Re: Bar Stool Pundits Week 5
[ QUOTE ]
[ QUOTE ]
that might be kelly/2 correct if one uses previous win rate as true value of pick
[/ QUOTE ]
Scary but true I think.
[/ QUOTE ]
I had the unfortunate luck of discovering the Kelly criterion before learning about sample sizes. Some dude on the TheRX (Spreadbeater I believe) was hitting over 60% on his NBA picks over a sample size of (ZOMG!) ~100 games, so I decided to start tailing him. First night I tailed him he posted 5 picks (a lot for him), so I was pretty excited. I threw down 8% of my bankroll on each game (full kelly sizing for a 60% win rate, simultaneous wagers be damned). After a 1-4 night...good-bye 25% of my bankroll. [img]/images/graemlins/grin.gif[/img]
|