Limitations of GSAx

As with any statistic, there are some things that are shown well and other things that become problematic. I have written about how GSAx appears, on the surface, to be a massive step forward for goalie statistics, and it certainly is that. At the same time, it is not without controversy. 

Some people a lot smarter than me have written about variance in expected goals (xG), a stat which seems to go into calculating GSAx (xGA). This article at medium.com covers it. Dom Luszczyszyn referenced it when trying to figure out Igor Shesterkin's next contract with the Rangers. His article is behind a paywall, but I for one do recommend subscribing to The Athletic for all your daily sports news and analysis. 

Lastly, Hockey-Statistics.com has an excellent breakdown on the rink bias prevalent in goaltending statistics. You will notice Shesterkin's predecessor, Henrik Lundqvist, referenced there prominently. You can also expect to see him featured prominently in my GSAx annual and cumulative regular season and playoff top 10s that I plan to start soon. 

Next I will start to reveal the great performances of the GSAx era, i.e. 2008 to the present.



Comments

Post a Comment

Popular posts from this blog

Introduction

Ever since I can remember I have been passionate about hockey, and in particular goaltending. Growing up my position of choice in our local street hockey games was in front of the net. I'd use some old padding I had found somewhere if we played with hard plastic orange road balls or actually hockey pucks, or go without it when we chose to play with tennis balls. I loved keeping the puck/ball out of the net, and contributing to a win. As I grew older I stopped playing but my love of hockey continued, and I combined it with my interest in analyzing statistics. I'd keep large spreadsheets for season and playoff performances and update them after each season had finished. I still do that now, and I do it for hockey, as well as my other passion, baseball.  Baseball is much more of an individual sport than hockey, and as such it is a lot easier to separate out individual from team performances. So my files on baseball stats, for both pitchers and hitters, have grown a lot more comple
Read more

GSAx - some background information

  I do not claim to be an expert on this stat, just intrigued by it. Traditional measures of evaluating goalies have been proven to be ineffective. Obviously the W-L record is the best example of this for reasons we do not need to get into. Save percentage seemed better, and of course it is, but it also does not include within it so much of what goalies actually face, as all shots against are considered equal, no matter the situation or where they come from on the ice. Other stats weight the save pct of a goalie against the league average. This may account for trends in goalscoring across seasons and decades, but does not account for the performance of a goaltender against every shot in a game.  GSAx measures the potential scoring outcome of each shot given several factors, arriving at a total number of expected goals the goalie might allow, and then compares this to the number of actual goals allowed.  Read more here -  Adjusting how we Evaluate and Analyze Goaltenders This article by
Read more