Recently, I wrote about Matt Vinc and Nick Rose and how their career stat lines are quite similar, other than the Goaltender of the Year awards they’ve won, where Vinc is ahead 8-0. My initial thesis was that their stat lines were very similar but Vinc has had more post-season success, which implies that the teams he’s been on have been better than Rose’s. So can we conclude that Rose, with similar stats on worse teams, is actually the better goalie? The conclusion, based on the GSAA (Goals Saved Above Average) stat, was that no, Vinc has been the better goalie.

But how does the GSAA stat tell us this? Specifically, if Nick Rose’s save percentage is very close to Matt Vinc’s, why is Vinc’s GSAA so much higher than Rose’s? Today, we’re going to see if we can answer that question by looking at this stat, how it’s calculated, and what it really means.

For the examples below, we’re using numbers from the 2023 NLL season. In 2023, Rose had a save percentage of 80.47% and a GSAA of 23.50. Vinc had a save percentage of 80.30% and a GSAA of 25.16. We can see that Rose’s save percentage was actually higher than Vinc’s, but he had a lower GSAA. Let’s find out why.

Calculating GSAA

To calculate the GSAA, we perform the following steps:

• We calculate the average save percentage for the entire league. In 2023, there were 10,651 saves made on 13,729 shots, which comes to a save percentage of 77.58%.
• Next we subtract that number (as a decimal) from 1 to give the percentage of shots that DID score. In this case, we take 1-0.7758 which gives us 0.2242.
• Then we multiply that by the number of shots faced by the goalie in question. This is calculating how many goals an average goalie would have allowed on the number of shots faced by this goalie. This gives us 182.5 for Rose and 207.2 for Vinc.
• Finally we subtract the number of goals that the goalie DID allow during that season to find out how many more or less than average they allowed.

Analysis

In plain terms, here’s what the stat means. In 2023, a league average goalie would had a save percentage of 77.58%. If such a goalie was in net for the Bandits and faced the 924 shots that Matt Vinc did, he would have allowed 207 goals. Vinc allowed just 182, which is better by 25. Similarly, that goalie in net for the Rock and facing the 814 shots that Nick Rose did would have allowed 182 goals. Rose only allowed 159 which is better by 23.

In 2023, Rose had a slightly higher save percentage but a slightly lower GSAA than Vinc. Why? Shots faced. Vinc faced an incredible 110 more shots than Rose did over basically the same amount of time (1,056 minutes for Rose, 1,047 for Vinc). So even though Rose saved a slightly higher percentage of shots, the total number of shots he faced was less than Vinc’s. When we do the math, it turns out that Vinc prevented more goals than Rose did.

The thing to understand here is that this stat is not “if each goalie faced the same number of shots, how many saves would they make (or how many goals would they allow)?”. That’s save percentage; the idea of that stat is to be able to compare goalies regardless of how many shots they faced. But this one is “how many less goals did this goalie allow than if an average goalie faced the same number of shots?”. We’re not factoring out the number of shots here, we’re explicitly using it in the calculation.

One advantage of comparing to the average goalie is that offenses, defenses, and strategies differ over the years. In some years, the average number of goals scored in a game was 27, in other years it was 23. A goalie’s GAA and shot percentage varies accordingly. The GSAA compares against the average goalie in the season in question and thus takes that part out of the equation, so to speak. It allows us to compare goalies from the early 2020’s with goalies from the mid 2000’s (or earlier, if we had the stats available) and have the comparisons actually mean something.

But including the shots here is important. Say you have two goalies with an 85% save percentage. Both of them have obviously played very well. But if one of them has faced 100 shots and the other has kept up that same save percentage over 900 shots, who do you want to have in your net?

One drawback to this stat is that it can’t take the quality of shots into account. If Vinc faced many more shots but all the extra ones were from terrible angles or from the restraining line, then they don’t mean as much as if they were from the edge of the crease. However those bad-angle shots count in the calculation just as much as the close ones. All this means is that if you have a great defense in front of you forcing those bad angle shots, that might inflate your GSAA numbers a bit.

Conclusion

So getting back to the original question: if Rose’s save percentage over the last eleven years is close to Vinc’s, why is the GSAA so different? Again, shots faced. Vinc faced more shots than Rose in nine of the eleven seasons from 2013-2023. For the first 7,397 shots of the season, Rose and Vinc would have prevented about the same number of goals since their save percentages were similar. But Vinc faced an additional 1,440 shots – almost 20% more than Rose. More shots means more goals, but Vinc prevented a higher percentage of those than the average goalie would have. Thus in absolute numbers, Vinc prevented far more goals than Rose did. The save percentage by itself doesn’t give you the whole story.