July 8, 2009
Projecting pitcher performance is a difficult venture, but PECOTA gives it a shot every year, and it does a very good job of it. For a variety of reasons, it doesn't hit right on target with every pitcher, but its track record is good, even at midseason. Today let's take a look at some pitchers who PECOTA has pegged correctly halfway through 2009, as well as the ones where it has missed the mark in either direction.
Below we have the 11 pitchers with a minimum 80 innings pitched that are closest to PECOTA's forecasted ERA, out of the 98 total that have reached that innings threshold. While ERA is not the best indicator of pitcher value, it is in rate form, so it can be measured halfway through the season without too much trouble for either of us. You can also explain why a pitcher's ERA looks better or worse than the one forecasted by PECOTA easily, using more reliable information.
Actual PECOTA Name IP K/9 BB/9 ERA ERA Diff. Jon Garland 101.1 3.5 3.1 4.80 4.89 -0.09 Braden Looper 96.0 5.3 2.7 4.78 4.68 +0.10 Armando Galarraga 93.0 6.0 4.0 5.03 5.13 -0.10 Shairon Martis 85.2 3.6 3.8 5.25 5.35 -0.10 Brett Anderson 83.1 6.6 2.5 4.86 4.97 -0.11 Johan Santana 109.1 8.8 2.0 3.29 3.17 +0.12 A.J. Burnett 101.0 8.8 4.4 3.83 3.70 +0.13 Gil Meche 104.1 7.0 4.0 4.14 4.28 -0.14 Jorge de la Rosa 89.1 9.4 4.1 5.14 5.28 -0.14 Ryan Dempster 105.2 7.6 3.5 4.09 4.24 -0.15 Trevor Cahill 93.0 4.4 3.6 4.55 4.70 -0.15
In total, 25 of the 98 pitchers that have reached my 80-inning minimum are within one-third of a run of their projected ERA; the first 11 are here since Cahill and Dempster are tied for tenth with their being -0.15 runs away. There are not many surprises to be found in this group: it was going to be tough for Jon Garland to beat out his forecast given his home-run tendencies and the new park he found himself in. His adjusted numbers match up with his actual ones as well as PECOTA, so there's a good chance he will still be near the top of this list come year's end. Most of the pitchers on this list are low-strikeout guys that get by with good control or command. While sometimes one of these pitchers will outperform expectations (we'll get to specific examples in a bit), it looks like PECOTA has them pinned for the most part. Johan Santana and A.J. Burnett are the only frontline starters on this list; Gil Meche could be considered that, and probably should be outperforming his forecast by a larger margin than he has, but a case of dead arm has hindered that. The system also pegged the A's two rookies hurlers, Brett Anderson and Trevor Cahill.
Next up we have the pitchers who have done their best to put PECOTA in its place by outperforming expectations:
Actual PECOTA Name IP K/9 BB/9 ERA ERA Diff. Brad Bergesen 97.2 4.6 1.8 3.59 6.28 -2.69 Edwin Jackson 114.2 7.3 2.5 2.59 5.16 -2.57 Nick Blackburn 116.1 4.0 2.1 2.94 5.35 -2.41 Rick Porcello 87.0 4.9 3.2 4.14 6.52 -2.38 Jason Marquis 117.1 4.1 2.7 3.61 5.86 -2.25 Zack Greinke 121.1 8.9 1.4 2.00 4.01 -2.01 Scott Feldman 89.2 4.5 3.0 4.52 5.80 -1.89 Kevin Millwood 124.0 5.4 3.1 3.34 5.19 -1.85 Doug Davis 105.2 6.6 4.1 3.15 4.87 -1.72 Jarrod Washburn 105.1 5.9 2.1 3.08 4.71 -1.63
There is a bit more to dissect in this list than the previous one, as we need to identify why these pitchers have accomplished what they have. Brad Bergesen doesn't whiff anyone, but he keeps his walks to a minimum and keeps the ball in the park for the most part. It's tough to think his stellar ERA will survive the season though, as he's kept his BABIP down at .267 in spite of an Orioles defense that's converted just 68.1 percent of balls in play into outs. Nick Blackburn was covered in detail last week; unlike Bergesen, the defense behind him might help him keep up this charade. Porcello should not be as bad as PECOTA expected him to be, but he also should not be as productive as he has been; unless he starts whiffing more hitters or leaving more balls in the yard, you can expect him to drift in that direction by the end of the season.
Jason Marquis has not been as bad as you would think; sure, he's been a bit lucky on his way to an All-Star invitation, but he's kept the walks to a minimum, and the long ball that occasionally plagues him has been nonexistent. I would bet on some regression, but he should be able to beat out that pessimistic forecast due to his strong start. Scott Feldman and Kevin Millwood can both thank the Texas defense-give the unit a trophy for "Most Improved" after moving from dead last to sixth in the majors in Defensive Efficiency. Granted, this gives you reason to be skeptical that they keep it up, with Feldman posting a .237 BABIP and Millwood sitting at a more reasonable but still questionable .268.
Last, we have those pitchers who PECOTA adored, but who have also been unable to make that love anything but unrequited. PECOTA actually did pretty well in this regard, with just 18 pitchers of the 98 missing their forecast by one-third of a run or more:
Actual PECOTA Name IP K/9 BB/9 ERA ERA Diff. Andy Sonnanstine 81.2 5.5 2.3 6.61 4.46 +2.15 Ricky Nolasco 84.2 8.7 2.1 5.42 4.01 +1.41 Dave Bush 81.0 6.6 2.8 5.67 4.31 +1.36 Bronson Arroyo 103.0 4.7 3.4 5.85 4.57 +1.28 Francisco Liriano 96.2 8.2 4.1 5.49 4.38 +1.11 Cole Hamels 92.0 7.7 1.7 4.70 3.65 +1.05 Randy Johnson 91.2 7.9 2.8 4.81 3.78 +1.03 Todd Wellemeyer 96.0 5.7 3.8 5.44 4.52 +0.92 Joba Chamberlain 84.2 7.9 4.3 4.04 3.12 +0.92 Derek Lowe 106.2 4.7 2.7 4.56 3.72 +0.84
There are a few who missed by a lot, but none so much as Andy Sonnanstine. Sonny's control has not been as good as last year, which is a problem given his industry-standard strikeout rates. To make matters worse, he had problems with giving up homers, which is why he's now in the minors and no longer at the back end of a stacked Rays' rotation. He looks a bit more like the Sonnanstine of 2007 than the one who matured as an intelligent pitcher in 2008; it's possible he returns with his head on straight later in the year, but there's nothing he can do about it while he's farmed out. Ricky Nolasco's FIP is 3.38, and he has pitched much better as of late after a short stint in the minors. He'll be much closer to his PECOTA forecast by year's end if he keeps things up, though the Marlins' defense is not going to do him any favors while ranked 20th in Defensive Efficiency.
Randy Johnson's main issue has been his inconsistency, which many have ascribed to injuries. He's now on the disabled list, so if he can come back healthy he may find a way off of this list. Derek Lowe's placement here seemed odd at first, but that's until you take a look at his numbers-his strikeout rates have dipped from those of the past few seasons, when he was very successful out in LA, and he's walking a few more hitters than last year. The good news is that his FIP, despite the below-average punchout rate, is 3.70, meaning Lowe has a chance to match his projection if he can keep it together during the second half. He just needs more cooperation from a Braves defense that sits in the bottom third of the league defensively-that makes the life of a groundballer difficult, especially if he's struggling to whiff hitters.
Besides Sonnanstine, Bronson Arroyo (between a massive dip in strikeouts and a 5.98 FIP), and Dave Bush (2.0 HR/9) there isn't really anyone on this last list that deserves to be on it. Todd Wellemeyer is walking more hitters, sure, but his FIP is right around the same place as it was in 2008, which is right near his PECOTA forecast. Liriano has been a disappointment to those who thought his short stint last year was representative of his post-op ability, but those folks didn't realize how much the lack of command and drop in velocity would ruin him. His FIP is near his PECOTA forecast though, so you would have to think he can rebound somewhat. Cole Hamels' .352 BABIP is to blame for his problems, as he's striking out hitters just like 2008, walking even fewer, and giving up homers at about the same rate. That results in an FIP of 3.64, which is a smidge better than last year.
The true test of PECOTA's ability comes at season's end, but it's good to see that the system has not made many outright mistakes among those pitchers soaking up the most innings for their clubs. By the time October comes around, expect to see the list of those underperforming their weighted mean dwindle, while those over their heads regress towards expectations as well.
A version of this story originally appeared on ESPN Insider .