clock menu more-arrow no yes mobile

Filed under:

Staying Competitive: GA vs. xGA

The Devils have been excellent at preventing goals against, but their xGA says it shouldn’t last.

New Jersey Devils v Boston Bruins Photo by Steve Babineau/NHLI via Getty Images

Any fan of the New Jersey Devils this season would know that perhaps the biggest reason the team has seen more success than anticipated has been from excellent goaltending. Mackenzie Blackwood has been saving almost one and a half goals above average per 60 minutes at 5 on 5, tops in the league for any goaltender with more than one start this season. Granted he has only played a few games himself thanks to entering the covid protocol way before the entire team did, but he has still been a big part of that. And Scott Wedgewood and Eric Comrie have played well too in Blackwood’s absence, keeping the Devils in games they might not have been in last season.

Now, goaltending isn’t the entire cause for a team’s GA and xGA stats, but it is a big reason for it, so I decided to start there. However, across the board, the Devils have been playing significantly better than expected in terms of giving up goals against, whether it is from the play of the goaltenders or from better defense from the skaters. To date, the Devils have far and away the best ratio in terms of their actual goals against per 60 minutes versus their expected goals against per 60 at 5 on 5 play. Here are the numbers for the top 10 teams in GA/60 at 5v5, and how those numbers compare to their xGA/60 thanks to Natural Stat Trick: Note: these are from yesterday, before last night’s games. All numbers at 5v5.

As you can see from the differential, none of the other teams are even close. The Devils are expected to give up 2.34 goals per 60 minutes at 5 on 5 action, but through 11 games, have only given up 1.49. That is an exceptional difference, one drastically in their favor at 0.85. Montreal has had some real positive movement as well, but at only 0.24, their differential is nowhere near where NJ’s is. Dallas is crazy, with their xGA and actual GA being identical to the hundredths place, but even crazier is Boston, who ranks 5th with 1.88 goals against per 5v5 60, but they’re actually expected to be slightly better than that. You definitely expect the differential to be positive for the 5th ranked team in GA/60. Vegas and the Isles have also given up slightly more goals against per 60 than what they are expected to, despite still making the top 10. But as you can see, no one is close to coming to NJ with allowing way fewer goals than they are expected. Toronto is at 0.29 with their differential, next best, but that is still 0.56 less.

The question will be, then, can the Devils continue to outshine expectations when it comes to goals against. The other teams in the top 10 all have much smaller differentials between actual GA/60 and xGA/60, meaning that the analytics say they should remain near the top 10. Maybe Toronto and Montreal fall somewhat, but only somewhat. With New Jersey, however, it is expected that eventually they will fall heavily back into the pack. If the Devils had an actual GA/60 that matched their current xGA/60, they would rank 18th in the NHL in GA/60, just ahead of Chicago. Considering they are first right now, that is a monster difference, one that would have them losing way more games than they are winning.

As I mentioned above, the main reason for this extreme positive in differential has been the goaltending. Blackwood has been unconscious so far, posting a 0.972 save percentage at 5v5 through 5 games and perhaps an equally impressive 0.920 save percentage at 5v5 against high danger attempts. The only goalie better at high danger saves with more than one start has been Carey Price, at 0.921, but no one with more than one start has been better than Blackwood at overall 5v5 save percentage, which is so impressive early on. He gives up 0.78 goals per 60 minutes at 5 on 5, and if you want to know why the Devils as a whole only give up 1.49 goals per 60 at 5 on 5, that is the main reason.

It also helps that the backup play has been solid despite expectations, with Scott Wedgewood having a 0.913 save percentage at 5v5 so far, which is very good for a backup. Blackwood is the guy and has been unconscious, but Wedgewood has kept the Devils in games as well, and that has also contributed to the strong GA/60 numbers.

Further, the Devils skaters have also contributed to this as well, as if they were poor at allowing lots of shots against, you would see more goals against over time just by default. New Jersey ranks 12th in Corsi against per 60 at 5v5, allowing 51.75 attempts against per 60. That is not bad, and in the top half of the league, so you have to be happy with that. Furthermore, the team allows under 30 shots per 60 at 5v5 at 29.31. That is only good for 19th in the league, but it could definitely be worse, and it means that Blackwood isn’t being entirely peppered either.

But I think what you see there overall is why the actual GA/60 is so much better than the team’s xGA/60. First, no one can expect Blackwood to maintain a 0.972 save percentage at 5 on 5. That is near impossible to maintain over a full season, and will come back down to earth. He could still win the Vezina and have that drop down to 0.942 easily. That would add more goals against over time. Also, the Devils are still allowing a decent amount of shots against, at least as compared to the league. Ranking 19th in shots against per 60 means that the defense can be leaky at times, and the only reason the GA/60 is so low is thanks to goaltending bailing them out. The odds of that remaining the same over 56 games is low, and so the reason that the expectation is that GA/60 rises.

In the end, I still don’t think this all means that the Devils are a paper tiger that will eventually crumble. Is it possible? Yes, they could definitely fall apart. But the longer this positive streak goes, the more that this becomes what is expected. Yes Blackwood is not ending the season with a 0.972 save percentage at 5 on 5, but the more he continues to impress, the more likely he ends the season as a top 5 or top 10 goaltender, and that would give the Devils a great chance to keep winning games. And it gives the skaters more time to get better at preventing shots against, something that could help to mitigate any loss in save percentage.