clock menu more-arrow no yes mobile

Filed under:

Why Does Everyone Think The Devils Were “Lucky” Last Year?

The notion that the Devils were “lucky” last year is running rampant through the fanbase, forecasters, and message boards. But is it true? Did the Devils see any tangible benefit from simply getting lucky?

If you buy something from an SB Nation link, Vox Media may earn a commission. See our ethics statement.

Toronto Maple Leafs v New Jersey Devils Photo by Bruce Bennett/Getty Images

The Devils got off to a blazing fast start last year and people correctly said that they should regress because of a few really unsustainable performances — Schneider’s pre-injury renaissance, Jesper Bratt’s ridiculous start, and Brian Gibbons unsustainable scoring rate, were among the culprits of the suspicion. This was a suspicion that was, by the way, well-founded. As of December 27th, the Devils were 22-9-5, gainng 68.1% of the possible points. From that point on, they were 22-20-4, gaining only 52.2% of points. On a season long scale, that means they went from a 111-point pace to an 85-point one (remember that number).

This seems to have validated people’s thinking of the team. The Devils have the 21st best Stanley Cup Odds and in what’s probably the best season preview series available, The Athletic’s Dom Luszczyszyn has projected the team to finish 25th in the NHL with 85.3 points. In other words, Dom expects the Devils to be exactly the team they were in the 2nd half of the year. Emmanuel Perry’s model has us projected for 92 points — a modest 5 points drop from this past year, though still 2 spots out of the playoffs. After I wrote this piece on why it was fine the Devils did nothing this Summer, I noticed a similar skepticism in both the comments (1, 2, 3) and on twitter (another one) all referencing the same refrain that the Devils seemed “lucky” last year. In Dom’s piece he says:

“Eclipsing last season’s total of 97 is much less likely at just seven per cent. Essentially, regression is in the forecast and it’s unlikely the team can top last year’s surprising season. That’s something the Devils should be expecting this year. A lot went right for them in 2017-18, but I’m not sure many in the fan base were expecting it to go as well as it did.”

I want to challenge this notion because I don’t believe the Devils were consistent beneficiaries of any quantifiable form luck we’re of which we are currently aware.

My fellow AAtJ columnist, Alex Potts, will have to forgive me for blatantly stealing his format here, as this is a thing he has looked at once or twice, but I think it’s relevant to the current portrayal of the Devils. In Hockey Abstract, author Rob Vollman lists 5 common measures of “luck”: 1) PDO, 2) CHIP, 3) Special Teams, 4) OT record, 5) 1-goal-game record.

PDO

PDO is a team’s save percentage plus their shooting percentage. It has proven to be unsustainable and likely to regress to 100 over time. An adjustment was made by Manny Perry of Corsica to account for the danger of shots — he creates an xPDO which is what you’d expect based on the shot data, and dPDO (PDO - xPDO). A high positive means luck and a high negative meands unlucky. According the Corsica’s team stats page, the Devils 5v5 PDO was 99.6% (19th) and their dPDO was -0.66 (22nd). In all situations, the dPDO changes to -0.02 (-0.14, adjusted). The Devils had slightly below league-average puck-luck in 2018.

Injuries

CHIP is the “Cap Hit of Injured Players.” Created before WAR measures, and addressing the problem of trying to quantify the value of someone who is playing, by definition, an incomplete season, Cap Hit is an attempt at trying to find the “value” of an injured player. If you want to look at the Devils in this particular metric, check out NHL Injury Viz (Devils were 13th highest). However, to get a better idea of the actual value of the players performance rather than their contract, I’ve downloaded their database and merged it with Corsica’s WAR model. I used the past 2 years of data, and multiplied the players WAR per game times the number of games they missed. The results are below.

The Devils lost the 7th most wins from injured players. Injuries to Marcus Johansson and Kyle Palmieri were certainly felt up and down the lineup and I believe the injury to Cory Schneider may actually be underestimated here because he was never the same after the injury even though he returned from it. The Devils got bad injury luck

Special Teams

According to Hockey-Reference, the Devils were 10th in PP%, 7th in PK%, and 8th in combined percentage. Their PP expected goal numbers are not particularly good and their was a fair deal of reliance on Hall’s ridiculous PP performances so I think it’s fair to say some luck was involved there. The PK was simply extremely good. Their xGF% was the highest in the league and the only notable player lost from the unit is Brian Gibbons. It would not be shocking at all to see them repeat that performance. The Devils were likely in the top third of the league in special teams luck

One-Goal Games

The Devils were 11-9 in overtime and the shootout, and 14-16 in regulation one-goal games, making for a total of 25-25 in all one-goal games. This is exactly league average. The Devils were not notably lucky or unlucky in one-goal/overtime games.

The Devils Were NOT Lucky Last Year

The Devils lost a lot of time to injured players, they got terrible early hockey from Kinkaid and terrible late hockey from Schneider. They slightly outperformed their powerplay expectation and did not benefit from a disproportionate amount of points in one-goal games. So why is it that everyone thinks the Devils were lucky last year? Basically, because it “feels” like they were lucky. When someone does well that isn’t supposed to, the stats stick out of the scorecard like a sore thumb. When they do poorly, it can sometimes be very difficult to notice if you’re not actively looking. When the 29 year old AHLer and the 19-year-old 6th round pick both spend time atop your scoring list, people take notice. Less remember that they ended up disappearing for the entire second half. There was a lot of both of these things last year, but people remembering only the success clouds the reality of last season.

People remember Schneider, Bratt, and Gibbons ridiculous starts, but not that Schneider and Gibbons got injured and never really came back, and Bratt got so bad that they sat him at the end of the year. People know that Hall has never been a 90-point player, but seem to forget just how good he’s been this whole time. People remember Kinkaid’s absolutely unreal performance down the stretch, but they forget how TERRIBLE he was early — his final Sv% was 0.913 (career: 0.912).

Listen, I get it. The Devils weren’t supposed to be good last year and it’s surprising that they were. But the fact that we didn’t expect it doesn’t mean that it was a fluke. The statistical evidence that the Devils success was due to falling on a fortuitous spot in the probability distribution is scarce. And most of their best players are either in their prime, or before it.

The Devils weren’t lucky last year — they were good.

Your Thoughts

Do you think the Devils were lucky? Do you think their success from last year is sustainable/surpassable? Thanks for reading as always, and leave your thoughts in the comments section below.