"If a theory has a lot of parameters, you adjust their values to fit a lot of data, and your theory is not really predicting those things, just accommodating them. Scientists use words like “curve fitting” and “fudge factors” to describe that sort of activity. On the other hand, if a theory has just a few parameters but applies to a lot of data, it has real power. You can use a small subset of the measurements to fix the parameters; then all other measurements are uniquely predicted. " Frank Wilczek