...and they mean less than an Alfredsson guarantee (For Wysh's blog, see here: Wyshynski).
Now, this normally wouldn't be worthy of comment, except for the inordinate number of people across the web on various forums who:
a) Seem to believe they mean something.
b) Seem to believe the oddsmakers are trying to accurately predict a team's final tally of points.
So arm yourselves fellows. This is why Wyshynski and anyone else who believes Las Vegas is predicting playoff teams, or even worse, a team's point total is loonier than Eric Lindros after a Scott Stevens hit (Is that possible?).
The Over/Under is a very basic betting system. The oddsmakers sets a projected total. The better bets that the real total will be over or under the projected total. Usually the odds for over and under are identical or nearly identical. Ideally, the oddsmakers want to set a projected total whereby 50% of the betters will vote over and 50% will vote under. If this is not the case (e.g. more people betting over), the projected total will move up or down (e.g. up) so more people bet on the other side to compensate (e.g. under). The way Vegas makes money is by giving you less than 2 to 1 on your money.
So for instance, if they give you 1.91 for every dollar you bet and 100 people bet 1 dollar with 50% betting over and 50% betting under, then Vegas makes:
$100 (the bets) - $95.5 (the payouts) = $4.5
So you see Vegas doesn't care what the actual result is. All they care about is that 50% of the people will believe the team will do better and 50% will believe it will do worse. Because that for them, is guaranteed money.
But sure you say...maybe they're not trying to predict the standings. But maybe their over/under totals actually do predict point totals.
Well for that, let's examine the Eastern Conference last year.
Vegas's Over/Under (Points) for the 2008/2009 NHL Season
1. Montreal 99.5
2. Pittsburgh 98.5
3. Devils 97.5
4. Ottawa 94.5
5. Philadelphia 94.5
6. Washington 93.5
7. Rangers 93.5
8. Sabres 91.5
9. Boston 91.5
10. Lightning 91.5
11. Carolina 89.5
12. Florida 82.5
13. Atlanta 79.5
14. Toronto 76.5
15. Islanders 74.5
How did the point totals really turn out?
1. Boston 116 (+25.5)
2. Washington 108 (+14.5)
3. New Jersey 106 (+8.5)
4. Pittsburgh 99 (+0.5)
5. Philadelphia 99 (+4.5)
6. Carolina 97 (+8.5)
7. NY Rangers 95 (+1.5)
8. Montreal 93 (-5.5)
9. Florida 93 (+10.5)
10. Buffalo 91 (-0.5)
11. Ottawa 83 (-11.5)
12. Toronto 81 (+4.5)
13. Atlanta 76 (-3.5)
14. Tampa Bay 66 (-25.5)
15. NY Islanders 61 (+13.5)
I thought about doing the Western Conference in detail but saw that Vegas' top 4 were:
109.5 DET (112)
102.5 SJS (117)
100.5 ANA (91)
99.5 DAL (83)
and I mean...why bother.
In short...Vegas' lines aren't predictive at all. They were only within 4 points (+2 or -2 wins) 4 times (in the East). Their point totals had no consistency......a full 40% were over 10 points more or less than they predicted. From a statistical point of view, it's just brutal.
Granted, I did only look at one year. However I'm fairly certain that if you went back to past years, it'd be much the same.
To prove how poorly Vegas' predictive values were, I ran a thought exercise. I challenged Vegas to see who could predict the most teams with 10 points of their actual point totals (e.g. an over/under of 70.5 would mean 61-81). (Note: 10 points is a large margin, but given Vegas' lack of accuracy, I figured it was fair)
I took my mind back to the beginning of last year. And I thought:
"I know San Jose and Detroit are going to tear it up. They have for the past 4 years (Detroit for longer). So I'm going to predict that they'll top 100 points easily. My over/under will be 109.5 points."
"The Islanders and Thrashers are brutal. I doubt either will top 80 points. So I'll set their over/under at 70.5."
Everyone else...no clue. But what I do know is that most teams finish with between 70 and 100 points. So I'll set everyone else's over/under somewhere in the middle at 85.5 points.
Vegas: They were within 10 points of a team's actual standings for 19 out of 30 teams. A pretty shoddy performance if they're actually trying to predict actual point totals. But again, they're not. They're just interested in getting the homers to bet one way, the haters to bet another and the sweet, sweet sound of profit and their over/under numbers will reflect the public perception of the team instead of it's actual potential.
Me: I was within 10 points of a team's final standings for 20 out of 30 teams. My results, were a fairly unimpressive 66% given that most teams finish within such a limited range
But that's right, I got one more right...and I didn't even really try.
So if you read Wysh's blog, or see someone mention this elsewhere and you see people taking it seriously...have a hearty laugh at their expense (HaHa).