December 11, 2017

A Case For Advanced Goalie Stats

I relied heavily on for this article.

I’ve always liked goalies. Growing up in the seventies, I loved looking at those lonesome warriors in their interesting masks. I was an average house league defenceman who got bailed out by his goalie more than once, so perhaps I appreciated what they did on a deeper level than most. Someone once told me that the stitches on Gerry Cheevers’ mask represented shots that had hit him in the head. I assumed that the stitch to shot ratio was one to one, and I wondered how many of those stitches came courtesy of shots from Guy Lafleur.

Like most of the kids in my neighbourhood I collected hockey cards and committed to memory many player names and statistics. I was fascinated by shutouts. The idea that a player could so dominate a game that he could single-handedly determine the outcome intrigued me. I knew that the career record for shutouts was 103 by Terry Sawchuk, and I was assured that it would never be broken. I also learned early on not to mention the S-word if a goalie had one going in the third period, unless it was against the Montreal Canadiens. In that case, I had an obligation to try to jinx it.

Today we have advanced statistics. Admittedly, these numbers hold less fascination for me on a child-like level than traditional stats like the shutout. But as a lover of math, I admire the logic and creativity that they require. Advanced statistics are sophisticated. Rather than stand defiant against the winds of change, I’ve decided to accept that it can’t be the seventies forever. Heck, even Terry Sawchuk’s unbreakable shutout record was eventually broken.

Goals Against Average: Besides shutouts, this is probably the most significant traditional goalie stat. Goals Against Average (GAA) is defined as the number of goals that a goalie has allowed, divided by the number of minutes played, multiplied by 60. After World War II, the career leader in GAA was Dominik Hasek at 2.20. He was the first goalie in nearly 40 years to win the Hart Memorial Trophy as the NHL regular season MVP, and he earned the nickname “The Dominator”.

The most recent goalie to win the Hart Trophy was Montreal’s Carey Price, in 2014-15. Price’s GAA that year was 1.96. He had 44 wins in 66 games, and he recorded nine shutouts. He had a save percentage of .933, meaning he stopped 14 out of every 15 shots he faced. He was the consensus choice for best player in the NHL that year. But, just how good was he?

Goals Saved Above Average: This is a modern stat designed to determine how many more goals would have been scored on an average NHL goaltender than the one being considered. To calculate Goals Saved Above Average (GSAA), we multiply the number of shots faced by a goalie by the average league save percentage for the season. This tells us how many shots an average NHL netminder would have stopped. From there, we simply subtract this number from the total number of shots faced to determine the goals that an average NHL goalie would have let in. Then, we subtract the goals let in by the goalie being considered.

Let’s consider Price’s 2014-15 season. He faced 1953 shots that year. The league-wide average save percentage was .915. Multiplying these two numbers means that an average goalie would have stopped 1787 shots, thus allowing 166 goals. Price allowed 130, meaning he saved 36 more goals than an average goalie would have. Because of some rounding issues in our analysis, Price’s GSAA that year was actually 33.70.

Considering Price played 66 games that year, he saved his team an astonishing half a goal per game. Considering how many NHL games are decided by one goal, one can see that Price contributed many points to his team’s final total of 110. Wouldn’t it be great if there was a stat that could tell us just how many?

Goalie Point Shares: Point shares is a statistic similar to baseball’s Wins Above Replacement (WAR) in that it represents an attempt to determine how much a player contributed to his team’s final record. Traditional statistics are considered in this analysis, but so are concepts like marginal goals and marginal goals per point. For example, the league-wide number of goals scored divided by the number of points accumulated by all teams is about two and a half. Therefore, one could argue that for every two and a half goals above average that a goalie stops, the team earns one extra point.

If we divide two and a half into Carey Price’s GSAA of 33.70, we could say that Price directly contributed about 14 points to his team’s total that year. This is close to his Goalie Point Share for that season, but the computations for Goalie Point Share are more complicated. In reality, Price’s Goalie Point Share that year was 16.2. If we round that to 16 and subtract from Montreal’s point total of 110, we see that Montreal likely would have gotten about 94 points with an average NHL netminder that year. Frankly, that seems about right. Montreal was a better than average team even without Price, but as they showed when Price went down in Game One of the conference finals against the Rangers, they were not elite without their star goalie.

Drawbacks and Advantages: Both of the advanced stats that we’ve looked at here rely heavily on shots against and on save percentage, but neither take into account shot difficulty or time in the game. Ask any Edmonton fan from the 1980s whether Grant Fuhr had the ability to make the big save when it counted and they will answer in the affirmative. Also, these newer advanced stats don’t consider the impact that a goalie has on his team’s psyche. Players that play in front of elite goalies will tell you that they feel more confident when they know that their netminder will probably back them up should they make a mistake. Advanced stats fail to take this into consideration.

In fairness though, traditional stats also don’t consider such concepts either. Traditional stats don’t take shot difficulty or time of the game into account. Also, there is no measurement anywhere for the confidence that an NHL player inspires in his teammates. Some of the critics of advanced statistics in any sport fail to recognize that the traditional stats are not addressing their concerns either.

Conclusion: I think that Goals Saved Above Average and Goalie Point Share are both welcome additions to the analysis of NHL goalies. Granted, the calculations are not something that the average fan could do. But then again, how many of us actually bother to calculate the traditional stats? I love numbers and even I don’t calculate them. I just assume the accuracy of someone else’s calculations. Then, if the numbers support my preconceived notions I use them. What can I say? I’m a fan.

In short, I like being able to say that my team’s goalie likely saved my team 34 more goals and gained my team 16 more points than they would have had with an average goalie. It’s just not as much fun as keeping track of shutouts was in the seventies.

About The Author

Jim Beland is a Math Teacher and sports fantatic from Windsor, ON. A devout HABS and Blue Jays fan, Jim uses sports as a metaphor to teach his students about mathematics.

Leave a Reply