It appears the New Jersey Devils have hired Sunny Mehta as their new director of analytics.
— Jonathan Willis (@JonathanWillis) June 12, 2014
While this was certainly uplifting news to Devils fans, one has to immediately wonder how much influence Mehta will have reporting to Devils General Manager Lou Lamoriello.
This was a move clearly pushed for by the Devils’ new ownership, as just over a year ago, owner Josh Harris introduced Sam Hinkie, “one of the leaders of the NBA’s advanced analytic movement,” as the Philadelphia 76ers new general manager and vice president (Harris owns both the Devils and the Sixers). It should not be a surprise he’s moving quickly to ensure the Devils also have someone whose sole responsibility is to provide a different voice in the room when making hockey decisions. The question is, again, how large will that voice be?
After all, this is a team coached by Pete Deboer, who admits he’s “not a stats guy.” David Conte, who runs the Devils’ scouting, thinks advanced stats aren’t worth our time: “I don’t read that stuff. I think it’s all cosmetic. I don’t think it means anything.”
Mehta’s most difficult task might be to cultivate acceptance of hockey analytics in the Devils’ front office before he even shows anyone a spreadsheet. Regarding the actual Devils roster, what should Mehta tackle first?
Using some of the more widely recognized advanced stats today, I looked at the top three areas the Devils could use improvement in. All of these are easily accessible through sites like Extraskater.com, Stats.HockeyAnalysis.com, and BehindtheNet.ca.
To start, the Devils are one of the league’s best possession teams, controlling 53.3 percent of the shot attempts in even-strength situations, which is good enough for fifth in the league. The Devils were only behind the Los Angeles Kings, Chicago Blackhawks, San Jose Sharks and St. Louis Blues, and ahead of the Boston Bruins and New York Rangers. Elite company.
If possession isn’t the problem, perhaps Mehta can start elsewhere.
Looking at the team’s .915 even-strength save percentage, the Devils were 24th this past season despite having Cory Schneider post a .923 even-strength save percentage. Deboer opted to start Martin Brodeur despite his .905 even-strength save percentage. In case you’re wondering where this ranked leaguewide, Brodeur finished 54th among goaltenders playing 21 games or more. One has to wonder that if someone pointed this out to Deboer, the Devils would have made it to the playoffs riding the play of Schneider.
With the Devils re-signing Marek Zidlicky, they look to have another logjam on the back end as they have eight defensemen either signed or hold their restricted free agent rights. On such a strong possession team, we can use a player’s Relative Corsi (Corsi when that player is on the ice minus when he’s off the ice) to evaluate their impact on the rest of the team.
Pending unrestricted free agent Mark Fayne was third on the Devils’ blueline with a plus-1.5 percent Relative Corsi percentage, meaning the Devils control 1.5 percentage more of the shot attempts when he’s on the ice as opposed to when he’s off it. He’s someone the Devils should re-sign immediately as he faces the second-fewest offensive zone starts and second-strongest quality of competition and still comes out ahead in possession. He’s a new-age defensive defenseman and will be highly sought after should the Devils fail to sign him.
But how to make room? Perhaps Mehta will advise a compliance buyout of Devils captain Bryce Salvador. He’s last on the Devils in Relative Corsi at minus-5.2 percent. In fact, not only is he the worst defensemen in terms of Corsi, but he’s more than twice as bad as the next-worst defenseman. He has the worst Corsi percentage on the blueline at 50.4 with the nest worst at 53.1, so there’s quite a gap there as well. Perhaps an analytical approach will prevent the Devils from making a mistake and keeping Salvador at the expense of Fayne.
Despite their strong possession numbers, the Devils’ shot totals were in the basement. Similar to their ESSP, the Devils were near the bottom of the league in their total shots. At even strength, they were 27th in shots for and 29th in goals for, only outscoring the lowly Buffalo Sabres. Clearly, the Devils need to score more goals. Getting more shots on net would certainly help. But if the Devils are so strong in possession, why are they left behind when it comes to shots and goals?
In the 50 games I’ve tracked passes and shots generated by passes for both the Devils and their opponents, the Devils are not nearly as efficient as their opposition. The Devils Shot Attempts Generated, or SAG, was only 45 percent, a full 4 percent lower than their opposition. What does that mean? It means the Devils have to work 4 percent harder than their opponents to generate shots.
Say the Devils and their opponents each generate 20 shots from their passing. The Devils will need to have generated 44.4 shot attempts, whereas their opponents only need to generate 40.8. Is every other team that much more skilled than the Devils? Or is it a systems question? Why is this important? Through the 50 games I’ve tracked, teams that finish with a higher Shot Attempt Generation Efficiency rating win 82 percent of the time. Each winning team averages a 52.3 percent SAGE rating, significantly higher than the Devils 45 percent average in those 50 games.
This is the dilemma Mehta will face from Day One. He steps into a role with one of the league’s best possession teams, but they have finished outside the playoffs two straight seasons in a league where over half the teams make the playoffs. He’ll have to dig deeper than Corsi and isolate phases of the game (zone exits, zone entries, passing, etc.) to solve the Devils’ problems.
In an organization where your head coach “is not a stats guy” and your director of scouting doesn’t “read that stuff,” Mehta has his work cut out for him in order to get his message across.
Read more about the New Jersey Devils:
Ryan Stimson is a contributor at InLouWeTrust on SB Nation and has been a lifelong New Jersey Devils fan. He believes that the future of hockey analytics is in analyzing phases of the game to reveal tendencies of winning teams. Follow him on Twitter @RK_Stimp