Sunday, September 27, 2009

Clutch Hitting Over Time (1952-2008)

The Retrosheet site can tell you the league averages for various situations going back to 1952 for the NL and 1953 for the AL (plus some earlier years). Here I show the league differences in close and late (CL) situations for both batting average and isolated power. Each league is done separately and each data point is a 3 year average (just a simple average, with the exception being that the first AL data point is just 1953-4).

The first graph shows the NL AVG difference. That was calculated by the non-CL situations minus the CL situations. For example, in the NL in 1952, the non-CL AVG was .251 while the CL AVG was .248, for a difference of about .003. The next two years had differences of about 0.007 and 0.008, respectively. So the first NL data point was (.003 + .007 +.008)/3 or .006.

Notice how the difference is growing over time. This might be due to the greater specialization of relief pitching. But that is just a guess (although the correlation between the 3 year AVG difference and the % of games not completed is .74 in the NL-it was also about .74 for ISO-in the AL those correlations were .85 & .93, respectively-the last graph shows NONCL ISO MINUS CL ISO AS A FUNCTION OF GAMES NOT COMPLETED IN THE AL). Now the same graph but for isolated power (or SLG - AVG).

I wish I could explain what was going on in the late 1960s and early 1970s. It does not seem like it was that much harder to hit in CL situations than non-CL situations in the NL. Now the AL graphs.

The graph below shows how closely related the difference between clutch and non-clutch hitting is to the % of games not completed. The fewer games completed (the more relievers are used), the harder it is for hitters to keep up their normal power hitting in close and late games.

No comments: