"outliers" in stats

[davidcampbell]

hey guys need some stat brains here please i have the kernel of an idea growing for a system (dont want to disclose much just now) and im looking to backtest it over the past few seasons uses average odds that i can access for various matches now lets say i have for example a mtch where out of the 5 bookies odds i have available prices are 2.65, 2.5, 2.4, 2.4 and 2.35 this gives an average odds of 2.45 but obviously the first one - 2.65 is pulling that up somewhat as without it the average is 2.41. question is how far does one term have to "outly" (dont know if thats the right terminology - sounds vaguely familiar) before you dismiss it? is there a "given" answer or does it not matter too much? if this doesnt make sense ill try to explain a different way thanks

[Bruise Pristene]

Re: "outliers" in stats work out standard deviation of data. anything outside 2 standard deviations of mean can be dismissed as an outlier so if the s.d of your data ( i havent worked it out) was 0.05 say and the mean is 2.45, then anything above 2.55 and less than 2.35 is an outlier so neligable