May 7, 2008
Last week, I covered four starters that QuikERA said were as good as or better than their April statistics. Based on the e-mail response, this week's article will take a look at the other end of the spectrum: the pitchers QERA expects to see drops in production from. Before I get into that though, I wanted to cover one of the e-mails I received, because it offers a good lesson we should all learn and apply to when we make roster decisions, and it even pits two of my favorite statistics against each other. S.B. wrote in to say:
…I calculated the correlation coefficient between the pitchers BABIPs and (ERA-QuikERA) and found it was .61, so there's a fairly strong relationship there. There were some exceptions though (of course). Take Tim Lincecum-his BABIP of .374 suggests his ERA of 1.73 could be even lower, but his QuikERA is 3.39, significantly higher. Which is right? Over just five starts weird things can happen to skew numbers, but do you have a feeling which of these methods is less easily skewed?
The components that make up QERA-namely, K%, BB%, and GB%-all stabilize comparatively quickly in-season compared to BABIP, which can sometimes go an entire season without regressing to the mean. That's usually how you get your "fluke" seasons, where you will see the player return to their normal production level the next year. I prefer QERA for smaller samples, though BABIP has its uses in regards to valuing pitchers as well. In Lincecum's case, his performance has not been negatively affected by his lofty BABIP-he's stranded over 83 percent of all baserunners, which is above the league average-so I wouldn't think a drop in BABIP would mean his ERA is going any lower. He's also given up four unearned runs, which has helped keep his ERA down. His home park and low homer rate mean that you can slice a bit off of his QERA, but expecting him to remain under two runs per nine is asking a bit much.
Now, on to the guys to cover this week. First, let's look at Fausto Carmona, who has an ERA of 2.95 over his first seven starts of the year, but his QuikERA is 6.88; the 3.93 difference (QERA minus ERA) is the largest of anyone with 20 innings and a start this year. Carmona is striking out just 3.4 batters per nine innings and handing out seven free passes per nine, which is not a long-term formula for success. Carmona has started to put the ball on the ground more often though, as his G/F ratio is up to 3.8 from 3.0 in 2007; this increase in grounders has contributed to his lowering his homer rate down to 0.2 per nine, and his HR/FB from 11.2 percent for his career to just 4.3 percent.
With ERA on one extreme and QERA on the other, figuring out where Carmona's true value lies is going to require a deeper look at the numbers. Carmona has increased the rate at which groundballs are induced by upping the number of sinkers he throws at the expense of his other pitches. The problem is that he's thrown far more sinkers well out of the strike zone than he has inside it, jumping his walks up considerably. Though his sinker is filthy enough-it's breaking over eight inches horizontally and six inches vertically-to reduce the number of well-struck balls in play, the walks are going to catch up to him. Carmona has not had a start where his sinker wasn't on besides his April 12 outing against Oakland, so his numbers haven't suffered too much. If he gets in a funk though, we'll see his QERA move closer to being a reality, at least on a start-by-start basis. I wouldn't give up on him entirely, since his walk rates can't stay this high all season, and he's succeeding despite them in games where his sinker does its job.
Gavin Floyd almost threw a no-hitter last night, so you might think the timing is off for me to tell you that you can expect a decline in his performance. That's the case, though, as Floyd's ERA of 2.50 does not sync up with his 6.22 QERA. Floyd is walking almost as many hitters as he is striking out (4.1 BB/9 versus 4.3 K/9), which is two problems at once. That's far too many walks, and nowhere near enough punchouts. Combine that with only 35 percent of all batted balls as grounders-in a park designed to give power hitters a boost-and you can see why QERA is pessimistic about Floyd's future, even though it isn't park-adjusted.
Floyd has succeeded so far thanks to only allowing 4.5 hits per nine innings, which translates into an opponents line of .149/.250/.291. The White Sox have played some excellent defense behind him, better defense than their rank of 13th in Defensive Efficiency lets on. Take a look at this chart from First Inning to see how low the averages are around the diamond:
He's had problems with balls to right, but every other spot on the diamond is turning into an out. Even if his outfield numbers last, it's tough to believe that the opposition is going to hit under .100 on groundballs for the season. So, you can expect Floyd's H/9 to increase, bringing his ERA up with it. Considering last night's performance, it's the perfect time to sell to an optimist.
Daisuke Matsuzaka has earned praise for adjusting to hitters in April and posting an improved ERA, but has his performance been a real step forward? Looking at his 5.85 QERA tells a different story than his 2.43 ERA. At 7.3 K/9, his strikeout rate is solid, though not as high as last year's totals. The problem is in the walks, which Dice-K has been allowing at a ridiculous pace, averaging 6.0 walks per nine innings pitched. Like Floyd, his low hit rate has staved off poor numbers for now, but no matter how filthy his pitches may be, Matsuzaka will not sustain a rate of 4.9 hits allowed per nine innings, and we will see a change in his production.
Whereas Matsuzaka threw a multitude of pitches last year, nibbling away at hitters when he could have put them away with his better offerings, this year he has cut down considerably and dealt primarily with a fastball, a slider, and his changeup. The nibbling has not stopped though, as Matsuzaka has thrown his share of balls this year, as indicated by the lofty walk rates. The one positive to note from the drop in number of pitches used is that Matsuzaka has reigned in his home run rate, dropping it from 1.1 to 0.7 per nine with just 5.8 percent of fly balls going for homers.
Even with the drop in homers, Matsuzaka is in line for some poor production. He's managed to strand 85.2 percent of baserunners so far, but that will be harder to do as his 1.20 WHIP climbs when the hits start dropping in. Matsuzaka struggles with his control more pitching from the stretch than he does from the windup, as his .163/.338/.265 line with runners on versus .156/.269/.289 line without runners shows; this small sample is a larger skewing than last year's Isolated Patience lines of .084 with runners on and .075 without runners, but it's something to pay attention to going forward. The good news is that this is most likely something mechanical that the Red Sox can fix, meaning you will want to hold onto Dice-K, though watching his progress on how many walks he allows is going to be key for your own success.
Another Sox pitcher has outperformed his QERA considerably during the first month plus of the season: Jon Lester. Lester's 5.79 QERA is a far cry from his 3.94 ERA, and is the product of a 5.1 BB/9. Though his fastball has hit 92 mph consistently, a return of some of the velocity he was missing last year (his first returning from cancer treatment), Lester has had trouble finding the strike zone. Not only has he walked more than his share of hitters, he's only struck out 5.3 per nine, putting his BB/K ratio too close to an even one for comfort.
Lester has succeeded so far thanks to a high strand rate, starting with almost 81 percent of his baserunners not crossing the plate. The rate at which homers occur has dropped as well, even if his HR/FB is a percentage point higher than it was last year. He's increased his ground-ball rate and jumped his G/F to 1.3 from 0.8, which has also helped his cause. Even with the increase in grounders and the Sox's third-best Defensive Efficiency, Lester will need to drop the walk rate, as a 1.49 WHIP does not bode well for his future ERA given his other peripherals.
He's shown he can strike major league hitters out with his 7.1 K/9 last year, and his velocity has come back, meaning he should be striking out at least the same number of batters, if not more. A reliance on his fastball over his secondary pitches has messed with his numbers for now, though with only a handful of starts to work with it's difficult to tell if this is just due to a small sample or a new approach to hitters. If Lester dropped his walk rate considerably, he would still be a league-average pitcher. He's going to need to bump up his strikeouts and drop his walks at the same time in order to make this work. If you lack the patience for that, sell high, but hold onto him if you need the pitching help.