Toggle between a simple and complex model, and between poor and good features, then see which change lifts accuracy more.
Toggle the model and the features independently. See which one moves accuracy more.
A simple model with poor features barely beats a guess.
