Is WSAA Valid?
November 19, 2003
I’ve explored a new concept (at least, new to me) on this site: Win Shares Above Average (or WSAA). The question I haven’t answered directly, yet, is whether this approach passes our validity test. In this article, I suggested that the key criterion for any modification to Win Shares should be its fit with the Marginal Runs concept. So how does WSAA do?
To answer the question, I first computed average Marginal Runs per out made for all batters in the National League (if you’re curious, it was 0.09 marginal runs per out). I then calculated each batter’s marginal runs per out, and compared it to the league average. For WSAA to pass the validity test, the correlation between this figure (Marginal Runs/Out Above Average) and WSAA should be very high.
Here’s the graph of the result:
The correlation coefficient of these two variables is .995, which is about as close as you can get. I don’t mean to brag, but WSAA is a near-perfect expression of the underlying Marginal Runs concept, and a fitting extension of Win Shares.
Footnote: the marginal runs used in this calculation were already modified by my two previous recommended modifications: home park factor and negative marginal runs. By making these modifications, Win Shares is actually more true to its roots.
It’s just about time to move onto pitching and fielding. This is going to be WAY more complicated.
| <<Previous Article: Money for Nothing | Next Article: Money for Nothing, Part Two>> |