Sidebar

[mickey_arnold]

Steve Matthews for Newsday was already cited...The NY Post had a selector who specialized in longshots...Can't recall who.. Maybe Matthews worked for the Post at one time..

Was always puzzled with the number of points assigned to each of the choices, i.e 6 for 1st, 3 for 2nd, 1 for 3rd or 6,2,1 for each of the ranks.

If they were trying to mimic the % of winners in over the long run by finish position or ant other factor(s), there is a world of difference between a 6,3,1 scale versus a 6,2,1 scale.

I think Sports Eye used to go 4 positions deep for each selector.

I think I used a 10 handicapper consensus to widen the possible "contenders". This is where the scalar differences could be problematic so I would stick to newspapers or publications with the same point scale. get consensus total for each horse , adjust for scratches, get a simple average for each horse, go 1 or two decimals deep in the scores, square the scores, derive each horses share of the grand total of points to use as a probability of winning, cutoff point for final contenders and look for overlays among them.

If found that over about the long run, between using 6 or 10 handicappers there was little difference in the final probabilities of winning for each horse or the profitability.

Could using a little more sophisticated analysis of race type, scoring system and final selection of contenders have helped, who knows?

Over the long run, my method was better than picking random, produced a better net return compared to the average of all handicappers or any individual handicappers. But of course still less than profitable and too much manual work for too little return... With computerization and today's rebates, still anybody's guess.