Another Way to Look at the New Jersey Devils’ Clearly Awful Goaltending Problem

On Monday, I wrote a post about how often the New Jersey Devils have faced a goaltender who was playing incredibly well, how often they have lit up a goaltender, and more. Since then, the New Jersey Devils were beaten soundly by Vancouver and Calgary in 3-6 losses. Nico Daws started both games, pulled in both games, and Jon Gillies was not much better. The 16 to 18 skaters the Devils had on the ice in each game were largely poor. The two losses were cruel reminders that the 2021-22 New Jersey Devils are a Bad Team and how they lost in each game demonstrated how they are a Bad Team. The most glaring problem in each of these losses, as witnessed by anyone who watched the game live or after-the-fact with a recording, has been the goaltending. Once again, it must be discussed.

Rather than just post everyone’s save percentage and put up some commentary, I am going to show it in a different way. We know that the Devils have been “goalie’d” 15 times so far this season and they have lit up the opposition’s starting goaltenders 15 times - leading to three times being pulled. This lends to an obvious question, one posed in a comment by Tasmar Vanderweghe to Monday’s post, what about the perspective of the Devils goaltenders? Has New Jersey “goalie’d” anyone? How often have they actually been wrecked? Let us look at how the Devils goaltenders have done as starters.

The Approach

For this post, whether someone is a “starter” or a “1A/1B” goalie does not matter much. The idea is to look at how each goaltender has done for the Devils. As with Monday’s post, I only care about whoever starts the game for the Devils. The Devils have pulled their goaltenders quite a bit; twice in the last two games, for recent examples. By my count, head coach Lindy Ruff and the staff have replaced their starters 10 times this season. Nine of them due to performance and one of them due to a potential concussion issue from the November 14, 2021 game against Our Hated Rivals. To that end, I want to know how well or how badly the goaltenders and the team did when a certain goaltender has been in net for New Jersey from the start of the game.

As with Monday’s post, I counted up each game the Devils played in this season and filtered them out by who started in net for the Devils. The Devils have used six different goaltenders this season: Mackenzie Blackwood, Nico Daws, Jon Gillies, Jonathan Bernier, Akira Schmid, and Scott Wedgewood. Each has started at least two games this season. As the questions relate to how well the Devils have or have not done with them, I decided to collect all situations data from Natural Stat Trick from each of the 61 games played so far this season. Again, I want to know how the Devils have fared in terms of what they allowed in total with each goaltender. That includes power plays, shorthanded chances, and overtime situations. Everything but shootouts and empty-net situations by the opposition.

The 2021-22 Devils and What They Allowed by Starting Goaltender

Thanks to Lindy Ruff giving Nico Daws the start in Vancouver and Calgary back-to-back, the 21-year old rookie is now second on the team in games started. He has a chance at catching Mackenzie Blackwood if Blackwood remains out, Blackwood returns to the team late in the season, and/or Daws stays healthy and/or decent-ish. That stated, the last two starts for Daws were rough. As a result, no Devil goaltender currently has an all situations save percentage above the league average of 90.8% from their starts alone this season.

Natural Stat Trick

A quick legend for the stats in the above chart: CA is Corsi Against, or shooting attempts by opponents. SA is shots on net against the Devils; SCA is scoring chances against the Devils; HDCA is high-danger scoring chances against the Devils; xGA is expected goals against the Devils; and GA is how many goals the opposition scored against the Devils scored. QS is a bit different. It stands for Quality Starts, a concept Rob Vollman popularized through his Hockey Abstract books. His 2014 book defined it as a goaltender who had a performance better than the league average save percentage or posted a save percentage better than 88.5% if there were fewer than 20 shots against. This season, the league average save percentage is 90.8% and the Devils have held an opponent to fewer than 20 shots in a full game just once this season: the infamous 1-4 loss to Arizona on January 19, 2022 that ended up being Blackwood’s last game this season (so far).

You can quickly tell from the gross total that, in general, the Devils have not received good goaltending in all situations from whoever is starting for them. Every goaltender, when they started a game for New Jersey, gave up more goals than they were expected to allow. Out of the 61 games played this season, only 21 of them can be claimed as a Quality Start. (By the way, the Devils are 14-5-3 in those games). Daws is the best of the bunch but even he is behind the league average of 90.8%. Even he has as many Quality Starts as Gillies. While Blackwood has frustrated many, he remains well in the lead in terms of both results and Quality Starts. Bernier could have been a fine veteran backup or 1B but his last game was awful enough to drag him down beneath Blackwood. Wedgewood has absolutely recovered in Arizona, enough to make one wonder why the Devils waived him and not someone else, but his spot duty showed no immediate signs of being kept around. Despite how you may feel, it is unquestionable that Akira Schmid has been the worst goaltender on the team based on his starts this season. Schmid is among league leaders in the AHL in save percentage, so that could be more of a case of Schmid not being ready at all for the NHL but being more than ready for the AHL. In other words, it is too soon to write off the 21-year old rookie.

Natural Stat Trick

When you look at the above on-ice stats as a per-game rate, you can see that it is not a case that the Devils skaters did not show up for certain starters. If anything, they have been at their stingiest for the two rookies. Granted, with only so few starts for each, it is possible that the opponent and situation may confound those results. Still, the Devils have certainly tried to make life a little easier for Daws and Schmid. The team’s rate of attempts, shots, chances, and expected goals allowed are lower than the other four. The results, well, they vary.

This does highlight one of the not-so-obvious issues with the Devils. Even for Blackwood, Gillies, Bernier, and two starts of Scott Wedgewood, the Devils have not been giving up a ton of opportunities. They have been good at keeping opponents from firing away a lot and especially from those “high danger” areas. Their rate of expected goals against are all quite low except for two games for Wedgewood - and that was just two games. Yet, the goalies are conceding a lot of goals. Part of it is due to the quality, or lack there of, of the goaltending. Part of it is that what ails the Devils skaters does not really show up in these kinds of stats. There is a world of difference between the Devils giving up a shorthanded breakaway and the Devils in a scrum in front of the net; but both attempts would likely count as HDCAs if taken closely to the net. There is a world of difference between an opponent getting wide open in a faceoff circle because someone miscommunicated and did not stick with the player and an opponent firing from the faceoff circle with traffic in front; but both could be counted as scoring chances depending on where they are in the circle. Shooting attempts and shots on net (or chances for that matter) do not account for whether a Devil is screening their own goaltender. Yet, all of the goals count the same. The not-so-obvious issue is that when the Devils are beaten in defense, they are often beaten badly and sometimes in an avoidable way. The stats do not suggest this; but the video tape does. (This is also why 31 teams are not constantly bothering Tom Fitzgerald for any of their defensemen despite these really favorable numbers.)

The above is not to say that these stats should be ignored. Just that there are legitimate reasons why it is not simply just the goaltending’s fault. If you want a nice mix of goalie errors and skater errors, then just re-watch the last two games and you will get your fill of both.

Starters Getting Wrecked

Before you go and do that, here is some pain. As stated at the beginning of this post, one of the requests was to see how often the Devils’ starting goaltenders get wrecked. I define this as a difference between expected goals allowed and actual goals allowed of more than one. Recall that the Devils wrecked 15 opposition starting goaltenders this season. As you may expect on a team with bad goaltending, it has happened more than 15 times to the Devils. Here is the damage report:

Natural Stat Trick

The count: 26 times. The Devils’ goaltenders have been under expectations 26 times. As you may expect, the Devils won only three of those 26 games. The rest were all regulation losses. They went 3-23-0. Ouch. Similar to Monday’s post, even if the Devils’ goaltenders played closer to NST’s expected goal model and gave up fewer goals, they likely lose a lot of these games anyway. But some of them could have turned into wins or even post-regulation losses. It would make this look a lot less painful. Alas, it was what it was.

This list includes nine out of the ten times the Devils pulled their goalie for the backup. That happened to Daws three times, Blackwood twice (his other pulling was for injury reasons and it is not on this list), Gillies twice, and Schmid and Bernier once each. I do not think anyone can argue about the pulls in any of the nine cases. They were wrecked in their short time in the net. Each starter that was pulled gave up at least three goals and, coincidentally, were expected to give up at least two fewer goals. The two worst performances by starters this season through this method were Bernier on December 3 in Winnipeg (the 4-8 loss) and Gillies against Tampa Bay on February 15. Both gave up six total goals before getting yanked. It was rough to witness as it was happening and it is not much better in retrospect.

Schmid’s total of four starts all went badly and all four are on this list. That stated, Gillies’ game in Chicago on February 25 was the worst compared to expected goals against among all Devils starters who did not get pulled. The model expected him to stop around three fewer goals; he actually gave up six in the 5-8 loss. Even two of Gillies’ three wins as a starter is on this list. The offense really did have to carry him to both of those results: the 7-4 win at St. Louis on February 10 and 7-4 win over Carolina on January 22. Since that loss in Chicago, Gillies has not started a game for the Devils as of this writing. While Daws is not completely clear from this list, Daws’ worst games have not been as bad as Gillies’ worst games. Do not be shocked if Daws gives way back to Gillies if he continues to have nights like he did against Vancouver and Calgary.

One could make a case of Blackwood’s awful night against Arizona, the 1-4 loss on January 19, as being the worst. It was the lowest save percentage posted by a starter that played the whole game at 76.47%. Neither is all that good and the difference between actual goals allowed and expected goals allowed was just a bit higher in two other starts by Blackwood: The January 4 game in Boston the Devils lost 3-5 and the December 31 game against Edmonton that the Devils won in overtime.

As a final commentary on this chart of bad performances by the starting goaltender, here is a count of who was involved:

• Blackwood: 10 GP (~45.5% of games started)
• Daws: 5 GP (~38.5% of games started)
• Gillies: 5 GP (41.6% of games started)
• Schmid: 4 GP (100% of games started)
• Bernier: 2 GP (25% of games started)

Blackwood has the most appearances on this list one should not want to appear on and the fifth worst xGA-GA difference on the list. That is not good. Daws’ last two starts put him a tie with Gillies for appearances. Also not good. All of Schmid’s starts as a Devil are on this list. Bernier did so twice with his last one in Winnipeg and the one night the Devils got shut out in Anaheim, back on November 2.

Devils Goalie-ing Their Opponents

OK. The Devils’ bad goaltenders were wrecked a lot. 26 times. What about the opposite? Did they shut down any opponents? Did the Devils goalie their opponents? Yes and yes. It is a shorter list, but it did happen.

Natural Stat Trick

Specifically, eight times. The last time was just this past Saturday in the 2-1 shootout win over Anaheim. Seeing Daws on this list three times combined with his not-the-worst statline fuels some of the People Who Matter wondering if he is taking a NHL spot for the future. I think it is still too early to determine that for the 21-year old rookie. I think the last two starts have doused some of that fire, too. But it is good to see Daws’ shutting down the opposition beyond expectations. Both Gillies and Bernier appear on this list; Gillies did it in the 7-1 rout of Montreal back on February 8 and Bernier kept the Kraken to a minimum on October 19.

What is also notable is that Blackwood is on this list three times and each are among the best goaltending performances the Devils have received this season. His shutout of the Islanders on November 11 is even more remarkable when you consider that he stopped 42 shots, 39 of those 78 attempts were scoring chances, and the Isles were expected to score almost as many as the Devils actually did. He allowed nothing. Blackwood’s other shutout is on this list, although it is not nearly as impressive as Philly played a rather sad game on December 8. The non-shutout appearance was in the rare 3-1 win over Columbus back on January 6. Columbus attacked quite a bit but Blackwood was the difference maker. It was probably the last time that the People Who Matter felt Blackwood was going to be OK. Since that game, Blackwood and the Devils as a whole have not been OK.

Also: The Devils went 8-0-0 when the goalie the other team. It did require a rare shootout win and an overtime win for that perfect record, but they did do it.

Games Where the Opponents Were Expected to Score At Least Three Goals on the Devils

As a final area of comparison and to complete the request from Tasmar Vanderweghe, here is the list of games where the Devils were expected to allow more than three goals by Natural Stat Trick’s model. While the model is not perfect, it is a good as a rule of thumb to identify the games were opponents brought an offensive storm on the Devils. How well did the Devils do in those games?

Natural Stat Trick

Surprisingly, not that bad. It happened nine times and the Devils are 4-5-0. I expected worse, but only one of these games - the December 2 loss in Minnesota - ended up on the ugly list of 26 games where the Devils’ starter got wrecked. There were more cases of the opponent getting goalie’d by the Devils than that. Those were covered in the preceding section.

That the Devils have had only nine games where the starter had an xGA of three or more goals does speak well to the Devils’ overall defense. They are not getting gashed over and over by opponents. Again, I think this is where the not-so-obvious issues come up again. However, we cannot say that the Devils are bleeding opportunities for opponents to attack if opposing teams are racking up attempts and quality attempts such that their xGA is over three just nine times out of 61 games.

The five games not on either list are worth some quick thoughts. The non-goalie’d win is Blackwood’s start against Florida on November 9 where the score ended up at 7-3. Surprising to me, Blackwood did better than maybe I would have felt at the time. However, the other performances were not as strong. Bernier, Wedgewood, and Gillies may have not given up much more than what the model expected, but they still posted sub-90% save percentages in their games on this list. Gillies’ three goal allowed night against the Isles on January 13 was almost right to expectation and he still posted an 88% save percentage in the process. That was definitely not good for his overall save percentage. Ditto Bernier’s and Wedgewood’s games on November 13 and October 21, respectively. Daws’ start against Our Hated Rivals on March 4 at least cracked the 90% plateau. He was a save or two away from that being a quality start. Alas, it was what it was.

Conclusions & Your Take

The New Jersey Devils need goaltending help. There is a possible future with Daws. The last two starts have brought him down to Earth, to use a phrase, but he has had enough good performances to at least keep starting over Gillies for now. Schmid’s AHL numbers should warrant another look; his four starts this season should not sink him completely. I do not think the Devils can afford to wait to find out if Daws can keep this up or if Schmid improved with more experience.

Given that both Blackwood and Bernier are injured, this is a trickier situation. While they have both turned in some stinkers as starters, they have also provided some legitimately good games too. This is what is lost when looking at a total save percentage over a season. A couple of truly bad games can sink as much as a couple of truly awesome games can elevate it. Blackwood has shown within this season how amazing he can be and how dreadful he can be. Knowing he has not been 100% and he was still getting starts by Ruff really confounds the situation. If he does return, will he actually be fully healthy or healthy-enough in Ruff’s eyes? More importantly, would he turn in better performances than what he has been doing? If the answer to that question is no, then New Jersey has to move on. Ditto for Bernier; although the Devils cannot see him play until the next season. Whether Fitzgerald does acquire a goaltender will be rather telling in terms of how he sees both Blackwood and Bernier. Yet, Fitzgerald certainly cannot stick with the status quo if we agree that the goaltending has been awful - which it has been. I do not envy his position.

This is also why I’m sympathetic to calls for Dave Rogalski to be held accountable. If nothing else, it is a change the Devils should strongly consider making regardless of whether they change the goaltenders or not. That the starter has got wrecked 26 times as opposed to styling in 8 games is reason enough to consider a new voice on that end.

I also think that this also reflects poorly on the skaters as well. It is true that the Devils are not asking their goaltenders to stop a ton of shots or chances or a high number of xGA. And it is notable that they tried to clamp things down for Daws and Schmid compared to the other four, more experienced pro goalies. However, when they falter, they falter badly. Some of those errors are a result of how they play and how well they play. Goalies do not always get pulled because they themselves have been bad. The stats show that it was warranted in nine out of the ten times the goalie was pulled. However, some of those pulls were made to also send a message to the rest of the team to get their act together. Such as when Daws was pulled against Vancouver and Calgary recently. That certainly did not happen as the team went on to lose in those nine cases it happened.

Ultimately, this post shows what you already know: the New Jersey Devils goaltending is a problem. It is a big reason why they are a Bad Team. And to do nothing about it would be foolish. This is just another way to present it.

By the way, the player who has crushed the Devils the most is a tie: Auston Matthews and Oliver Bjorkstrand each has scored four goals against the Devils this season. Yes, the reputation is real; Bjorkstrand is a Devils-killer. Matthews, well, he has been scoring against everyone with 45 goals in 58 games.

What do you make of these findings? Are you surprised that the Devils have goalie’d opponents eight times this season? Or that opposing teams have crushed the Devils’ starting goaltenders 26 times this season? What do you even do about the goaltending this season? Please leave your answers, reactions, and other thoughts about the goaltenders in the comments. Thank you for reading.