Earlier today, I published Score-Adjusted Fenwick (SAF) estimates for the last seven regular seasons. In the comments, Darryl Metcalf of Extra Skater noted differences between my numbers and his, and mentioned that his method of calculating SAF used team-specific TOI by score situation, whereas my numbers had used league-average TOIs. I stuck with the league-average numbers to stay faithful to the standard calculation first posted by Eric Tulsky at Broad Street Hockey, and to ensure comparability with other SAF estimates that used Eric’s formula. However, insofar as Eric notes in his post that the use of league-average TOIs isn’t ideal, and comparability with Extra Skater isn’t a bad thing either, I’ve revised my calculations with team-specific TOIs.
Using team-specific TOIs meant that I couldn’t simplify the math by grouping all the constants into a separate term. The equation I used for these calculations is as follows:
So, if you’re interested in SAF numbers that compare well with those you’ve seen published by other analysts, you probably want the ones from my earlier post. If you want numbers comparable to those available on Extra Skater, and truer to the intention of Eric’s method, go with these.