At least Cole is back

With a nine strikeout performance today, while allowing only a single walk, it's safe to say Cole is back.  Today's game was certainly terrible to watch, but a little distance from it, and I can realize that at least Cole has pitched well.  If you think about his season so far, he's had six starts.  He had a clunker to start the season when he was still getting healthy, but after that, he's been reasonably good to great in all five starts since then.  Against San Diego in his second start, he gave up three homeruns (at least a couple of them were lucky CBP HRs) but struck out four and walked one through six and was doing very well most of the time.  In his third and fourth starts, he was dominating before he had freak injuries.  And in his fifth and sixth starts, he has been very strong.  The ERA will tell you that he's at 5.04, which looks pretty bad, and for the next four days, we'll hear that "only Brett Myers has an ERA under 5" which will technically be true.  The fact is that Cole has 31 strikeouts and only 7 walks in 30.1 innings, and is maintaining a groundball to flyball rate of exactly 1.0.  That's 9.2 K/9 and only 2.1 BB/9.

There are a few more careful ways of evaluating pitchers than just looking at their ERA over 30 innings.  One statistic invented by Tom Tango, called "FIP" or "Fielding Independent Pitching", is based on the theory that pitchers only control HR, BB, and K.  Batting average on balls in play (BABIP) has little to no correlation between seasons, which gives sabermetricians a feeling that pitchers have little to no control over it.  I take a less extreme view of this theory, but the general point is that you can evaluate a pitcher based on their HR, BB, and K rates.  Roughly:

FIP=3.2+(3*BB-2*K+13*HR)/IP

Cole has an FIP of 4.42 this year.  In other words, if you neutralize his luck on balls in play and timing, Cole's ERA should be at about 4.42.

But even that statistic is not quite right.  Cole has surrendered 6 HR in 30 flyballs.  That is bad luck.  Pitchers actually have little to no control over what percent their flyballs go over the wall.  The hitter's power generally controls how far the flyball goes.  It's the pitchers responsibility to keep it from getting up there, but it's the hitter's job to get it out of the park.

So two other statistics are more useful in evaluating what a pitcher's ERA "should be" if they were not the victim of bad luck.

QERA, invented by Nate Silver (of FiveThirtyEight.com and BaseballProspectus.com), uses the three things that pitchers control most to predict what ERA without luck should be-- Groundball/Ball in Play(including HR), Unintentional BB/PA, and K/PA.  The statistic looks like this:

(2.69 - 0.66*GB% + 3.88*UBB% - 3.4*K%)^2

By this measure, Cole's ERA should be 3.42.

More simply, an adjustment to FIP can be made-- just making the HR/FB the normal 11% that it is in an average park on average.  That's called xFIP

xFIP=3.2+(3*BB-2*K+13*(0.11*Flyball))/IP

Cole's xFIP is 3.26.

In other words, if a few of those flyballs landed on the warning track instead of the first row, and if a few of those hits were spread out a little bit, Cole would have an ERA in the low 3's.  For the record, in 2008, his regular season QERA was 3.59, his FIP was 3.69 and his xFIP was 3.67.