The problem with backing horses is... they're horses
For the 2014 Cheltenham Festival, a friend recruited me as his partner in a massive "league" in the local pub, based on all the results of four days. Depending on the odds, you got a certain number of points for a win or a place, the scores were recorded on a big board on the wall of the pub and there were cash prizes for the top points scorers.
Though I had retired from betting on horses some years previously, my friend insisted that I make the selection for each race, in tribute to my fine judgment and because he could not be bothered. And for the first three days I failed him horribly. Drawing on all my knowledge and experience, we were so far down the table that on the fourth day there was simply no point in trying to do it right, the arithmetic told us we needed several horses to win at huge odds to have any chance at all.
So I just picked out a bunch of unlikely candidates at big prices, based on nothing but the fact that they were at big prices, and not only did we end up picking the winner of the Gold Cup - Lord Windermere at 20/1 - by the end of the day using this "method" we'd had so many ridiculous winners we shot up the table to finish in second overall.