June 12, 2008
More on QuikERA
It's time for another look at one of the more useful pitching statistics out there, QuikERA. For those who are just joining us, QuikERA (or QERA), is made up of strikeout, walk, and ground-ball rates. These three components all stabilize relatively quickly, which makes QERA useful for analyzing even small sample sizes. QERA is a better future predictor of ERA than ERA itself, which lends it plenty of utility for the fantasy owner searching for a way to evaluate pitchers. Today, we'll look at a few starters that QERA thinks aren't reaching their true performance level, for better or worse.
Bronson Arroyo has been one of the better starter options around during his two seasons with the Reds thanks to solid contributions in terms of wins, his ERA, and his strikeout totals. This season Arroyo has increased his strikeouts considerably to 8.7, which would be a career high for the 31-year-old starter, and two whole strikeouts above last year's rate. His walk rate has also increased from 2.7 to 3.7 per nine, but with the bump in strikeouts his K/BB ratio has remained almost static. Historically, Arroyo is a fly-ball pitcher, and that trend has continued this season, with just 36 percent of his batted balls going into play as grounders. Despite this, thanks to his excellent strikeout rate, Arroyo's QERA is 4.28, almost one and a half runs lower than his actual ERA.
What keeps Arroyo out of trouble in Cincinnati despite his lofty fly-ball rates is the sheer number of shallow fly balls and infield popups that he induces. For his career, almost 14 percent of his fly balls have been popups, and this year is no different, with a staggering 17.6 percent. The difference between this year and past seasons though is that Arroyo is giving up 14.3 percent of his fly balls as homers, and also allowing 25.3 percent of balls in play as line drives. Those two numbers make it difficult to assume that Arroyo is pitching at the level QERA tells us he is, since an argument could be constructed using those figures to say how hittable he has been.
Arroyo normally handles right-handers well while struggling against lefties, but hitters on both sides of the plate have beat him around this year, to a collective tune of .316/.378/.525; essentially, Arroyo is turning every hitter into the 2008 version of Adrian Gonzalez, which is holding him back despite some solid peripherals. It's difficult to pinpoint exactly what his issue is: he has plenty of movement on his pitches, and his release points are similar enough that he shouldn't be tipping his pitches. His velocity has not diminished at all from 2007 though, and we've seen him increase his strikeouts significantly, so the conclusion that brings us to is that he is in fact struggling to keep hitters surprised and off balance. Until the Reds and Arroyo figure out what's the matter-and things have been worse in June (.405/.457/.833) after a productive May (.271/.340/.372)-I would shy away from using him, but in a year where his strikeouts are way up, it's tough to justify dropping him outright, even taking his recent struggles into consideration.
Andrew Miller's 5.63 ERA is the same as his 2007 mark, but he's made his way to it differently this time around. Though the strikeout rates (7.9 to 7.5, or just three fewer in a third of an inning less) are similar, Miller has actually been a much better starter this year than last thanks to a drop in his walk rate from 5.4 to 4.2, and a dip in home runs allowed from 1.1 to 0.7. That change in homers is the equivalent of going from 25 homers in 200 innings pitched to under 16, a significant improvement for any pitcher, but a necessary one for a starter who walks as many batters as Miller. So if he is pitching better-and QERA says he is as well, with almost a full run difference between his 4.75 QERA and his 5.63 ERA-why have we not seen the improvement in his production?
For one, Miller's BABIP is once again lofty, this year coming in at .369. Though that's higher than his expected BABIP of .333, neither figure is one that a starter should be pleased with. Between allowing opponents to hit over .300 against him and walking 4.2 batters per nine, Miller's WHIP is a dreadful 1.71. It's not hard to understand why he's allowed an above-average 36 percent of his baserunners to score once you take a look at those figures.
There is a glimmer of hope though, and that's in Miller's performance during the past two months. Much like Kevin Kouzmanoff last year (who hit so dreadfully in April that it took almost all season for his line to even out and represent his performance level), Miller's .417/.473/.609 opponent showing in April have handicapped his numbers for the season. Since the calendar turned over to May 1, Miller's improved considerably, as his opponents have hit a much more subdued .221/.308/.307. Assuming his ground-ball rates remain static, his QERA for that period is an impressive 4.29, thanks to twice as many strikeouts as walks. If Miller's still around in your league thanks to the ugly overall numbers, make sure to pick him up or trade for him.
Nate Robertson has never been much more than a mid-rotation guy at his best, but this year we have seen him struggle considerably. His 6.03 ERA is ugly, but his 4.76 QERA gives him the third-largest difference between his actual and hypothetical ERAs among any starter with 50 innings this year. This is mostly because Robertson doesn't walk too many hitters (2.8 per nine), and he punches out batters at an above-average rate (6.1). What Robertson doesn't do well is keep the ball in the park: for his career, he gives up 1.3 homers per nine-that's over 28 for every 200 innings-and he's on that same pace this season.
What is it that causes Robertson to be so hittable, and make him look attractive via QERA? Pitch F/X data helps to detail his problems. First, Robertson is essentially a two-pitch pitcher, in that he only uses his fastball and slider most of the time. He mixes them up well early in the count, and uses his changeup then as well, but he will never throw the changeup when he is in a hitter's count. When the batter only needs to pick between "fastball" and "slider" when neither of them are particularly impressive-Robertson tops out around 88-89 mph, and his slider is of the low-80s variety without much movement-it's easy to see why he has the problems he's had. Josh Kalk has created charts at his blog that detail this very issue. Take a look at how little movement there is in his slider relative to his fastball, when the release points for the two are fairly consistent:
This is the combined movement of his sliders, vertical and horizontal. The batter does not have to adjust very much if he picks up on the pitch being a slider; this has allowed opponents to hit .304/.357/.494 against Robertson despite Detroit converting 70.7 percent of balls in play into outs as a team.
Matt Garza has been knocked around a bit this year, in part thanks to an early injury, but he's settled in with a 4.38 ERA over 61 2/3 innings pitched. QERA is not a fan of his performance, though, as it expects him to be 1.1 runs above that rate at 5.48. Considering Garza doesn't keep the ball in the park all too well-he's on pace for 26 home runs allowed over 200 innings, or 1.2 per nine-we should look into why QERA isn't a fan.
First, Garza is a neutral pitcher, with a G/F ratio of 1.0. His 42 percent ground-ball rate has been a considerable help for his performance despite not being anything special, thanks to a Tampa Bay defense that ranks third in the majors in Defensive Efficiency. Outside of the context of one of the top defenses in the league though, Garza's groundball-inducing ways are nothing to write home about, and QERA recognizes this.
Secondly, Garza's seen a significant drop in his strikeout rate while failing to curb his already lofty walk rate. Whereas he struck out 7.3 hitters per nine for the Twins in 83 innings last year, and 9.3 for their Triple-A squad, he's down to just 5.0 this season. Toss in 25 walks versus his 34 strikeouts, and things start to look pretty iffy. On May 28, Garza struck out ten batters while walking just a pair; for the rest of the season, he has 24 strikeouts against 23 walks, an unacceptable rate to expect a quality performance from in the long run. Garza has potential-he was ranked as the fifth best talent under 25 on the Rays by Kevin Goldstein, and that organization is filled to the brim with players who fit that bill-but if he isn't going to miss bats, his potential is cut back considerably. If you need help at other positions, or already have a plethora of starters, moving Garza before his ERA balloons may be something to look into.
Though QERA is a statistic you can get plenty of mileage on in your league, it's a good idea to delve deeper rather than relying on it by itself, as pitchers like Robinson-who look good when you only talk about the components that make up QERA-are the kind of guys who could ruin your team's stats for weeks at a time, despite the occasional promising performance. This isn't a knock against QERA either; it's an excellent statistic, but one should never settle for using a lone number to guide their decisions. As a guide and complementary piece for evaluating pitcher's performance, though, it can't be beat.