Diversity Prediction Theorem

[Inspirited]

Do you have thoughts about this in relation to sports betting?

[hutennis]

It looks like a perfect explanation why implied probability does not need to be an actual probability to be a very hard to beat benchmark.

As long as difference between implied and actual is tiny enough (and video shows that it always is) to fit within margin protected by juice and stay in that margin most of the time ( my guess it is there 95+ percent of the time) , making money in SB by trying to outsmart collective wisdom of the market is very hard task indeed.

[Salmon Steak]

So using multiple models, looking at the same game in a different ways, will produce a better result. I read that book. I thought it was a good book. It made me think about my sport betting differently.Thanks for posting the video.

[hutennis]

So using multiple models, looking at the same game in a different ways, will produce a better result.

I think you are missing a point a little bit.

Using multiple models, looking at the same game in a different ways, will have a great chance to produce a better result if you are competing with a single individual who uses a single way of looking at the game. Then your diversity will give you an edge. In a sense, by virtue of using deferment models you become a "crowd" which single guy can not beat.

But since no matter how many models you personally would use it still will be only a drop in a bucket comparing to a total diversity produced by all participants (aka market), the best you can do is to be ALMOST there, and ALMOST is almost NEVER good enough.

[cyberbabble]

Good find - interesting video.
The other videos in the series are interesting.