I'm not a mathematician and I'm also rather new to handicapping. I know the basics of excel and I've been testing out various formulas using a database of 500 CBB games, focusing exclusively on totals.
A couple questions which will probably be very easy for those of you with background in statistics:
1. What percentage would I need to get right from a sample of 500 to have confidence that my results are statistically significant? Obviously I want to be hitting about 53% or better longterm. But if I just get 53% of 500, it could easily be accidental. If I get 60% (300/500), am I justified in having a high degree of confidence that it's not random? 55%? Because if it is random, I'll be losing a lot of money long-term if I keep following it. So what is the best way to calculate statistical significance from a given number of games?
2. The vegas cappers are obviously light years ahead of me with their final numbers. I try to use the information of their line as a starting point for my own excel tests, comparing that number to various other numbers and then seeing whether the result was over or under.
What hope is there, realistically, that their lines could be soft? I know that there's a bit more value in unders long-term, so that's a starting point, but that by itself won't even take you to 51%. If their lines are sometimes a bit off, how much of an edge would I need? If I could get just a half-point advantage long-term, would that be enough?
Thanks for any responses or Tips you can provide.
A couple questions which will probably be very easy for those of you with background in statistics:
1. What percentage would I need to get right from a sample of 500 to have confidence that my results are statistically significant? Obviously I want to be hitting about 53% or better longterm. But if I just get 53% of 500, it could easily be accidental. If I get 60% (300/500), am I justified in having a high degree of confidence that it's not random? 55%? Because if it is random, I'll be losing a lot of money long-term if I keep following it. So what is the best way to calculate statistical significance from a given number of games?
2. The vegas cappers are obviously light years ahead of me with their final numbers. I try to use the information of their line as a starting point for my own excel tests, comparing that number to various other numbers and then seeing whether the result was over or under.
What hope is there, realistically, that their lines could be soft? I know that there's a bit more value in unders long-term, so that's a starting point, but that by itself won't even take you to 51%. If their lines are sometimes a bit off, how much of an edge would I need? If I could get just a half-point advantage long-term, would that be enough?
Thanks for any responses or Tips you can provide.