Thursday, August 13, 2015

Using Leverage Situations To See If Team Clutch Correlates From Year To Year

This is based on a post from last week about the 1969 Mets. They had a positive OPS differential in High Leverage Situations but it was negative in the Medium and Low Cases. The 1969 Mets were one of only 10 teams to have an OPS differential of at least .100 in High Leverage cases while having a negative one in BOTH Medium and Low Leverage situations. Of those 10, only the 1969 Mets made it to the World Series. Also, the 1969 series was the greatest mismatch since 1914. See How Overmatched Were The Mets Against The Orioles In 1969?

What I did here was to find the difference in OPS differentials for each team since 1960 between High and Medium Leverage (HMD)  situations and then between Medium and Low (MLD). Then I added those two numbers. Then I found the correlation of each team's total with and their total the next year (I have Dan Levitt to thank for this suggestion, although, of course any mistake are solely due to me).

The correlation was just .03. So if a team did alot better in High Leverage situations than Medium and alot better in Medium situations than Low situations one year, it did not repeat itself the next year. This can be seen as a measure of clutch performance. Doing better the greater the leverage would be a definition of clutch. But it looks like this does not persist year to year for teams.

I also tried multiplying  HMD by 2 and then adding it to MLD. That way High Leverage cases are weighted more importantly. But the correlation was still low, this time less than .01.

No comments: