i sympathise with this debate. for many years i tried times and developed with a friend a complicated pascall based program to calculate the times of meetings and to work the cyberhorse meetings automatically.
many top selections win. but the difficulty was there is a top selection in every race and which ones to back and not to back is always the perennially question. its a problem that will alwasy confront us devising a method that does not take too much work, too much betting and is reasonably reliable.
i have done studies also on various ideas as said in another topic on sunday i backed donarch in the coleraine cup even though it got beat by 31 lengths last start.
i also was talking to the local chemist today and he said he was at caulfield on sunday and had $20 on it because it was paying $18 as against 8/1 on the books.
simply on the overs he cracked it. my system is automatic (programmed (perl language ) cyberhorse form based and selected two horses only from coleraine.
prince iluka and donarch which ran 1st and second. i developed this method after years of trying times and weights.
exasperating is the description i use and to give an example of the silliness of my fall backs in great glens race on sunday i looked only at the top 4 horses and backed the two longest prices. i got the winner.
today i had four selections in the last at canterbury, time deposit, britts best, porquoir and forbill.
i backed the two longest prices ones. they finished 2nd and 3rd the even money favorite and britts best 6/1 never featured.
maybe there's no formula. except to determine a safe factor of horses that brings regular good priced winners.