In the lead-up to Game 1 of the Stanley Cup Final on Wednesday night, NBC’s television coverage noted on several occasions that the Los Angeles Kings were “overwhelming favorites” to beat the New York Rangers in the series. And if you pay attention to informal straw polls such as this one, it seems like that’s the case. As FiveThirtyEight’s Eric Tulsky pointed out in this preview, the media have been near-unanimous in predicting the Kings will win the Cup. Indeed, the Kings prevailed in Game 1.
But they didn’t dominate, and the percentage of respondents who predict a certain outcome is not equivalent to the predicted probability of that outcome — especially when there’s little to no accountability for failed predictions, and the real possibility that herd behavior will produce non-independent picks. The most accurate and unbiased predictor of a given sporting event is usually the Las Vegas betting line, not a pundit. And according to Vegas, the Kings did not enter the series a markedly dominant favorite relative to other pre-series Stanley Cup front-runners.
Unfortunately, archived futures odds like these, which gave the Kings a 59.6 percent probability of winning the Cup, are not available for past seasons. But the useful site SportsDataBase.com does offer money lines for individual NHL playoff games going back eight postseasons. Using those for Game 1 of every final since 2007 (combined with the assumption that, in the NHL, a home team will beat an evenly matched road opponent about 55 percent of the time), we can infer the probability of each team winning a game at home and on the road — and thus the probability of winning the series.
Prior to the Kings’ Game 1 victory, they had a -146 money line in Vegas, and the Rangers were listed at +135. Converting those numbers to probabilities and accounting for the “juice” that bookmakers add to each line to make a profit, Vegas thought that Los Angeles had a 58.2 percent chance of beating the Rangers on home ice. Armed with that number, we can rearrange Bill James’s log5 formula to extract the implied probability that the Kings would beat New York at a neutral site (53.3 percent), on the road (48.3 percent) and in the series (58.7 percent). (The difference from the 59.6 number listed earlier is due to using data from different sportsbooks.)
Using SportsDataBase.com, we have data for eight Stanley Cup Finals played. If we apply the method above to them, the average expected win probability for the favorite in those eight series was 61.3 percent — higher than the Kings’ pre-series odds this year. Here’s the rundown of all eight series:
Instead of being “overwhelming favorites,” the Kings were actually less favored than the typical Cup front-runner. I expected them to be favored before the series began, and their odds have certainly improved after winning Game 1. The idea that this is a notably one-sided matchup, though, just isn’t supported by the market.
