A new math rule makes some statistical models provably reliable

When researchers build statistical models, they often use a grid of points to estimate smooth curves from data. This paper provides a new mathematical rule for how dense that grid must be to make the model's predictions truly reliable.

Building a statistical model often means guessing how many data points to use in your grid. This paper replaces those guesses with a hard mathematical rule. If you want your model's predictions to be provably correct, you now have a specific standard to meet.