The Wisdom of Crowds and Las Vegas Lines
My football pool has 107 people who have made picks every week, which is a decent sample. It's safe to assume that while few of these folks are professional gamblers, most of them are football fans of one kind or another. I was curious to see how successful we are as a group at picking games, so I took the average correct picks for each week. There are no ties and 16 games per week so far.
Week 1: 8.1
Week 2: 7.2
Week 3: 7.7
Total: 23.0 out of 48
As a group we've made a total of 5,136 picks at 47.9% accuracy. It will be interesting to see how closely we track to 50% as the year proceeds. Based on this rather limited set of data, it seems that the people who make the lines know what they're doing.
posted by Jerry @ 12:07 AM
2 comments
2 Comments:
I subscribe to a stats journal- a few years back they did a study on “how good were the lines” (i.e. leading to 50-50 outcomes, defeating the “wisdom of the crowds”).
They found two. Big favorites in MLB tend to be underbet (I think 12-14 and worse). And gigantic point spreads in college football (five TDs and more) tend to be not generous enough.
Top jockeys at regional tracks also tended to be overbet- people just looking for action take the familiar. But the take-outs were too big to overcome the bias.
I am not surprised the disconnections are at the extreme outliers.
I’ll try and find a link.
did you read the book entitled "the wisdom of crowds"? it is a good one. you can also use those people in your sample to predict the weight of an ox. this is useful if you need to know how much your ox weighs.
i made my class guess how many pages were in the complete works of william shakespeare and the average was closer than any one student's guess. i look forward to your math.
Post a Comment
<< Home