I don't think that statistics are useless. It is how you use them which determines whether they will be of assistance. One statistic on it's own is useless, you have to give varied weight to each combination.

As becareful says, rating chances according to combinations of information ensures that you bet when there is value about a specific runner.

For example, did anyone hear the rap the racecaller gave Citichy today as it was being loaded into the stalls - he gave the various races he had won. At $78/1 surely he was over the odds, but even more so 15/1 for the place!