I personally find that trying to optimize too much leaves me stuck in crummy local maxima. By the time you're attempting to use mathematics and modeling to determine the maximum return for happiness, you've lost sight of the bigger picture.