It’s often a question when evaluating goaltenders: How can we filter out the clutch-ness of a goaltender to actually know how good he is? How can we know a goalie’s going to come up big in the most important situations? To do this, we can observe the Save % for every game and average out the SV% of the winning goalie as well as that of the losing goalie. Doing this with the 2012-2013 season, I came up with this:
Winning Goalie Save%: .941 Losing Goalie Save%: .869
From these 2 numbers, I hope to make a linear regression equation that’s decent for predicting what a goalies Point % should be. Here, we assume that a goalie with a .941 save percentage over the course of the season would be good enough to win every game (a Point % of 1.000), since his numbers average out at .941 each game, which is the average SV% of the winning goaltender. Likewise, a goalie with an overall .869 SV% would result in a Clutch Adjusted Point % (ClAP%) of .000, since he’s averaging the SV% of the average losing goaltender. The reason it’s called clutch adjusted is because what we get from it is the predicted Point % of the goalie when their record is not affected by clutch situations. The linear regression equation is y = 13.889x - 12.069.
Using their SV% and P% from the 2012-’13 season, I calculated the difference of these goalies actual P% with their ClAP% (an * denotes goalies who’ve only played 19 games or less).
As you can see, the goalies with positive residuals (Actual P% - ClAP%) have stopped the puck when their team needed it most, but also let the biscuit in at the times it didn’t matter to much with securing the win. These guys can be seen as clutch goalies, where they ultimately do better in the situations they’re needed in, and conserve energy (like they need to) in the situations they’re not. The goalies with negative residuals show one’s who’s records don’t live up to their save percentage. These goalies aren’t there at the times their team needs them most, despite doing well in games that are hopelessly done with. What’s annoying about the residuals is that they tend to make the incredibly good goalies look bad, because it’s nearly impossible for some teams to return as good of a performance as Craig Anderson gives the Senators, so we can just stick with normal ClAP% for the most part.
An important thing to understand is that ClAP% is what a goalies point % would be had they been given the average difficulty of shots every game, along with an average offense and a consistent save %. Clearly these are all factors that never stay consistent, which is why it’s rare that anyone’s ClAP% residual is zero.
The lowest P% of any team to make the Eastern Conference playoffs in ’13 was .573 (New York Islanders), while in the Western Conference it was also .573 (Minnesota Wild). I thought it would be interesting to come up with an imaginary playoffs for that year given that is was goaltending that decided the games, and every team played their top goaltender all the time. So theoretically, the Sabres, Lightning, and Hurricanes should’ve all made the cut, as well as the Stars, Oilers, and Flames. It’s also funny to note that Craig Anderson would’ve lead Ottawa to an undefeated season, and if Scott Clemmensen played the entire season, he would’ve only won 3 games and gotten 1 overtime loss.