Matthew Grieco wrote:
I have to go with Wabber on this one. ERA gives you a result -- but stats anterior to it give you a reason for that result and a clue about the future.
Let me go at this one more time, because I believe we agree on most points. I agree, for example, that the stats cited by Wabber can help spot raw talent and predict future performance.
But can we agree that the job of a pitcher is to prevent runs, not to maintain a high G/F ratio? If so, the best measure of a pitcher's current effectiveness would tell us, on average, how many runs he gives up in a game.
ERA purports to tell us that, but Matthew and Wabber say it fails to take into account things beyond a pitcher's control, namely park effects and fielding. (Is there anything else?) I concede the point. Total Baseball's adjusted ERA, however, takes into account park effects, and it makes a difference. Koufax won the ERA title each of his last five seasons, but he topped the league in adjusted ERA in only two of those seasons (1964 and '66).
Adjusted ERA does not take into account fielding, but I question whether teams' differences in fielding would have a large impact on a pitcher's ERA over the course of a season.
Take double plays. In 1992 (last year in my TB), the Giants turned 174 DPs, most in the league. The Expos turned 113, fewest in the league. That's a difference of 61 DPs between the best and the worst. Let's say a pitcher worked 210 innings for the Expos. That's about one-seventh of the total innings played. So we might expect that, over the course of 210 innings, the Expos would fail to convert nine DPs (61/7) that would have been converted by the Giants. How many of those unconverted DPs would lead to earned runs? I'd say between three and four--enough to bump up the pitcher's ERA about 0.13 to 0.17. But that's the worst-case scenario. In most cases, the DP differences among teams might lead to an extra run or two being charged to a pitcher over a season. The effect on ERA would be negligible.
Of course, I can't be sure that fielding does not have a greater impact on ERA than this superficial analysis suggests. I'd be surprised if no one has studied the matter. But if fielding does indeed have a signficant impact, I don't see why it could not be incorporated into adjusted ERA. That would tell us what we want to know most.