Except the issue is how certain we are of these percentages. In poker, the win rate of AA against 72 heads up is well known. In reality, we don't have that. We

Except the issue is how certain we are of these percentages. In poker, the win rate of AA against 72 heads up is well known. In reality, we don't have that. We get whatever ideas have the largest marketing budget behind them, and we just have to trust them. It's more like "nobody gets fired for buying IBM (or pick your favorite pundit)" than running a poker scenario through a solver. And that's fine, I don't think you can expect much more than that out of people.