Login

Duff101 · 10-13-2020, 03:57 PM

NOTE: I am by no means a stat guy and this is by no means a perfect evaluation of goalie performance and skill. I put this together over a few days. Shoutout to the Atlanta LR for helping me out and checking my data. This is still very much a work in progress.

INTRO: GAA IS A BAD STAT

GAA is a terrible stat to measure individual goalie performance for a multitude of reasons. The main issue with it is unlike save percentage (which is imperfect by itself) is that it doesn't take into account the volume of shots a goalie faces. A goalie that faces 40-50 shots a game will naturally have a higher GAA than one who faces 20-30. As such if there's one thing I want any potential SHL or SMJHL award voters to take away from this article it's this:

Stop using GAA as a metric of individual goalie performance. It is a team stat.

A good goalie on a bad team will have a high GAA. In order to counter this, I wanted to focus on a stat that focused mostly (you can't entirely eliminate goals against because it's kinda the only way to measure saves) on shots stopped rather than shots let in.

DEFINITIONS

The first important definition to understand and the stat that forms the bedrock of this entire article is Shots Against Average (SHAA). This stat represents the amount of shots a goalie will face in the average 60 minute game. It's calculated in the same way as GAA with shots replacing goals.

[Image: download_2.png]

What this gives us is how shelled each goalie gets. Generally, goalies on worse performing teams will have higher SHAA as their team allows more shots. (For example the highest and second highest SHAA so far this year belong to the SFP goalies, and in S54-55 the title was held by a TBB goalie. All of these teams were or currently have the worst record in the league.) So essentially one can look at SHAA as the quality of defense around a team.

The second definition is expected saves. Expected saves is a simple metric that takes the SHAA and multiplies it by the goalie's save percentage, giving the amount of saves the goalie makes on an average gameday.

[Image: download_3.png]

A goalie with a higher SHAA will stop more shots than one with a lower SHAA even if his save percentage is significantly worse so this stat on its own means fairly little and is mostly used as a building block for the final equation.

THE EQUATION

[Image: unknown.png]

GoalieScore, the worst named stat in history, weights a goalie's performance based on where they rank among starters (>1000 MP). The "score" part of GoalieScore is calculated by taking their save percentage rating and tripling it (because save percentage is more important than shots against average, it should be weighted higher so goalies on bad teams that get shelled don't get an undeserved bump). 10 was chosen as the number to divide it by mostly because it just makes the numbers have a more significant gap. The weighting has a twofold effect. Firstly, it cuts down those who face a high shot volume and therefore have a high amount of expected saves, and raises those who face less shots but save more of them.

THE RESULTS

I did this for every completed SHL season in FHM (everything in the index except S56). I also color coded McBride winners in red and nominees in yellow. The results are all in the spreadsheet below.

https://docs.google.com/spreadsheets/d/1...sp=sharing

S53:
[Image: unknown.png?width=1025&height=311]

[Image: unknown.png?width=1025&height=311]

S54:

[Image: unknown.png?width=1025&height=334]

S55:

[Image: unknown.png?width=1025&height=295]

S53's best GoalieScore is Luca Del Vecchio, S54's is Harry Carpet and S55's is Markus Tegernako. Out of this group, only one of these goalies received a McBride nomination and none of them won.

MY OPINION AND THE FUTURE

While this metric is still very much a WIP I felt I had enough done to share with everyone (also I need money). I've done some preliminary looks into what effect shots against has on a goalie's save percentage but I don't have enough data right now to really come to a conclusion as to whether there's a direct correlation, and there is a significant amount of standard deviation. As of right now the trendline is negative. The other thing is changing the weighting of save percentage. While right now I'm fairly comfortable with the results, I do think that goalies with a high sv% and low SHAA are being punished a little bit TOO much but I don't want this metric to just turn into "the best sv%." I'll give an update at the end of this season, and give the season results as well as who i think should win the McBride.

Credit to Geck, Ragnar and Juni for sigs

ml002 · 10-13-2020, 04:35 PM

Carpet the goat tbh

But more seriously it's nice to see the goalies that aren't always getting Ws getting deserved recognition

Credit to Copenhagen, Wasty, FlappyGiraffe, InciteHysteria, and caltroit_red_flames
[Image: MM4nqx6.png]

***Chevy*** · 10-13-2020, 04:49 PM

Solid work Samat.

Duff101 · 10-13-2020, 05:05 PM

10-13-2020, 04:35 PMml002 Wrote: Carpet the goat tbh

But more seriously it's nice to see the goalies that aren't always getting Ws getting deserved recognition

Carpet posted the best overall score but bc its relative to the season I would have to do a score adjustment to determine the actual correct FHM-era best goalie. I would hazard a guess to say it's probably S55 Tegernako

EDIT: After adjusting the all time rankings it is indeed S55 Tegernako S54 Carpet is second place.

Credit to Geck, Ragnar and Juni for sigs