Hi, KV.

Someone a few years ago, it my have been on this forum, had/has a large data base and said using the default settings that the top selection had the same strike rate as faves. I've never forgotten this, and if the results still hold true that would mean that the loss using default settings would not be as great as backing every fave. Based on this, I'm assuming that there must be some merit with the neurals, maybe not much but I'm hoping.

As previously stated, I've looked at uncountable methods but have yet to discover a profitable method. But I'll never know if I don't try. And there are so many combinations, and even more when combined with non-neural factors that it might just be possible. And despite the disappointments it is a challenge and can be enjoyable if expectations are not too high.

Was it Erasmus? Pity it was'nt a suicide bomber.