Ahhh, the epsilon greedy strategy... I forgot I had read this until I was a few paragraphs in. If you have rare positive results, it gives you better statistics

Ahhh, the epsilon greedy strategy... I forgot I had read this until I was a few paragraphs in. If you have rare positive results, it gives you better statistics on your highest signal tests, while still evaluating the others (i.e. more than one alternative).