What to Do About Earned Run Averages, Part 2
What to Do About Earneed Run Averages, Part 2
By Carlos Bauer
Being Stupid for a Long Time
Still Beats Being Stupid For Ever
Finally it occurred to me that, maybe, I didn’t have enough data points with just using the league totals, so I decided to use team totals. What I did was put down every major league team’s fielding average and what percentage of earned runs they gave up, just like I had done with leagues.
I plotted, averaged the same way as before, and it started to show a marked progressively, but it still wasn’t good enough. In other words, it was meaningful— but not useful. There still happened to be what statisticians call “too much noise.”
Then I had an inspiration (or just plain dumb luck). It occurred to me that the difference between .955 and .956 was minute. Why, I asked myself, not average all the earned run percentages in 5 point increments, say between .951 and .955, .956 and .960, etc.
The chart on the following page is the product of that research. With each increment of fielding average, the percentage of earned runs increased progressively. (The only place where there was a problem was between .876 and .800, where the percentage turned out to be out of line at 43%. I plugged in 50% because I only had four data points, and it fit in with the percentages before and after. I felt that if I had had more data points this figure would move up to the 50%-51% level.)
How It Works & Does It Work
To apply the chart on the following page [below in this case], one needs the following information: Runs Allowed by a specific pitcher; Innings Pitched by that pitcher; the Fielding Percentage of the teams the pitcher pitched for.
Next, one goes to the chart, looks up the team’s fielding average and finds the estimated Earned Run Percentage (ER%).
Then one has to multiply the ER% by the Runs Allowed by the pitcher to get an estimated number of Earned Runs Allowed.
Finally, one calculates the pitcher’s ERA in the normal way.
Let’s go through a couple of examples taken form The Minor League Register:
In 1904 Charlie Chech had a fine 27 and 8 record for St. Paul in the American Association. Chech gave up 103 runs in 311 innings; St. Paul fielded at a .954 clip. Looking on the chart we find that .954 FAVG (all team fielding averages are taken from the guides) gives an ER% of 71%. Multiplying .71 times 103 R gives us an estimated 73 earned runs. Calculating the ERA in the normal way we come up with an ERA for Charlie Chech in 1904 of 2.12. It seems like a logical ERA for a pitcher with a 27-8 record in 1904.
Next I calculated Oscar Graham’s ERA for 1903. Graham was 28 and 29 for Oakland in the Coast League in ‘03. His estimated ERA turns out to be a relatively high 3.44, in line with what one would expect from a .500 pitcher in that time frame.
Third, I took Rube Vickers 1911 season. (I would have loved to use his 1906 season, but there were not any fielding records for that year.) Rube pitched for Baltimore in the Eastern League that year, and wound up with a fine 32-14 mark. His estimated ERA comes in at 2.09. That doesn’t seem to be out of line, either.
By Carlos Bauer
Being Stupid for a Long Time
Still Beats Being Stupid For Ever
Finally it occurred to me that, maybe, I didn’t have enough data points with just using the league totals, so I decided to use team totals. What I did was put down every major league team’s fielding average and what percentage of earned runs they gave up, just like I had done with leagues.
I plotted, averaged the same way as before, and it started to show a marked progressively, but it still wasn’t good enough. In other words, it was meaningful— but not useful. There still happened to be what statisticians call “too much noise.”
Then I had an inspiration (or just plain dumb luck). It occurred to me that the difference between .955 and .956 was minute. Why, I asked myself, not average all the earned run percentages in 5 point increments, say between .951 and .955, .956 and .960, etc.
The chart on the following page is the product of that research. With each increment of fielding average, the percentage of earned runs increased progressively. (The only place where there was a problem was between .876 and .800, where the percentage turned out to be out of line at 43%. I plugged in 50% because I only had four data points, and it fit in with the percentages before and after. I felt that if I had had more data points this figure would move up to the 50%-51% level.)
How It Works & Does It Work
To apply the chart on the following page [below in this case], one needs the following information: Runs Allowed by a specific pitcher; Innings Pitched by that pitcher; the Fielding Percentage of the teams the pitcher pitched for.
Next, one goes to the chart, looks up the team’s fielding average and finds the estimated Earned Run Percentage (ER%).
Then one has to multiply the ER% by the Runs Allowed by the pitcher to get an estimated number of Earned Runs Allowed.
Finally, one calculates the pitcher’s ERA in the normal way.
Let’s go through a couple of examples taken form The Minor League Register:
In 1904 Charlie Chech had a fine 27 and 8 record for St. Paul in the American Association. Chech gave up 103 runs in 311 innings; St. Paul fielded at a .954 clip. Looking on the chart we find that .954 FAVG (all team fielding averages are taken from the guides) gives an ER% of 71%. Multiplying .71 times 103 R gives us an estimated 73 earned runs. Calculating the ERA in the normal way we come up with an ERA for Charlie Chech in 1904 of 2.12. It seems like a logical ERA for a pitcher with a 27-8 record in 1904.
Next I calculated Oscar Graham’s ERA for 1903. Graham was 28 and 29 for Oakland in the Coast League in ‘03. His estimated ERA turns out to be a relatively high 3.44, in line with what one would expect from a .500 pitcher in that time frame.
Third, I took Rube Vickers 1911 season. (I would have loved to use his 1906 season, but there were not any fielding records for that year.) Rube pitched for Baltimore in the Eastern League that year, and wound up with a fine 32-14 mark. His estimated ERA comes in at 2.09. That doesn’t seem to be out of line, either.
posted by Carlos Bauer at 8:14 AM
0 Comments:
Post a Comment
<< Home