didn’t allow a single home run between 1979-1981 pitching as multi-inning reliever for the San Francisco Giants).
Robertson has given up roughly 8 home runs for every 100 fly balls throughout his career, which means he should have given up at least one by now (1.44 to be exact). Spitballing it, he’s likely due for 3 or 4 HRs by the time the season concludes if his HR/FB ratio regresses to normal and he continues getting FBs at a 35% rate. It’s not the worst thing in the world, but it likely will happen and it likely will happen with inherited runners on base, given that that’s usually when he’s brought in. Count this one as a minus.
Strand rate: 87.2%. In his career Robertson has been a 77% strand rate pitcher, meaning he’s stranding about 10% more this year than in the past. Relievers don’t normally maintain strand rates north of 85%. It’s very rare. Mariano Rivera comes in around 80%, as do Joe Nathan, Billy Wagner and Francisco Rodriguez. Plenty of pitchers sit lower.
Robertson is a very good pitcher with the ability to get loads of strikeouts, so his ability to maintain a high strand rate is in some sense believable. At the same time, he’s not likely to maintain this high of a rate for the remainder of the season or the entirety of his career. Could he? Sure. Anything could happen. Robertson could also throw another 25 innings with a BB/9 over 6, a K/9 over 14 with a BABIP of .350, even though only one other pitcher in baseball history has managed to pull that off before (Kenley Jensen, this year) and no pitcher in baseball history has ever done it over 50 innings. But in the absence of some intervening explanation as to why we should expect this to happen, I’m far more comfortable going forward with a reasonable expectation of regression based on probability.