There are so many more inputs and outcomes to model in sports than in politics, I'm not surprised that it's a more interesting subject matter for a statistician

There are so many more inputs and outcomes to model in sports than in politics, I'm not surprised that it's a more interesting subject matter for a statistician. He never seemed too much into the wonkier side of politics. The science of predicting election outcomes may continue to be improved, but perhaps it's essentially solved in a way that sports prediction isn't.