Regarding goal prevention in hockey, we can split it into two parts: skater defence and goaltending. In terms of defending, the responsibility of the skaters is to limit scoring chances for the opposition as much as possible, while the job of the goaltenders is to stop the shots that do get through.
However, when a team allows many goals against, there always seems to be endless debate regarding whether the team defence or goaltending is the primary culprit for those results. In my experience, I feel many people struggle to separate the two objectively.
This is what is currently happening with the Oilers.
For yet another season, Edmonton has had an underwhelming start to the year, and once again, a long-standing question is asked: are the Edmonton Oilers good enough defensively?
So far, the Oilers rank 21st in the league in goals allowed per hour, which is unsatisfactory for a Stanley Cup contender. To address this issue, it seems that Edmonton’s management is leaning towards acquiring a defenceman over a goalie at the current moment. Per OilersNow host Bob Stauffer, Edmonton may be pursuing a top-four defenceman who can play the right side, while according to NHL insider Pierre LeBrun, the Oilers are going to “sink or swim” with Stuart Skinner, Edmonton’s starting goaltender.
This certainly does make sense, as Edmonton’s right-defenceman depth is a massive weak point after Evan Bouchard. They have also played the likes of Troy Stecher and Travis Dermott in the lineup consistently, who are more suited to be 7th or 8th defenceman on a contender.
But what if I told you that Edmonton’s overall team defence has been fine, if not very good?
My go-to statistic for evaluating team defence, separate from goaltending, is expected goals against. This metric has a controversial reputation among fans, but it’s not as overly complex of a stat as some assume.
In a way, expected goals can also be called weighted shot attempts. To compute xG, every shot is assigned a specific value based on the quality of the shot and the probability that the shot will result in a goal. For instance, a shot in the inner slot may have an xG value of 0.5, meaning it has a 50% chance of going in, while a weak wrister from the point may have an xG value of 0.01, meaning it has just a 1% chance of going in. The sum of all these values equals the team’s total expected goals. It is not a perfect stat, but I prefer it much more than simple shots for and against, which fail to account for shot quality.
Now, different data sources can have slightly different methods of calculating expected goals, as they have their own models. So, here is how three commonly used models rank Edmonton’s skater defence:
All three models comfortably rank Edmonton as a very solid defensive team.
Additionally, Natural Stat Trick has a scoring chance model that separates shots into three categories: low, mid, and high-danger. Using this model, Edmonton has allowed 8.73 high-danger chances per 60 minutes of play, the lowest in the entire league.
On paper, Edmonton’s defensive core certainly has room to improve, but overall, the Oilers have done a fine job at limiting quality scoring chances against this season. In this aspect, I feel like they are very underrated.
So, despite their strong chance suppression results, why do the Oilers still rank considerably below average in goal suppression?
The simple answer is goaltending.
Now, I typically try to avoid placing excessive blame on the goaltending, as it’s often lazy and superficial analysis to simply scapegoat a goalie for an entire team’s struggles. To clarify, I am not entirely blaming Edmonton’s net-minding for their underwhelming record; the team’s finishing and power-play have also significantly contributed to Edmonton’s disappointing start.
But, specifically regarding goal suppression, the overwhelming body of evidence suggests that it is more of a goaltending issue rather than a defensive one.
To evaluate goaltending, I typically use GSAx – goals saved above expected. I am not a fan of save percentage, as unlike GSAx, raw SV% fails to account for the quality of shots a goaltender faces.
Edmonton’s starting goaltender is Stuart Skinner, who has appeared in 14 games, while their backup is Calvin Pickard, who has appeared in 8. Here is how they rank relative to the league in cumulative GSAx using three commonly used models:
Safe to say, these results are not pretty.
Interestingly, MoneyPuck rates Skinner significantly worse than the other two models and Pickard slightly better, but regardless, all three models combined strongly indicate that Edmonton’s goaltenders have been well below average. As of November 17, TopDownHockey’s model also ranks Skinner dead last in GSAx (-7.7).
What Skinner and Pickard’s GSAx indicates is that with league-average goaltending, the Oilers would roughly allow ~11-13 goals less, which is pretty significant in just 20 games.
To strengthen this analysis, we can also use data from Clear Sight Analytics (CSA) – a proprietary model. This is the type of model that NHL teams will pay thousands of dollars to access.
The biggest flaw with publicly available models, such as the ones mentioned above, is the lack of puck movement data. In simpler terms, public models are primarily reliant on the location of shots to determine quality. Still, the NHL does not release pre-shot movement data, so for example, public models can’t determine if a pass preceded a shot, if it came off the rush, etc. So, certain teams that allow a high amount of rush chances or shots off dangerous passes can be overrated by public data, while the goaltenders on those teams can be underrated.
However, proprietary models like CSA do incorporate a variety of different pre-shot movement variables, such as passing and situational details of shots (you can view the exact list here). Luckily for us, CSA releases a limited amount of data publicly, such as the top-five and bottom-five teams and goaltenders using their model.
Using CSA’s data, Stuart Skinner ranks dead last in the league in goals saved above expected with a -7.7 GSAx. This roughly matches EvolvingHockey’s model, which ranked the Oilers as the fifth-best defensive team in the league.
Therefore, we conclude that Edmonton’s team defence is actually quite solid, and the goaltending has made it look worse than it actually is. Moving forward, Edmonton’s management should strongly consider acquiring a goaltending upgrade at the trade deadline.
