Welcome to one of the few annual posts I make. Usually after each year, I write up how faceoffs went during the season to get a gauge on how the team was doing with trying to gain initial possession. One of the other posts I usually create annually has to do with luck. I started doing this post years ago after reading about how Rob Vollman tried to quantify luck in hockey in his book Hockey Abstract. I was intrigued by it, and so I applied his standards for determining luck to the New Jersey Devils. Since then, it has been an enjoyable topic to dive into.
According to Vollman’s work, there are 5 categories that help to determine whether a team was lucky or unlucky over the course of an 82-game season, and they are: shooting and save percentages (PDO), the impact of special teams (STI), injuries in the form of the total cap hit of injured players (CHIP), a team’s record in games that go beyond regulation, and a team’s record in one-goal games. All 5 of these stats, to a degree, are luck-based, and when you combine them, give a decent snapshot of a team’s luck. Of course, it isn’t perfect, as there is no perfect way to quantify luck. But it still works to provide a general picture of a team’s luck, so let’s see how it was for NJ this past year.
PDO is perhaps the easiest of the 5 to find out and see when compared to the rest of the league. PDO is simply the combination of a team’s save and shooting percentages. It sort of works like OPS works in baseball, in that it takes two disparate stats and puts them together to paint a better picture. With PDO, the idea is that a luck neutral team should be around 1.00. Anything drastically under 1.00 would constitute an unlucky season, and anything decently over 1.00 would showcase luck. Of course, if you just plain have bad goaltending, like NJ did last season, then it is going to be sub-1.0 because of a lack of talent as well, but the thought process is that over 82 games, a team of average luck should end up with something like a .910 team save percentage and a 9% shooting percentage, which equals the luck neutral PDO of 1.0. Skew the numbers however you like from there, but it ends similarly.
According to Natural Stat Trick, the Devils had a team shooting percentage of 8.57% and a team save percentage of .897 (lol) for a PDO of 0.983, good for 31st in the league, ahead of only Seattle. But again, that shooting percentage ranked NJ 13th in the league, so this PDO number is solely thanks to that sub .900 save percentage. And it is up to you to say how much of that was due to bad luck. I would have to say that some of it is indeed bad luck, as had Jonathan Bernier stayed healthy, it would have certainly been higher, but in the end, the guys who were out there were simply bad, sub-replacement level many nights. Plus, Bernier being hurt will be showcased later in the injury section of this luck review, so in the end, while PDO says the Devils were very, very unlucky last season, I don’t necessarily agree. They just had extreme difficulty finding any competent goaltending.
Special Teams Impact
STI works just like PDO does, but instead of combining shooting and save percentages, it combines a team’s power play and penalty kill percentages. The closer a team is to 100%, the more luck neutral they were. Conversely, a super low number would indicate some unlucky business, while a high number might mean some luck. Again, not all special teams is luck. Some teams have great setups, great movement, or great snipers. That all requires skill and teamwork. But as you also know, not all goals are created equally, and some go in off of lucky bounces and whatnot, so a team that can’t score at all on the power play might also be snakebitten, aka unlucky.
According to NHL.com, the Devils had a power play percentage of 15.6% this past season, and a penalty kill percentage of 80.2%. Add those together, and the Devils’ overall STI was at 95.8%. That, once again, is considered unlucky. This is mostly due to that low PP%, which ranked 28th in the league. Their PK percentage was 14th in the NHL, right near the middle, so it was not a major impediment, but they did have trouble scoring with the man advantage. Which, let’s be honest, seems to be a struggle every year for this team. However, you have to like the power play units that are being thrown out there now, with some real quality offensive talent, so it isn’t like they are icing some B-level talent on the power play. They just couldn’t put it together last year, and some of that could be considered unlucky, especially over the course of 82 games. I would be more willing to chalk this stat up to being unlucky than I would the PDO we just saw.
Next, and perhaps most interesting, is trying to determine luck with respect to injuries. The best way that Rob Vollman determined to figure this out would be to look at the cap hit of injured players. This takes the per game cap hit of players that are not playing because of injury and compiles it. This is an extra layer beyond simply looking at man games lost due to injury. If Team A and Team B both had 100 man games lost due to injury, you might think their luck the same. But if Team A lost their superstar for 50 of those 100 games, while Team B only lost replacement-level players, well that makes a huge difference. This is where the cap hit comes in. It isn’t perfect, as there could be superstars still on their ELC deals, but it’s still better than not looking at it at all.
One great place to check out this stat is the “NHL Injury Viz” blog. Here, you can find beautiful charts that showcase CHIP from this past season. You can see the overall view, per team, compiled together, and you can also look at each team and their per game CHIP. Overall, the Devils had the 7th highest CHIP last season, which is definitively unlucky. Their overall CHIP equaled $16.44 million, whereas say, the Rangers were at $6.93 million, ranking them as the 4th luckiest in this stat. Check out the Devils’ specific chart (use the link above to see it on the blog itself):
You can see that in the overwhelming majority of games, the Devils were over the league average in per game CHIP. In fact, between games 26 and 82, they were only under the league average 3 times. The blue lines represent forwards who were injured, the red line is the defensemen, and the gold is the goaltenders. Early on, they were pretty lucky, then of course Dougie Hamilton was hurt, Jack Hughes was hurt, and that top gold line you see for most games is Bernier. Even Nico was out for games here and there, and his cap hit was a huge blow to this chart. In the end, there is no doubt that with injuries, the Devils were unlucky last season.
The final two stats here constitute game scores. This first one is a team’s record in games that go beyond regulation. This, of course, can be either games that end in overtime, or games that go to a shootout. So many of these endings are luck based. You get the right bounce in 3v3, and all of a sudden it's a 2-on-0 the other way, and game over. Yes, there is skill involved in these times, but over the course of 82 games, teams that have records that are significantly better or worse than .500 in these games definitely had luck come into play.
Last year, the Devils played in 17 games that went beyond regulation. In those games, they had an 8-9 record. That, in reality, is as close to luck neutral as you can get without being directly on it. They actually had a positive post-regulation record heading into the last month of the season, then proceeded to go 0-4 in post-regulation games in April. So in reality, you cannot look at these games and say they were lucky or unlucky.
Finally, we have one-goal games that ended in regulation. The reason to include this is similar to the reason for post-regulation games. In any one-goal game, you’re looking at a bounce one way or another that seals it. There are so many close calls in any game, and to have it end within one goal, either way, can often be due to luck. Again, not all one-goal games might be solely attributed to luck, but over a full season, there will be some luck involved in these games. So really, any team that is way above or below .500 in these games can look to luck as one factor.
Here, we see some major level of unluckiness for the Devils. They played in 14 one-goal games that ended in regulation. In those games, the team was a whopping 2-12. You can try to spin that however you want, but some bad luck is absolutely involved there. You’re telling me that a bounce here or there couldn’t have gone the other way? Throw in a couple of those, and they would still have been 4-10 or 5-11 in those games. That really is next level in terms of poor luck, almost on the level of the Devils team from nearly a decade ago now that couldn’t win a single shootout. Yes, you probably wouldn’t expect a bad team to have a winning record in these games, and some of this was absolutely due to bad play, especially bad goaltending, but still, come on. They were also unlucky here.
So in the end, on the surface, the Devils had 4 categories where they were unlucky, and one where they were luck neutral. Even if you want to chalk up PDO to bad goaltending, and I would agree with that sentiment, you’re still looking at 3 downright unlucky categories in STI, CHIP, and one-goal games. And personally, I think that the CHIP and one-goal game numbers show a team that was undoubtedly unlucky. I don’t think that a change in luck would have meant a playoff berth. They were a bad team no matter how you shake it. But that unlucky streak definitely helped them end as low as they did, and it helped them win the lottery. If they were purely luck neutral across all 5 of these stats, they would have been much closer to the middle of the league in the standings and would have had a way lower percentage to win the draft lottery. So, in a way, you can make the argument that the team was actually lucky, as without the bad luck, they would never have picked second overall and gotten Simon Nemec.
And that’s it for my yearly dive into luck and the Devils! What do you think about these numbers? Do you agree that they were kind of unlucky this past season, especially with regard to CHIP and one-goal games? Or, do you have other reasons for this outside of luck? Do you think there are other categories we can look at that can show whether or not the team was lucky or not? Did they have anything that was lucky last year at all, outside of a lottery win? Please leave your comments below, and thanks for reading!