There has been much discussion of Livan Hernandez's declining stats. (Aaron Gleeman's post was as good as anyone's.) But of course, you can't throw an entire season's worth of pitches in one game, so I thought it might be nice to look at quality starts for Hernandez over the past couple of years.
A quality start is defined, traditionally, as one with 6+ innings pitched and three or fewer earned runs. Bill James, in his creation of "game score" as a statistic, defines it instead as a start with a game score of 50 or over. I've included both below, with QS using the traditional definition and QSgs using game score.
Here's Hernandez's numbers for the past ten years:
Year GS QS QSgs Avggs NM
1998 33 16 16 48 2
1999 30 14 13 48 2
2000 33 21 18 53 3
2001 34 12 13 45 1
2002 33 19 17 49 2
2003 33 22 21 57 1
2004 35 20 22 56 1
2005 35 22 22 51 3
2006 34 19 12 47 7
2007 33 19 14 44 5
"NM" stands for "near-misses"; these are the games that qualified as quality starts by the traditional definition, but had a game score in the 40s. I speculate that the increase in this is due to the decrease in Hernandez's strikeout rate. (In the game score calculation, a strikeout counts as 2 or 2.667 points, depending on the inning; a regular ho-hum out counts as 1 or 1.667.)
By way of comparison, roughly half the starts in the league in any given season will have a game score of 50 or above. Slightly fewer - in 2007, between 47 and 48 percent in both leagues - will fit the traditional definition of a quality start.
Throughout his career, Hernandez's ability to slog through six innings while giving up three or fewer runs has been relatively constant. This is not to say, though, that he has not experienced a decline since 2005; every stat shows that he has. You can see from the Average Game Score column that the average Hernandez start has become less and less dominating.
Do these numbers make Hernandez look better or worse to you? Or is it all just statistical foofaraw?