If a system is doing really well over 50-100 plays, is it likely to fall apart?

[eddie0vedder]

hey peeps, i am pretty new to all this but i have been tracking back the last month of nba looking at various ways to bet and i have come up with 2 systems that look pretty good. In the last 4 weeks one is 51-21 71% and the other 70-39 64% (this one i tracked back a month of NCAAB and it was 59% over 300 plays). I am strongly considering adding to my bankroll and starting to use these 2 systems but i am afraid they will start going the wrong way on me. I cant go any further backwards as the teams need to have at least 10 games played.

Is 1 month of data enough or should i track another month and then if its still going good start playing?

[Whippit]

think you answered your own question:

"I am strongly considering adding to my bankroll and starting to use these 2 systems but i am afraid they will start going the wrong way on me."

backtest last couple seasons & let me know if it works.

Ill be President and make you Vice President in charge of production.

[Jmayor]

Honestly you need at least 5 yrs of back testing to be legit. Let us know man

[hoya]

Definitely needs several seasons of testing. Nba can be weird with early season trends.

[hockeynut]

Yeah I'd say you need to test it for several seasons. Some people hit 60% one season and below 50% the next. You gotta keep it going for a while.

[Whippit]

gotta be predictive dude

PRE EFFIN DIC TIVE

[Whippit]

^^^ that's unpossible btw...just ask any weathergirl


in all seriousness brah you can datamine yourself a million ways into the grave

wouldn't waste me time backtesting cuz what you got is probably rubbish & total waste of your time