Keith Scherer has an article on Rob Neyer’s website arguing that the new Baseball Prospectus wisdom that a pitcher is as much responsible for his unearned runs as his earned runs (see Michael Wolverton’s article) is a fallacy. I’m generally on Scherer’s side - I think the evidence does show that defense is more responsible for unearned runs than earned runs - but I was intrigued by Wolverton’s secondary argument that ERA as a stat generally favours knuckleballers. Scherer quotes Tangotiger:
“If (Wolverton) had divided runs by unearned runs he would have found that the typical pitcher gives up 10% of his total runs as unearned (from 1950 through 1990, anyway.”
I decided to take a look at the percentages of unearned runs given up by knuckleballers:
18.2% - Hoyt Wilhelm
13.9% - Phil Niekro
13.0% - Tom Candiotti
12.9% - Tim Wakefield
12.5% - Charlie Hough
11.7% - Joe Niekro
It seems from a quick eyeballing of the data that knuckleballers do give up more unearned runs even as a percentage of their earned runs, if we take Scherer’s 8-12% range as the standard range for major league pitchers. (Joe Niekro was only a part-time knuckleballer, which probably explains his lower percentages.) Admittedly, only Wilhelm and Phil Niekro are above the 13% unearned run percentage of such non-knucklers as Tommy John and Bob Friend, but still it generally seems to trend high. This is probably due to the most obvious fact: passed balls - which knucklers give up in greater numbers - count as errors. It would be interesting to see a stat for ERA that doesn’t include PBs - but until then, ERA seems like a solid stat, albeit with a necessary adjustment for knuckleballers.