One of the most common complaints from the People Who Matter regarding the 2021-22 New Jersey Devils is goaltending. It has been terrible. It has cost the team loads of games and points in the standings. This is not opinion. The Devils objectively have a one of the worst performing team save percentages in the NHL. Prior to the games on April 15, the Devils have the worst 5-on-5 save percentage in the league at 89.57% and the 31st team save percentage out of 32 teams in all situations (empty nets included) at 88.18%. It may be frustrating. It may be enraging. It may be depressing. It is a fact that the Devils goaltending has kneecapped the 2021-22 campaign.
This has led several of the People Who Matter to argue or wish or hope or cope with the the following thought: If the Devils had league average goaltending, then they would be fine. If the Devils can figure out their goaltending position and/or get a league average goaltender for next season, they can be a playoff team. There are variations to this, but this is the central thesis. The Devils need to improve is simple: just a goalie. Or goalies, depending on who you ask.
As a hockey blogger for many, many years, I feel compelled to address this point. I get it. I hear you. I sympathize with it. But I do not agree with that. League average goaltending does and would make a difference - just not enough to turn the 28th placed Devils into a team in a playoff spot in 2021-22.
I can even prove why - at least in theory. I know I have done this earlier in the season. Since the thought persists, often when a Devils goalie gets lit up (e.g. Andrew Hammond getting wrecked by Montreal on April 7) or a Devils goalie actually has a very good performance (e.g. Nico Daws in Dallas and Arizona last week), I feel like I should do it again. So let’s do it.
Proving Eric Tulsky Correct
There has been a long-standing rule of thumb in the stats community that a goal differential improvement of 6 goals - whether it is six more goals scored, six fewer goals allowed, a combination that would add up to six - would be worth a win in the standings. Eric Tulsky, former Broad Street Hockey blogger and now current assistant general manager of the Carolina Hurricanes, explained how that came about in this post back in March 2011. (Aside: He was a finalist for the Chicago GM job a little while ago. Do not be shocked if Tulsky gets a full GM job somewhere soon.) He took the league standings, plotted each team’s points from the standings against their goal differential, and fitted a trend line on it to show the math behind a change in goal differential by 6 “guaranteeing” a win. He used “a few years’ worth” of standings and a spreadsheet program to do that. Rather than just taking it at face value, I want to see if that is still true eleven years later.
I followed the same approach. I went to NHL.com, copied and pasted the league standings from every season since 2005-06, and plotted the results. I decided to go all the way to the beginning of the cap era as that, more than anything else, has driven how teams approach building and icing their teams. I did this on Friday afternoon prior to the games on Friday, Saturday, and Sunday. As you will see, it will not matter much if I did include them.
I immediately found out that including the 56-game pandemic-shortened 2021 season and the 48-game lockout-shortened 2012 season really skewed the results. Enough to get a linear fit with a formula similar to Tulsky’s but with a much lower regression value. This makes sense as the 48-game and 56-game seasons allow for more variation with fewer games played than a normal season. When removing the two shortened seasons, my R-squared value (the correlation value, or how best the formula fits the data) jumped up to a stronger 0.847 with a formula of 0.345x + 89.87. If I remove this season, which is not yet over, and the 2019-20 season that was forced to end early, then I get a formula of 0.348x + 91.56 with a R-squared value of 0.915. In general, an R-squared value above 0.75 has a good correlation and anything near or above 0.9 is a really strong one. There is always going to be variation; a team can over or under perform by a few points just on a few breaks alone. But it shows that there is a relationship between goal differential and points in the standings and it is a close relationship. And it is looking good for Tulsky’s 2011 post and those before him who spoke of this concept.
To explain the formulas, the coefficient - the number next to the ‘x’ - is what each goal differential is worth. If a team had a goal differential of just +1, then that gets multiplied by the co-efficient and that would be worth 0.345 or 0.348. The intercept, which is the number being added to after the multiplication, is a baseline value. A team with a zero goal differential would be worth that many points in the standings. If you include the last two seasons, that’s 89-90 points. If you do not, then it is 91-92 points. In other words, it shows that a team with an even goal differential will at least be decent. Perhaps not enough to be a playoff team in some seasons - like this one - but not bad.
Now that the formula for the linear fit is explained, we can get into where the magical number of six comes from. A win is worth two points in the standings. It does not matter if it is in regulation, in overtime, or in a shootout. One win equals two points. We know that goal differential has a strong relationship with points in the standings. We know a co-efficient to multiply goal differential to add or remove points from the baseline. Therefore, we can divide the number of wins in the standings by the coefficient to get a goal differential amount that would be worth a win. Since I have two formulas, I’ll do it twice.
- If you include this season and 2019-20, then 2 divided by 0.345 equals 5.797. A goal differential change of +5.797 is worth a win.
- If you do not include this season and 2019-20 and only use 82-game seasons, then 2 shall be divided by 0.348. That equals 5.747. A goal differential change of +5.747 is worth a win.
Of course, no one can improve goal differential by 0.747 or 0.797. In order to meet that 5.747 or 5.797, we need to round it up to 6. This means that a goal differential improvement of 6 is worth a win. The rule of thumb has been proven. Tulsky’s post from 2011 and those prior to him stating that rule of thumb are still correct. This is still applicable in 2022. If nothing else, this is something you can takeaway from this post.
(Aside: Does this mean plus-minus has value? No. Because plus-minus does not cover all goals. Power play goals are not included. Only even strength and shorthanded goals. Further, it is only assigned to players who were just on the ice when a goal happened. It says nothing about whether the player did anything to make it happen.)
(Aside #2: I wonder if I should do this with other team stats. Just to show there is not much of a correlation between, say, hits and winning. It will be a long offseason, so watch this site for that.)
Proving that League Average Goaltending Would Not Likely Be Not Enough for the Playoffs for This Devils Team
The Devils have a goal differential of -47 per the NHL.com league standings after their shootout loss to Seattle. We know that an improvement of 6 goals is worth a win. We know the Devils’ goaltending stinks. What does league average goaltending look like?
This is where I made a mistake from the last time I did this. Hockey Reference does include a running list of league averages for each season. However, I missed that it does not include empty net goals. The shot count from Natural Stat Trick - which I used in previous attempts - for all situations does. And the goal differential count in the standings absolutely includes empty netters. In order to be consistent, I decided to calculate the league average team save percentage after yesterday’s games from Natural Stat Trick’s all situation team stats. The league average save percentage for a team is 90.24%.
That may not seem that great, but that would rank 19th in the NHL compared to the rest of the league. (Note: Averages and medians are not the same! The league median is a bit higher at 90.37%.) Remember that the Devils’ all situations save percentage is a woeful 88.18%. An improvement of nearly 2.06% may not seem like much of a big deal. However, there are over 2,000 shots allowed in a season. Even an improvement of 0.5% really makes a difference. Now, we can do some more math to see where league average goaltending would take the 2021-22 Devils assuming all other things are equal. (e.g. the shot count would not change if the goalies were better, the Devils would not score any more or less goals, penalties are still called as they were, etc.)
- The Devils have allowed 2,310 shots this season in all situations per Natural Stat Trick. The standings has the Devils with 278 goals allowed, but five of those are from shootouts - which are not factored into team or goalie save percentage. We have to use the number of all situations goals allowed from Natural Stat Trick, which was 273 goals allowed. Now we can go into theory.
- If the Devils had a team save percentage of 90.24%, then they would have saved 2,084.54 shots out of 2,310. No one can save 0.54 of a shot, so let us round that down to 2,083. The theoretical 2021-22 Devils would have 2,083 shots saved.
- 2,310 total shots faced minus 2,083 shots saved yields 227 goals allowed.
- Since the Devils gave up 273 actual goals let us subtract 227 from that. 273 minus 227 goals is 46. Therefore, Devils would give up 46 fewer goals with league average goaltending. The Devils’ goal differential would improve from -47 to -1 before factoring in shootouts.
- The Devils lost five shootouts and won three. After including those, the goal differential in the standings would become -3. That gives us a goal differential improvement of 43 goals.
- 43 divided by 6 results in 7.1667 wins. I will be generous and round that up to 8 more wins since a team cannot win part of a game in the standings. That’s the gain from league average goaltending alone assuming all other parts of the 2021-22 season are equal.
The gains from league average goaltending over the collective trash-fire of the real-life goaltending of the Devils would be huge. The Devils would immediately go from being deep in the red in goal differential to breaking just above even. The Devils would have many more victories that are guaranteed. (Note: Technically, they could have even more than eight - or less, we can not ensure the distribution of 46 fewer goals in a game-by-game basis.) The Devils’ record would not be sitting fifth from last place in the league with 58 points before the Seattle game. No, they would have 75 points. They would be in the mix with Columbus (76 points) and the New York Islanders (79 points) for fifth place in the Metropolitan Division. They would be on the edge of the bottom-ten (22nd, they would be ahead of Anaheim and their 72 points) and well ahead of the very worst teams in the NHL. The variation in the formula could mean the Devils could have even more points - or less. For the sake of argument, let us consider this to be their improvement.
I cannot overstate how much the situation would change dramatically in New Jersey with this theoretical improvement. Remember that the Devils finished third-from-last in 2021, finished sixth from last when 2019-20 ended abruptly and missed the Return to Play tourney that even Montreal and Arizona made, and third-from-last in 2018-19. A team on pace to earn somewhere between 89 and 92 points in an 82-game season would be a legitimate improvement over those teams. It would objectively and clearly show that General Manager Tom Fitzgerald and the team he built is heading in the right direction. It would further support that the coaching staff led by Lindy Ruff may indeed know what they are doing. It would keep people more engaged with the team longer and it would absolutely mean a more competitive team in terms of the scoreline instead of just in on-ice metrics.
I would be happy with that since I wanted the team to make some real improvements. That would have been marvelous. I could (and would) champion that as an actual step forward. Praise would be given. I would like to think many more of the People Who Matter would be also very happy about a theoretical season where the Devils may still end up seventh in their division but could clear Philly by nearly 20 points. Fitzgerald’s promise of meaningful games in March could have come to fruition. Far fewer fans would be calling for Ruff to be replaced (Mark Recchi would still need to be dumped like a bad habit) and feel a lot more comfortable with Fitzgerald in charge. More would be willing to be patient with a fourth straight playoff-less season and a sixth playoff-less season since Lou left in 2015 seven seasons ago. It would make a huge difference for the direction of the team and the patience of the fanbase.
However, that would not be enough for playoffs. Not in 2021-22 at least.
The owner of the last playoff spot in the Eastern Conference is the Washington Capitals, who has 94 points as of Saturday night and seven games left to play. In current reality, the Devils are behind the Capitals by 35 points. A gain of eight wins is a big gain, which is 16 points. The Devils would still need 19 more points to catch up to the Caps. That is a massive gap. So much so that I would not consider the Devils to be on any kind of bubble. I certainly do not think the Isles or Jackets are, two teams who have had playoff odds below 5% for months and below 1% for weeks now. (Blue Jackets were mathematically eliminated on Saturday, Islanders could be on Sunday.) Would the theoretical Devils be among the best of the rest in the East with league average goaltending? Yes. Better than the last three seasons? Absolutely. A playoff bubble team? No - even with a generous definition of what a bubble team is.
It is true that eight more Devils wins would mean fewer points for the teams they would apparently beat. That would not matter with the Caps. Yet, even if the Devils ended up sweeping the three games against them in this theoretical world and didn’t give them a point in their one actual win (in real life, they went 1-2-0), then that would make the gap 13 points. That is better but not nearly close enough to think the Devils are some kind of bubble team. Especially not with just two weeks to go in the season barring a total collapse by the Capitals.
I fully agree that league average goaltending would be fantastic for the Devils in 2021-22 and I wish they had it. But I cannot agree it would make them a playoff team outside of the absolute best case scenario involved. The potential gain from it would just not be enough for this season. The gap is that large and a huge improvement to one position with massive weakness is not going to close that gap.
So What Would Have Made the Playoffs Happen for the Devils in 2021-22?
If league average goaltending alone would have not made the Devils jump high enough in the league standings to be a playoff team, then what would it take for them to do it?
Let us go back to the formula and work from there. For simplicity’s sake, I’ll use the one from the 82-game seasons-only formulation. That is 0.348x + 91.56. Let us assume this is accurate and a team with a goal differential of zero would finish the season with 92 points - since a team cannot get 0.56 points in the standings. The Capitals, who have 94 points in 75 games, are on pace to finish this season with 103 points (102.77 to be exact). In order for the Devils to take that last playoff spot, they would need to be better than the Capitals. Therefore, the goal for the theoretical 2021-22 Devils is 104 points. Solving for x in the formula, the Devils would need a goal differential of +35.74 to get that many points. Again, we cannot get .74 of a goal, so let us round that up to +36. That seems like a lot until you look at the Metropolitan Division and see that the Caps have a goal differential of +37, the Penguins have a differential of +39, Our Hated Rivals have one of +44, and Carolina is in first with a +67 goal differential. All inclusive of shootouts, but the point remains: the Devils needed to out-score their collective opposition much more.
The question now becomes: What would it take for the Devils to have a goal differential of +36? We know that league average goaltending in all situations would get the Devils to nearly break even. At -3, they would need an improvement of 39 more goals - a combination or more goals scored and fewer goals allowed.
At the risk of seemingly like a smart aleck, the Devils could aim for better than league average goaltending. To put the all situations save percentage in perspective, Our Hated Rivals lead the league at 91.96% after Saturday’s games. If the 2021-22 Devils had that, then they would have allowed just 186 goals. That would mean an improvement of 87 goals; after factoring -2 for shootouts (remember, I’m assuming all other things are equal), . Shesterkin-driven goaltending alone would put the Devils above the needed +36 goal differential. Specifically, and assuming I did the math right, it would put them at +40. Of course, asking for the league’s best goaltending is more of a wish than anything more realistic like asking for league average goaltending. The point remains: any amount of save percentage above 90.24% would mean the 2021-22 Devils would need fewer goals scored to make up that +36 mark needed. Get a 91% team save percentage in all situations, and the Devils would not need six wins’ worth of goals scored to meet that +36 goal differential. There is a balance involved.
So let us figure out where the Devils could get more goals. It may not be so ridiculous to find, although the task is seemingly ridiculous enough. Consider some of the other known issues with this season’s team:
- The Mark Recchi Experience that is the power play has really held the team back. In real life, the Devils have scored just 34 goals out of 210 power plays for a success rate of 16.2%. That rate ranks 28th in the NHL. If the Devils had a success rate around the league median - 21.1% is 16th as of the morning of April 17 - and assuming they would still just have 210 power plays, then the Devils would have scored 44.31 goals. If we round down to 44, then that is a gain of 10 goals right there. (Or at least one win, in other words.) If the Devils could have drawn more power plays, then that would bump that improvement in goals up even higher. Thanks, Recchi!
- Here is the list of skaters by the Devils organized by goals as of April 17. Scoring depth is an issue and you can see it here. Let me put it in a form of a series of “What if” questions. What if Andreas Johnsson did not get ice-cold after November? What if Tomas Tatar and Pavel Zacha produced more often and performed better in various roles? What if Yegor Sharangovich was hotter on the scoresheet before February and March? What if Janne Kuokkanen did not hit a sophomore wall almost as hard as Ty Smith? What if Michael McLeod was anything more than a bad fourth-line center? What if the Devils never acquired Mason Geertsen and just put him on the ice for a few minutes in 26 games? You get the idea - and I did not even mention the defensemen specifically. Among these what-ifs, you can find a few extra goals with some differences in fortune and performances. Improvements at wing and in the bottom six can provide some notable gains the team may need for next season.
- Speaking of what-ifs, what if the Devils did not play so aggressively for such little gain in 5-on-5? As much as their on-ice metrics look solid, when they get beat in their own end, it is frustrating to watch. The emphasis on overloading on puck carriers and flooding the strong side (puck-carrying side) of the zone has led to a lot of opponents exploiting the Devils’ weak side and/or holes in their coverage. The goal allowed to Soucy in Seattle on Saturday is a great example of this. One of many, many examples from this season. More than just asking for better goaltenders, I think some gains, even marginal ones, could be made in goal differential in the Devils’ favor with a more “traditional” approach to defense. Likewise, not having defensemen activate as much as they do could yield more offensive opportunities for forwards and fewer counter attacking opportunities for opponents. That would also help a bit over the course of a season.
- Every team deals with injuries and I personally think it is not the greatest excuse to lean on when Pittsburgh had massive absences earlier this season and they will be a 100+ point team. Still, Jack Hughes alone had 26 goals in 49 games. That is an incredible average of 0.53 goals per game. Had he played in even just 66 out of the 74 the Devils have played in this season (same number as Nico Hischier’s games played) and maintained the same goal rate, Hughes would have 35 right now. That’s 9 more goals from him alone. Assuming they weren’t power play goals from the earlier point (I think some would be, though), that’s still a gain from just one player. Another reason why I call Hughes, The Big Deal. You could repeat this for some other players (Dougie Hamilton), but Hughes is the standout here.
- And since we’re on the topic of what-ifs, what if the Devils did not collapse in December? That 3-win month more than any other saw the Devils effectively drop out of the playoff picture, get entrenched in the basement, and make many of the People Who Matter start to check out of the campaign. Even a 5-5-3 month - like they had in November - would have slowed down the bleeding out of their 2021-22 season.
- It would also be great if someone else in the Metropolitan Division can take a step back. The Devils need one or two of Washington, Pittsburgh, Our Hated Rivals, and Carolina to suffer so they can rise up. I doubt Carolina will. I doubt Our Hated Rivals will unless Chris Drury botches this offseason. Pundits want to say the Pens’ and Caps’ windows are closed and each season they prove it is not. Still, their excellence means the bar is high for the Devils (and Islanders and Blue Jackets and Flyers) to clear for a playoff appearance. Something that I would hope the People Who Matter do not take for granted anytime soon.
If you put all of these bullet points together, then the Devils are closer to that vaunted +36 goal differential mark. A mark that could be lowered if the Devils’ goaltending is better than league average and teams in the division (and the East) drop off. It is not a guarantee, but this is what I think it would take to get the Devils from where they are in real life to where most of the People Who Matter want them to be in this thought experiment. Personally, I think if the Devils do get all of this, then they may be able to do a bit more than just make the playoffs and be there. But that is a different discussion entirely.
I will admit that this is a lot to request for one season. It is enough to make one just wish for the Devils to find the next Igor Shesterkin and just run back the rest of the season, hoping for the best. The Devils are not likely going to find the next Igor Shesterkin. It would be great but there is no reason to expect it. The main point I am making with this part of this thought experiment is that the 2021-22 Devils have to improve other areas of their roster to push this team towards the playoffs given how the standings have shaken out this season. Better goaltending, even the best goaltending the NHL in all situations, alone would not be enough. This makes sense. The 2021-22 Devils have problems beyond just the goaltender. Anyone honest about looking at this team would know and state that. Beware of those who claim it is only just the goaltending.
Goaltending is the Devils’ biggest issue. Do not misunderstand this post and think that I think goaltending is not their biggest issue. No. It is their biggest problem and it needs the most attention in this coming offseason. I will go as far as to say that Fitzgerald is probably not going to have a job for long if he cannot get at least decent goaltending next season. Not try to get decent goaltending, but actually getting it for 2022-23. But the point, again, is that alone would not be enough to drag the Devils into the playoffs unless a 91-92 point season somehow becomes enough to get in. Better goaltending will not make the power play functional on any decent level. Better goaltending will not address the real issues with depth scoring not being consistent or often enough. Better goaltending would not make the team healthier - even if all that means is more Hughes. Better goaltending would not address the issues with how the team plays regularly. Better goaltending does not open up competition in the division that is necessary to get a playoff spot. The Devils cannot control much about whether the other Metropolitan Division teams can fall apart, but they can address their power play, their scoring beyond their top stars, health, the gameplans the coaches put in place and enforce, and so forth.
Mike wrote recently that the Devils do not seem that far away. Perhaps. Possibly. Hopefully. But if management thinks that all they need to do for 2022-23 season is fix the goaltending - presuming we trust management to do that - then I have little confidence for next season. Which I really would not like to feel, but we have to live in the reality we have and not the reality one wishes they live in.
We should all want the Devils to get league average, or better, goaltending next season. They need it. It would not be enough though. The theory espoused by past hockey analytics people and explained in detail by Eric Tulsky that a goal differential improvement of 6 goals is worth a win appears to be true now as it was in 2011. Again, if nothing else, take that away from this post. Based on that, it is proof that the Devils really do more than just a better goaltender(s) if they want to make a real push for the playoffs in 2022-23. To do that, they will need to address their other issues that have been plain to see in 2021-22 and in past seasons. We should all want more than that for the Devils to avoid missing the playoffs (and finishing 26th or worse in the league standings) for a fifth straight season.
Now I want to know your take. Do you trust Fitzgerald and the front office to fix the goaltending? Do you think they will adequately address the other glaring needs on the team? Will the Devils make the gains they need to propel themselves up the division and show some actual, tangible, legitimate progress? We will get some answers to that later this Summer, so for now, I would like your opinion and speculation to answer these questions. Thank you for reading.