robert99
05-20-2009, 05:02 PM
Hi,
A single horse race can be MLR analysed in terms of the winner = 1 and the losers all = 0. For a single race you can get a set of constants for the various racing parameters you think are important for forecasting.
You can analyse many races but the constants will be different for each race and each will only be an "accurate" best fit for the race they came from.
How then do I get to the next stage of finding the best fit for a set of generalised constants to apply with minimum error for odds prediction for a set of future races? The papers I have read seem to gloss over this key stage.
Thanks
A single horse race can be MLR analysed in terms of the winner = 1 and the losers all = 0. For a single race you can get a set of constants for the various racing parameters you think are important for forecasting.
You can analyse many races but the constants will be different for each race and each will only be an "accurate" best fit for the race they came from.
How then do I get to the next stage of finding the best fit for a set of generalised constants to apply with minimum error for odds prediction for a set of future races? The papers I have read seem to gloss over this key stage.
Thanks