Back testing, how many games to be statistically relevant?
Does any one have opinions or links to how many games do you need to say that your theory is statistically relevant?
I back tested a theory and it went 123-87 from 08-11. So 210 games at 52.5 games per year. Is that enough to have any confidence in it this year? It hit 58.6%, best year was last year 30-18, 62.5% worst year was 7-6, 53.8%