Posted by Andy on August 16, 2007
So previously we looked at the number of pitchers used in wins vs losses in 2006. Here, we look at something slightly different. This graph shows the average number of pitchers used (in any game, not specifically wins or losses) in 2006, vs several decades earlier. Here, the trends moving away from complete games are alarmingly apparent:
Note that I used curved lines on the graph to make it easier to read, but realize that only the actual data points are meaningful.
You can see that in 2006, almost 4 pitchers are used per team, per game. In 1996, the average was a little over 3. As recently as 1986, it was most common to use 3 pitchers, and the average was even less than 3.
Interestingly, if you compare 1966 vs 1976, you can see that use of 1 or 2 pitchers in a game was actually more common in 1976 than it was in 1966.
This is clearer from the plot below:
Here you can see that 1976 actually saw fewer average pitchers used per game than 1966 or 1986. Since 1986, the number of pitchers has been increasing pretty linearly. Of course, this can't continue. In 1986, most every team carried 9 or 10 pitchers per staff. By 1996, most teams had 11, and these days all teams have 12 or 13. But, it can't go much higher than that, unless teams are willing to give up one of just 3 or 4 bench player positions.
It would be interesting to look at the average number of pitchers used per game normalized to the average number of pitchers per roster. I am guessing that the rate is pretty constant since 1966.