Devil’s Luck

(Andy Marlin, Getty Images)

(Andy Marlin, Getty Images)


In 2012, the New Jersey Devils made a surprise run to the Stanley Cup Final and were the only team to win more than one game in the playoffs against the Los Angeles Kings. What makes the run particularly surprising is that the Devils missed the playoffs in the previous and subsequent seasons.

The Devils finished the 2010-11 season with 81 points, jumped up to 102 points in 2011-12, then dropped down to 48 points in 2012-13, an 82-point pace. It’s a remarkable trio of seasons, as the Devils jumped from 11th up to 6th in the Eastern Conference, then dropped right back down to 11th.

It’s enough to make you wonder exactly what happened in those three seasons and what will happen in the future, with the Devils losing both Ilya Kovalchuk and David Clarkson, while also adding some significant pieces. To go from well out of the playoffs to the Stanley Cup Final and right back to being well out of the playoffs is a stunning reversal of fortune, and it’s fortune itself that I want to investigate.

Fortune, aka. luck, is a bit of a controversial topic in hockey. We have all seen luck affect a hockey game, whether it’s a centring pass deflecting off a defenceman’s skate into the net, a puck fluttering on end that causes a defender’s clearing attempt to go over the glass for a game-changing penalty, or a shot destined for the top corner hitting the knob of the goaltender’s stick and off the post. Everyone accepts that some games are partially decided by luck in some form.

Most people believe that luck washes out over time, with idioms like “Fortune favours the bold” or “You make your own luck” showing the human desire for life to fit the narrative that we get what we deserve. We want to believe that the best team wins the game or, at least, the majority of the games. But the truth is that there is still a fair amount of luck involved even over an 82-game season. Or, if the word “luck” bothers you, call it “applied chaos” or some other euphemism.

Skill trumps luck given enough time, but a dose of good or bad luck can still make a significant difference in the standings and playoff results. Consider the 2012 Kings: how did a team that finished eighth in the Western Conference steamroll their opponents in the playoffs? Were they a good team that was just unlucky during the regular season or were they a mediocre team that suddenly saw a run of good luck? Or is it one of those combined with some other factor?

Over at Hockey Abstract, Rob Vollman has a fun visualization of luck as it pertains to NHL teams. He identifies five basic ways that luck impacts a team: “injuries, shooting and save percentages (which is called PDO when added together), their power play and penalty killing percentages (which is called Special Teams Index when added together), their record in one-goal games (regulation time only), and their record in overtime and shoot-out games.”

He’s quick to clarify that many of these involve a combination of skill and luck to varying degrees, but all of these are of particular interest as they’re generally not repeatable from year to year. While something like injuries seems obviously random and wouldn’t persist from season to season, the same is generally true for these other factors that seem like they would have more basis in skill than luck. Shooting, save, power play, and penalty kill percentages tend to fluctuate wildly, however, and the same is true of a team’s record in one-goal games and past regulation. Certainly, very good teams tend to finish higher than very bad teams, but there isn’t much consistency from year to year.

Vollman has combined these five factors into a somewhat haphazard Luck Score. The visualization on his website lets you play with how each factor is weighted, depending on whether you think PDO is more or less dependent on luck than special teams, etcetera. But just looking at the basic score in relation to the New Jersey Devils over the last three seasons is interesting.

In 2010-11, the Devils were below .500 with a 38-39-5 record. Their Luck Score for that season was fourth worst in the league. The bulk of their bad luck came on the injury and PDO fronts, as they actually had a very good record in overtime and the shootout. Missing Zach Parise for 69 games was the culprit injury-wise and a league-worst shooting percentage of 6.31% at even-strength combined with a below-average even-strength save percentage gave them the second worst PDO in the league. Mediocre special teams and a slightly above-average record in one-goal games round out their Luck Score.

In 2011-12, they added 10 wins, finishing 48-28-6, the ninth best record in the NHL. Their Luck Score went from fourth worst to fourth best, with minimal impact from injuries, an improved PDO as their team even-strength shooting percentage skyrocketed to league-average (though their save percentage remained well below average), the league’s top penalty kill, a decent record in one-goal games, and the best record in the league in overtime and the shootout. The Devils went 16-6 in overtime and the shootout in 2011-12, meaning a third of their wins came past regulation.

Some will want to attribute that to the Devils being particularly clutch or determined that season, but given how overtime and shootout records fluctuate from season to season, it seems likely that there was a healthy dose of random chance involved. The Devils were a good puck possession team that season and certainly earned their success and trip to the Stanley Cup Final, but they were a good puck possession team in 2010-11 as well. Going from one of the unluckiest teams in the league to one of the luckiest certainly helped.

(Sidenote: the Nashville Predators had the highest Luck Score in the league in 2011-12, easily making the playoffs despite well below average possession statistics. They finished second last in the West this past season with still below average possession and the fourth worst Luck Score in the league.)

That brings us to the 2012-13 season, which, thanks to the lockout, was just the 2013 season. The shortened schedule meant that luck played an even more prominent role in the standings. 48 games is barely enough time for skill to win out over luck, while some might argue that it simply isn’t enough time, with randomness playing more of a factor than skill.

As for the Devils, luck was not on their side. They finished with a 19-19-10 record, good for last place in the Atlantic Division and 11th in the Conference. Their Luck Score, meanwhile, dropped from fourth best to second worst, ahead of only Florida who had truly execrable luck. While the Devils had a couple significant injuries, missing Ilya Kovalchuk for 11 games and Martin Brodeur for 10 games, they were even more unlucky elsewhere.

The Devils’ shooting percentage once again dropped down to one of the lowest in the league at 6.42% at even-strength, while their save percentage remained mediocre. Their penalty kill percentage dropped from first in the league to 16th, while their power play went from 14th to 21st. The biggest change, however, came in their record in overtime and the shootout. After finishing with the best record in the NHL in 2011-12 at 16-6, they finished 3-10 last season, easily the worst record in the league.

How does a team go from being the best in the league past regulation to worst in the league in the space of one season? Certainly, the loss of Zach Parise was significant, as he is a near-50% shooter in the shootout over his career, but that on it’s own doesn’t explain that big a shift. A large portion of it has to be considered luck, particularly with the shootout generally being little better than a coin toss to begin with.

The Devils deserved better last season, finishing with the third highest Fenwick Close in the league, a statistic that represents puck possession when the score is close and has been shown to have a high correlation with winning percentage and is highly persistent from season to season. Ahead of New Jersey in that statistic: Los Angeles and Chicago. Right behind New Jersey: Boston and Detroit.

The Devils should have been in the playoffs, if not for an aging goaltender past his prime, a terrible shooting percentage, and terrible luck past regulation. This coming season, however, with one of their best puck possession forwards gone in Clarkson and their best goalscorer gone in Kovalchuk, it may not matter how good their luck is in 2013-14. The additions of Cory Schneider, Michael Ryder, Jaromir Jagr, and Ryane Clowe may not be enough.

Comments (9)

  1. I think it’s great that, given how the Leafs ended their season, the luck visualizer has them listed as the luckiest team in hockey last year.

  2. The Devils were incredibly unlucky this season. I’ve never seen anything like it. Every game they’d hit the post 2-3 times and it wouldn’t go in. Brodeur got injured and Hedberg was atrocious. The team controlled posession almost all game, but just couldn’t finish. Clarkson started hot, but did virtually nothing the last 2/3 of the season. Add to that, the weird inability to win a shootout (Kovalchuk and Elias regressed to the mean bigtime after an insane run on the shootout the year before). This team was better than it looked in the standings. That’s not to say that they should have been back in the finals, but they should have been a playoff team. Are they now? Very hard to say given all the changes. Jagr’s a good posession guy and still has some offense left in the tank, Clowe could be an upgrade from Clarkson if he’s healthy (a very big if), and Ryder should give some solid offensive support. Ultimately, this team will either win or lose a lot of one-goal games. Having Schneider to fall back on if Brodeur isn’t looking strong (or gets injured) will help a lot. There’s also a bit of a wildcard on the team in Reid Boucher, who broke Stamkos’ goal scoring record in Jrs – if he can translate that to the Pros (he looked solid in a short AHL stint), he could contribute. If the Devils can still play that strong posession style and get solid goaltending, they may surprise a lot of the people that assume that they’re terrible without Kovalchuk.

  3. Since that luck meter (Vollman’s site) also includes skill as well, the question is, how much of the luck measure is actually b/c of luck? If you include all seasons within Vollman’s luck analysis (2008-09 to last season), the most “lucky” teams are the recent perennial playoff teams (Van, Bos, Pit, SJ, WSH, Ch, Phii) while the least lucky teams are those who have been not so good over the past few years (NYI, CBJ, TOR, COL, EDM, FLA). I would assume with this larger sample size that we would see a more sporadic listing of teams if it was actually all luck based, but it looks like the list just backs up the old adage that you have to be good to be lucky, and perhaps skill is a larger element of these luck metrics then some would have us believe.

    • That’s right–teams differ in their “true talent PDO” because of tons of offensive skill (Washington, Pittsburgh, for example) and/or goaltending (Vancouver, Boston…Columbus with Mason at the other end, perhaps). So assuming a PDO of 1000 isn’t exactly right. But within one season, the variance in PDO thanks to luck is bigger than the true talent variance, so it’s hard to judge teams’ PDOs based on a single season (especially when said season is only 48 games long).

      Also, I’d note that saying the teams that win the most have the highest PDOs is after-the-fact–winning is almost entirely GF and GA, which are heavily driven by percentages (PDO), so it’s not at all surprising that teams that have won the most games over the last few years have the best PDOs.

      • So what you’re saying is that over the log run PDO is not as much based o luck as it is o skill. What about the rest of the luck factors on that site? I tend to think luck is an overused terms in the analytics community and is a bit of a cop out, even though I know most of hockey can’t be explained by stats.

  4. Yeah, they were pretty unlucky this last season. As much as I loathe the phrase, the devils did “deserve to win” a lot of the games they lost during the middle stretch of the season. I love the guy, but Hedberg couldn’t stop a parked car during stretches. With even minimally better play from him and some better luck at the other end of the rink and the Devils would have been in the hunt this year.

    That being said, Clarkson may have been their possession leader, but I think he’s pretty easily replaceable. A lot of his time with the puck is idle and a lot of Devils fans know nothing is going to come from it. It’s either a toe drag, the clark-around, or a bounce pass to the d. Yeah, it’s better to have the puck than chase it, but when Jagr has the puck he has the talent to actually create something from it, instead of just banging it on net and hoping it goes in. He may be a shell of what he used to be, but he’s still Jagr and his strengths are well established. Losing Kovy is a different beast, but cleaning up the financial mess his contract might have created will benefit the team in the long run. The Devils haydays were also when they didn’t have a roster that looked that scary on paper. So maybe returning to their roots with a solid, if underrated, roster will help.

  5. advanced stats have no place here.. this aint baseball.. dunno how many times we gotta say this

  6. The Devils were definitely unlucky. You could add that the Devils had the best shot differential in the league, and the next dozen or so teams on that list all qualified for the playoffs.

    However, I don’t agree that losing Clarkson and Kovalchuk will have that much of an impact. Clarkson had one good season (30g in 2012) surrounded by mediocrity (every other season) and his career stats show he can’t pass the puck to save his life. Puck possession means little when you do as little with it as Clarkson. Kovalchuk’s departure will be felt, but he has been a shadow of his former self since the 2012 playoff run. He has been injured often and his motivation and effort left a lot to be desired. Ryder and Clowe are far more consistent, even if their scoring talent ceiling may be lower.

    The Devils are at their best with depth scoring, not when they’re relying on top end talent to carry them.

    • The loss of Kovalchuk will hurt, but I think that Deboer overplayed him, which wore him down. He was playing minutes like he was a top pairing D-man…that’s too much for a forward. Clarkson couldn’t pass and was very hot/cold. We’ll miss the goals for sure, but in truth, we’ve picked up three guys that replace a lot of the goals. With new ownership in place and Grabovski and Brunner still out there, we might not be done yet.

Leave a Reply

Your email address will not be published. Required fields are marked *