Hockey Outsider
Registered User
- Jan 16, 2005
- 9,213
- 14,714
Purpose - This thread is inspired by a recent observation from @overpass - namely, that we can assess how many Stanley Cups a player won, relative to the expectations of his team's strength, by taking into account the pre-season odds.
Method - I took the pre-season Stanley Cup odds from hockey-reference.com. They have a complete listing going back to 1985. I converted the odds each year into percentages. I adjusted the percentages so that they sum to 100% each year. (Without this adjustment, the percentages would add to over 100%, as there's a profit element built in. Interestingly, the profit element has decreased steadily over time - perhaps this is due to more online competition that drives the spreads downwards - I don't gamble on sports so maybe someone who's done that can confirm if that's correct).
As I usually try to do, I'll offer some self-criticisms of this approach:
Method - I took the pre-season Stanley Cup odds from hockey-reference.com. They have a complete listing going back to 1985. I converted the odds each year into percentages. I adjusted the percentages so that they sum to 100% each year. (Without this adjustment, the percentages would add to over 100%, as there's a profit element built in. Interestingly, the profit element has decreased steadily over time - perhaps this is due to more online competition that drives the spreads downwards - I don't gamble on sports so maybe someone who's done that can confirm if that's correct).
As I usually try to do, I'll offer some self-criticisms of this approach:
- This is a fairly crude analysis. It only looks at whether a team wins the Stanley Cup or not. It doesn't take into account whether they lost in game 7 of the Stanley Cup finals, or if they were swept in the first round. An obvious way to improve this analysis would be to get the odds series by series.
- Although I do some individual analysis, it's questionable if it's truly appropriate to apply this to individual players. It's likely not appropriate to ascribe all of the differences to the players I study (i.e., a team might lose due to poor goaltending, which likely isn't the top scorer's fault).
- The analysis is based upon pre-season odds. It would be more instructive to use odds right before the playoffs began, as that would take into account significant injuries, mid-year trades, etc.
- The purpose of gambling organizations is to maximize their profit (not necessarily to assess the strength of teams). To the extent that there's an efficient market, these would give similar evaluations. If you believe that gamblers exploit particularly optimistic fanbases, the results may be skewed.