Let’s chart out this St. Louis Blues “goalie controversy.”

[Author’s Note: The numbers for this article and the corresponding analysis was done prior to the Arizona game on January 20th.]

As I check my calendar I notice we are about halfway through the month of January which means it must be time to freak out bitch and moan hand wring take a critical look at the St. Louis Blues’ goal tender situation.

This year’s critical look is prompted by the “goalie controversy” in which the “starter” goalie is sitting while the “backup” goalie is getting a majority of starts. This comes after a perceived drop in performance on the part of Jake Allen (aka “the starter”). But is it just perception, or can we find some data to back up this claim?

This season’s trend to date

It looks like Jake’s struggles in net started around Jake’s 11th game when he was in net against the Flyers. That was the game where Jake let in a goal from the point while the Blues’ corsi for% for the game was 60%. The Blues went on to lose that game 2-0.

5v5 data from Natural Stat Trick

Jake’s save % going into that game was 94% while the current league average (5v5 minimum 600 minutes TOI) is at 92%. So Jake was riding an above average save% heading immediately prior to his decline. His current save% is 92% which is right at league average. But he took a hard fall before he was able to climb back up to league average goal tending.

At the same time that Jake started his decline, it appears the Blues also started to give up more shots against. Their shots against per 60 started climbing and was above team average for roughly the same time period that the team’s goals against per 60 started climbing and Jake’s save % starts declining. Jake Allen is an average goaltender and this is exactly what average goaltenders look like when they are coming down off a hot hand. Especially when they play for a team that, typically, has strong defensive discipline. When that defensive discipline slips, it results in more shots on net, and league average goaltenders will be league average goaltenders.

Even with that blip of increase shots against, the Blues still rank as one of the top teams in shots against per 60. Both Allen and Hutton are working behind a defense that allows some of the fewest shots against per 60 in the league (as seen in the chart below). And before you point out that the team’s SA/60 is better when Hutton is in the net, keep in mind that Hutton has not played as many games as Jake has this season. Allen has seen over 1600 minutes of 5v5 ice time while Hutton has only seen 685.

5v5 data from corsica.hockey

Allen’s career to date

While we have established that Allen is a league average goaltender at this point in the season, he is actually playing just below average for his career since 2014-2015. The 2015-2016 season were good seasons for him personally, but this season he is playing under his 4 season average and seeing the highest shots against per 60 out of the last four seasons between him and Hutton combined. Once again, an average goaltender is going to play like an average goaltender. Faced with more shots against, they are going to let in more goals.

5v5 data from Corisa.Hockey

Goals Saved Against Average

Goals saved against average (GSAA) is a simple concept and far more illuminating than goals against average. I am surprised we even still use goals against average knowing all we know about goal tender performance as it relates to a team. That is the beauty of GSAA. It takes the league average save percentage for goaltenders and then applies it to the number of shots a goalie has faced when in net. The result is theoretical number of goals a league average goalie would let in if placed in the same situation as that goalie. You take this number and subtract the number of goals against for the goalie. If the number is positive, then the goalie is performing better than a league average goalie would perform if placed in the same situation. If the number is negative, then the goalie is performing worse than a league average goalie.

Historically, Jake has actually performed slightly worse than a league average goalie by allowing more goals than a league average goalie would. He had a positive season in 2015-2016, but only slightly. And while he is negative for this season, he is performing better so far than he did in the previous season. I shaded the bars in this chart by the GSAA per 30 shots. This is calculated just like any rate stat is except instead of using TOI we use the number of shots. So for every 30 shots Jake faced so far this season, he was letting in .13 goals more than a league average goalie would.

5v5 data from corsica.hockey

Goals Saved Against Average by Danger

However, we all know that not all shots are created equal. A shot from the point that gives the goalie plenty of time to set up and get square to the puck is not the same as a shot from up close at a weird angle. Some shots are more dangerous than others in terms of their likelihood to cross the goal line. So Nick Mercadante took the GSAA concept one step further and applied it across each of the danger level of shot. This gives us a bit more nuance and ability to evaluate the strengths and weaknesses of a goaltender. [Quick note: Every site has calculated danger levels somewhat differently. I am using Corisa.Hockey data and he explains his danger calculations here and includes a map to illustrate the danger zones.]

In the past Jake has had problems with low danger shots while deftly handling high danger shots. In 2014-2015 and against in 2016-2017 Allen’s low danger save percentage was below league average and his low danger GSAA was in the negative. Meanwhile his high danger save percentage has been above average and his high danger GSAA in the positive for the last four seasons, keeping in mind last season was his worst season of the four in terms of GSAA HD. Where Jake consistently struggles, is the medium danger shots. These are the shots that originate from the “home plate” area around the net, but excludes a semi-circle around the front of the crease because those are the high danger shots. Conversely, the medium danger shots is where Carter Hutton has been proficient these past two seasons. Over the past 4 seasons medium danger goals have made up 35% of all goals scored against goalies at 5v5 (minimum 600 minutes TOI) while 34% of all shots were medium danger shots. For Jake Allen, 36% of his shots against have been medium danger, while 44% of his goals against were from the medium danger location.

It is not all on Jake.

This is not all on Jake, however. The Blues team as a whole is towards the top of the league in medium danger shots against, both in numbers, and in the rate of those shots. Jake Allen has also seen a much higher rate of those shots when compared to Carter Hutton. But, as we mention above, Carter has only seen half as much ice time as Jake has. The more ice time he sees, the more medium danger shots he will have to save. And while Carter had an average season last season in terms of medium danger shot saves, his two seasons prior to that were worse than Jake’s numbers for those seasons.

5v5 data from corsica.hockey

If I was David Alexander and Mike Yeo, I would take a good hard look at the shots originating from the medium danger area. How are these shots generated, and what is it about these shots that make it difficult for Jake to save. It is clearly his Achilles’ heel, and if they want Jake to be the number one goaltender, they have to solve this problem from both sides of the equation — skaters and goalie.