Grubauer — some more analytics!

Analytics are fun, so let’s look at some more basic ones as it relates to goal-tending!

I’ll lay out one way to judge a goalies performance, make a case for it briefly and then we’ll compare Gru and Otter in the Kraken vs Stars series.

The metric of choice, in my opinion at least, is Goals Saved Above Expected (GSAx). It’s formed from two other stats, a basic one and an advanced one.

One component is goals against (GA), this is the most basic stat possible, it’s how many goals were scored. In the context of a goalie stat, it’s how many goals were scored while the goalie was in the net.

The second component is expected goals against (xGA), which is an advanced stat which makes a quality adjustment against the actual shots taken. They look at stuff like where the shot started, was there a tip, was this a shot on a rebound, some angles, etc. Alison Lukan did a nice [writeup](https://www.nhl.com/kraken/news/seattle-kraken-analytics-with-alison-expected-goal-models/c-328756620) about this already if you’re interested in details. It basically will tell you how many goals would normally be scored given the quality of the shots involved.

To compute GSAx, we simply subtract GA from xGA. So if the expected goals were 3, and you let in 1, then your GSAx is 2. Positive numbers are good, negative numbers are bad.

It’s common to label the goalies performance based on GSAx. If you have a positive GSAx, we call that a quality start — you did better than expected. If your GSAx is larger than the margin of victory, we call that a steal — if you had performed merely average your team would’ve lost, hence you stole the game. Alison actually publishes a quick analytics summary for the team after every game, usually within about 10 minutes of game end, and this GSAx analysis is on there.

I like GSAx since it at least makes an attempt to isolate the goalie from the defensemen. If you have an amazing defense that is only letting in super easy shots to stop, that gets accounted for. If your defense is absolutely terrible and the goalie has to stand on his head all night, that also gets accounted for. xGA isn’t perfect, so neither is GSAx, but it’s at least quantitative and mostly objective.

So let’s look at the numbers for our round 2 series:

**Gru vs Otter Goalie Stats**

|Game|Gru – xGA|Gru – GA| Gru – GSAx|Otter- xG|Otter – GA|Otter – GSAx|
|:-|:-|:-|:-|:-|:-|:-|
|1|3.26|4|\-.74|2.78|5|\-2.22|
|2|5.3|4|1.3|1.87|2|\-0.13|
|3|1.79|2|\-.21|2.05|5|\-2.95|
|4|3.89|5|\-1.11|2.56|3|\-0.44|
|5|2.46|4|\-1.54|1.20|2|\-0.79|
|Total|16.7|19|\-2.3|10.46|17|\-6.54|

I did not include any stats for the relief goalies, which is why the GA’s in the table don’t necessarily match the game scores. After Gru and Otter were pulled the relief goalies came in, we just ignore everything that happened.

Looking at it through the GSAx lens, this has been a pretty rough series for goalies. there is only one positive GSAx in the table, so one quality start, and it belongs to Gru. Generally speaking though, he’s been doing just about a goal per game worse than he should’ve. Overall he’s outperformed Otter by quite a bit, though you might make a case that the way Otter’s numbers are distributed are a little better. Otter has 2 really bad games, the rest being not too far from 0, where as Gru is usually around -1. For a best of 7 series, you might rather a guy lay 2 stinkers but other wise be basically average, instead of generally be a goal worse than average.

Let’s look at Gru in the regular season. I’ll filter for goalies who played at least 30 games, and we’ll look at Gru against the best and worst in the League, and Martin Jones, because someone will totally ask. This time I’m going to jump to per 60 numbers, so basically a time weighted average. The regular season isn’t as much of a head-to-head affair as a single playoff series, and not all goalies play equal amounts of time. GSAx/60 is basically telling us what their GSAx would be for a normal 60 minute game, assuming they were in the entire time.

​

**Best, worst, Gru, and Jones by GSAx/60 – Regular Season min 30 games played**

​

|Rank|Goalie|GSAx/60|
|:-|:-|:-|
|1|Linus Ullmark|0.88|
|24|Phillip Grubauer (SEA)|\-0.02|
|33|Martin Jones (SEA)|\-0.16|
|42|Elvis Merzlinkins (CBJ)|\-1.0|

​

So in the Regular season, Gru was basically….. average. He’s right in the middle of pack by ranking, and he’s really close to zero GSAx/60, he basically saves exactly what a league average goalie would.

Let’s do the same thing, but for the playoffs. I’ll set the filter to 3 games here.

​

**Best, worst, and Gru by GSAx/60 – Playoffs min 3 games played**

​

|Rank|Goalie|GSAx/60|
|:-|:-|:-|
|1|Igor Shesterkin (NYR)|1.24|
|12|Phillip Grubauer|0.05|
|21|Vitek Vanecek|\-1.68|

​

Our boy is nothing if not average. Once again, middle of the pack with ranking, damn close to zero GSAx/60. Saving basically what an average goaltender would.

For grins, let’s compare the Kraken against the best and worst teams in terms of xGA. This will kind of quantify how well our defense does.

​

**Best, worst, and Kraken by xGA**

​

|Rank|Team|xGA|
|:-|:-|:-|
|1|Caroline Hurricanes|213.37|
|8|Seattle Kraken|242.75|
|32|Anaheim Ducks|320.76|

​

In terms of limiting xGA we’re actually pretty good! That probably suggests that overall we have a bit bigger issue in the net than we do on the blue line.

So where’s that leave us? Well, I’d say it’s pretty difficult to make the case that Gru alone is what’s holding us back. He is performing right at the average after all, so it’s not like the dude is a huge liability. He’s definitely a 1A in the NHL caliber goalie, just not an elite one.

Overall I’d have to give him a resounding, indifferent, meh. I’d love to see an upgrade at the position (sorry Gru fanatics), but doing something extreme to do it will only hurt the team (sorry Gru haters). Gru’s contract is basically immovable, so he ain’t going anywhere, and it’s an expensive proposition to pay for yet another Goalie. I think he is probably a weak link for us, but the cure would be worse than the disease if we try to make a move there.

by MartialSpark