Rateform And Value Odds

[cavello]

This week with little real horse racing I have spend time producing a linear regression trendline solution to relate Rateform* ratings to probability of results. Having eventually entered the data I produced a linear regression equation solution with an R squared figure of over 0.97(trend line correlates very well). Having produced this and with the assistance of Betdevil profiling analysis tool I can isolate teams who have odds that are higher then their Rateform figures indicated based on the trendline. Initially I will focus on home games. <?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

For those interested in the statistical data the trendlines equations are produced below:

HOME (y) = 0.0269x +42.5 with R² = 0.973

AWAY (y)= -0.0229x + 29.289 with R² = 0.979

I will post up all teams that meet the above-mentioned requirements later.

* Rateform ratings are a spin off of the ELO rating system to assess the ability of chess players and were devised by the Physicist Dr Elo. Later soccer analysts have adapted them in an attempt to rank the ability of soccer teams. In the Ratefrom system more points are obtained by beating stronger teams and less points lost by losing to stronger teams.

[PAULM03]

Re: Rateform And Value Odds What are y and x? I take it one is odds and one is rating? Have you tried fitting any other curves?