The study as described pretty clearly defines “best” as “gets the greatest reward as defined by the game”. But this is actually totally beside the point of the

The study as described pretty clearly defines “best” as “gets the greatest reward as defined by the game”. But this is actually totally beside the point of the article, which is that we’re bad at recognizing small differences in probability and over weight recent experience. The headline is misleading, it isn’t really about deliberately making bad choices. When the winning pattern was clear players chose it more often, and when it was less apparent they chose it less often.