For objective hockey analysts, we sure talk about luck a lot. Yet in his annual NHL Review, Alan Ryder reminds us that Wins are 94% predicted by goals for and against, marginal goals totals or goal differentials, suggesting that only the remaining 6% is luck.
And speaking of goals, theyre generated by getting into position to shoot, shooting, and the quality of the shot itself, each of which involve no luck, a little luck, and mostly luck, respectively. Overall, Gabriel Desjardins computed that 35% of a goal is luck.
Given these two findings it probably comes as no surprise that in a separate study, Desjardins, whose approach has been backed up by others like Tyler Dellow, discovered that 38% of winning percentage is luck.
If thats true, there are obviously teams that are getting the lions share of that good fortune. For example, Alan Ryder has always defined a lucky team as one requiring comparatively few marginal goals for each point in the standings, like the 2006-07 Boston Bruins or 2007-08 New York Islanders, while an unlucky team requires far more marginal goals, like the 2006-07 NY Rangers or the 2008-09 Minnesota Wild. Of course, Ryder is focusing only on the 6% of winning that cant be explained by goals. Is there a system that captures the other 32-35%?
The Kubaryk Approach
Recently, Adam Kubaryk published a provocative new look at luck (which you can download here), which essentially involved accepting the closing betting odds on a game as an accurate representation of who should win it, and then finding out which teams strayed particularly far from those expectations, chalking up the variance as luck.
Here are the teams whose actual performance varied the most from Kubaryks expectations.
Unluckiest Teams ExpPts Pts
2006-07 Philadelphia 75.5 56
2007-08 Tampa Bay 87.1 71
2006-07 Edmonton 86.9 71
2009-10 Edmonton 77.6 62
2008-09 Colorado 84.8 69
2008-09 Tampa Bay 78.4 66
2008-09 NY Islanders 72.4 61
2006-07 Phoenix 77.3 67
2009-10 Toronto 83.8 74
2007-08 Ottawa 103.6 94
The 2006-07 Philadelphia Flyers were doubly unlucky, when they lost the first overall draft choice in the NHL draft lottery to the Chicago Blackhawks, and had to settle for James van Riemsdyk instead of Patrick Kane. The lucky team here is the Tampa Bay Lightning, whose talent mismanagement was rewarded with Steven Stamkos and Victor Hedman.
Lets see which teams were luckiest from Kubaryks perspective.
Luckiest Teams ExpPts Pts
2009-10 Phoenix 88.5 107
2009-10 Washington 104.7 121
2006-07 Pittsburgh 90.9 105
2006-07 Vancouver 91.0 105
2008-09 Boston 102.3 116
2009-10 Nashville 88.8 100
2008-09 St. Louis 82.2 92
2007-08 Montreal 94.3 104
2009-10 Los Angeles 91.3 101
2009-10 San Jose 103.7 113
The 2008-09 St. Louis Blues were also the team with the worst odds of making the playoffs that actually didand by a longshot (over Columbus 2008-09 and the 2007-08 Bruins). On the flip side, the biggest shocker of a team missing the postseason was the defending Stanley Cup champion Carolina Hurricanes in 2006-07 by an almost equally sizeable margin, over a handful of teams including last years Calgary Flames.
Is It Luck?
It remains to be discussed if the teams that fall short of Kubaryks expectations are simply unlucky, or if its a reflection of the flaws in computing the odds of winning on a game-by-game basis. If certain factors are indeed being misweighted, any gambler who could isolate them could make a fortune!
If it is indeed lucky bounces, then it might be reflected in a teams shooting percentage and/or save percentages. Except for 2006-07, SkinnyFish has compiled the statistics we need here and Kubaryks unlucky teams are indeed in the leagues basement in even-strength PDO (shooting percentage + save percentage), while the lucky teams are at the top. The unlucky teams average a lousy PDO of .986, and over half of them were in the bottom two, while the lucky teams average a solid PDO of 1.014, and half the teams were in the top two.
While PDO can cover off most of the goal-based luck, Ryders data can get us the rest, and it also backs up Kubaryks theory. As a group, the unlucky teams required an extra 0.27 marginal goals per point relative to the teams in the lucky group. In fact, last years Predators are considered the fourth-luckiest team in the past four seasons by Ryders measure, and last years Oilers the fifth-unluckiest.
It may sound discouraging that a team can assemble the right talent, adopt the right systems, and execute the right plays, and yet 38% of their fate still remains outside their hands, and at the whim of a bouncing puck, a broken stick, or the blink of an official. On the other hand, perhaps thats what makes the game so exciting.
Thanks to analysts like Ryder, Desjardins and Kubaryk, at least we can begin to track and quantify lucks impact, and this greater understanding can ultimately only lead to a greater appreciation of the game.
Robert Vollman is an author of Hockey Prospectus.
You can contact Robert by clicking here or click here to see Robert's other articles.