Hockey Analytics: Goals Against Average is a meaningless stat

One of the most misused stats in hockey analytics today is goals against average, or GAA for short.

Welcome to the first in what may turn into a series of me obnoxiously screaming about things I don’t like in the hockey world. This series will be called: Hockey Literacy. In it, I will explain why certain stats are completely useless and terrible, and should never be used to as an indicator of talent.

Let’s start off with one of the easier stats to laugh at – Goals against average, or GAA. Now, on its own, GAA is not the worst thing in the world. It provides a very rough and flawed estimate of how good a goalie is. Depending on the team a goalie is on, however, GAA comes fearfully close to the line of uselessness.

More from Puck Prose

The primary reason it is not needed is because of save percentage. Save percentage tells you the same things GAA does and more. In scientific studies, scientists like to control things. Likewise, in hockey, this is the point of stats. They control things and show what they really are.

For example, goals tell us how many pucks when into the net. Thus, goals are a controlled form of shots. This same principle can be applied to save percentage and GAA. Let’s take a look at what GAA does and why it is so flawed.

Problem with GAA

As you probably know, GAA is a very simple stat. Just find the number of goals that a goalie has allowed and divide it by their number of games played. What’s wrong with that? Oh, there are many things wrong with it.

First of all, the number of shots a goalie faces varies, depending on which team you’re on. The team that allowed the least shots in the league this season were the Carolina Hurricanes. They allowed 2367 shots.

The team that allowed the most shots in the league this season were the New York Islanders. They allowed 2918 shots. This is a difference of 551 shots. Teams who allow more shots will tend to allow more goals, regardless of how good their goalie is.

If there was some sort of magical team that allowed 60 shots per game, then it wouldn’t matter how good their goalie is. Unless the poor goaltender has at least a .950 save percentage, they are going to allow more than average. Even a .920 save percentage would mean they’d allow five goals a game.

Obviously, there is no team that allows 60 shots per game. However, if you compare the Hurricanes and the Islanders, you can see some teams allow more shots than others. Because of this, some goalies will allow more goals through no fault of their own.

Now, why is save percentage so much better than GAA? Let’s take a look.

Save percentage and GAA

Save percentage is a bit more complicated to calculate than GAA, but it’s still relatively easy. Find out how many shots a goalie stopped and divide it by how many shots the goalie faced. If you don’t like decimals, you can multiply it by 100. Save percentage treats each shot equally instead of treating each game equally.

GAA and save percentage do some similar things. They both take into account how many goals a goalie allows. So what does save percentage take into account that GAA doesn’t? Shots. Save percentage acknowledges that a goalie who faces more shots will allow more goals, even if they are just as good as, or better than, the average goaltender.

Instead of showing how many goals a goalie tends to allow per game, save percentage shows how likely a goalie is to stop a shot. Save percentage does everything that GAA does, in addition to controlling for the number of shots a goalie faces.

For example, look at Jaroslav Halak’s 3.19 GAA. At first glance, you might get the idea that he is the worst goalie in the league. However, this is not the case because of save percentage. Halak actually stops 90.8 percent of shots he faces. Don’t worry Islanders, you didn’t have the worst goalie in the league. You just had the worst defense in the league.

Now, save percentage is far from perfect. A lot of people who are smarter than me are trying to find the best way to measure how good a goalie really is. However, it is better than GAA in every conceivable way. As a goaltending stat, goals against average is completely useless. It has more value as a team stat than a goalie stat.

Conclusion

Listing GAA beside save percentage is inefficient. It is no easier to understand than save percentage, and broadcasters should stop using it on TV. Using goals against average as a measurement of goalie skill is like showing someone the score of a hockey game and then putting your hand over one of the numbers.

Next: All-Decade 1980's Team

As a team stat, GAA is obviously helpful, but it is pretty useless as a goalie stat. Save percentage may not be the best thing either. But it’s the easiest to understand for new hockey fans, as well as being far superior to GAA.

Goals against average is irrelevant when discussing goaltenders. Everyone should stop using it. The only job it had is being done better by another stat.

Home/Editorials