Ratings,Trend lines,Odds and outliers

[cavello]

I subscribe to Betdevil that produce ratings for soccer matches-Rateform,Recent,League and Combo. I have pondered how to use these in a meaningful way to calculate match odds. Initially I produced a linear regression line analysis based on Rateform. But then I realised after awhile that Rateform figures grow as the season progresses and therefore render my input data invalid. Then how could I use the figures to isolate potential value bets? Then the penny dropped. Why not produce a weekly analysis based on the odds and rating values. If I take leagues which have completed approximately the same number of games and have similiar profiles in turns of Home/Away % I could produce a trend line and formula. This formula could then be used to calculate the value of individual matches. This week the following matches were indicated to be value Port Vale lost, Milton Keyes Won 1.9 , Kidderminster(draw),Tamwoth Won 3.25 .Cowdenbeath(won) 1.7 ,Forfar(lost),Stirling Albion(draw),Songdal(draw) and Tromdalen 2.04 won. A small lost incurred but 2 games involved last 2 minute reversals!! Obviously these prices that are bigger then then they should be based on ratings may have valid reasons for their apparent value(Derbies, injuries etc). But because this offers a quicker way of isolating potential value I certainly will be using this technique. The indicated value bets could be backed up by using RFO or The Times computer odds plus team knowledge or other statistical methods. I will produce the first set of outliers for next weekend. I expect this method to be more efficient after about 10 games.

[aliando]

Re: Ratings,Trend lines,Odds and outliers i think the level of maths is a bit beyond me, but can grasp what your trying to do - :hope for you mate!