The Best Five-Year Pitching Performances Since 1920 Based on Fielding Independent ERA

by Cyril Morong on Jul 17, 2006 9:05 AM EDT

That doesn't exactly roll of the tongue. But anyway, this article actually combines ideas from two earlier articles of mine (sees references at the end of this one). In one I found the best five-year pitching stretches based on runs allowed (adjusted for park effects and league averages). But since a pitcher's runs allowed can be affected by the quality of the fielders behind him, just using runs may not identify the best pitchers. There is a stat called "fielding independent ERA" or FIP ERA which calculates ERA only using pitcher controlled stats (HRs, walks, and strikeouts). I use FIP ERA here to find the best 5-year performances.

For the details on how FIP ERA is figured, see the article called "The Best Fielding Independent Pitching Seasons From 1920-2005." Besides using only pitcher controlled stats, all ERAs were normalized to a league with about a 3.70 ERA. Also, HRs allowed are park adjusted. All pitching seasons used had 150 or more IP. So if a pitcher was really great for a five year stretch but he had less than 150 IP in one of those seasons, it did not count.

The table below shows the lowest FIP ERAs over five-year periods (consecutive years).

I only wanted the best 25. But Newhouser had 3 stretches that overlapped with WWII years, when many of the good hitters were in the military. As you can see, some of these five-year periods overlap. The next table removes all overlaps. You might notice that Grove's 1932 stretch (1928-32) overlaps 4 years with the 1931 stretch (1927-31). So the table contains only completely separate and distinct periods.

Having a lower ERA is great. But if you don't pitch very many innings, you don't have much value. So the next table shows how many wins above average a pitcher had. This is based on what Bill James calls the "Pythagorean Formula." To get a winning percentage you square runs scored and divide it by (runs scored squared + runs allowed squared). I assumed that each pitcher was on a team that scored 3.696 runs per game. Their runs allowed per game is, of course, their FIP ERA. Then I divided each pitcher's IP by 9 to get a number of games pitched. That got multilied by the Pythagorean winning percentage to find a pitcher's games won. To see how many games that was above average, I first multiplied the game total (IP/9) times .5. That is how many games the average pitcher would have won. That got subtracted from how many games the Pythagorean formula predicted to get added wins.

The table below shows the leaders.