Methods
Bayesian Analysis
A statistical approach that describes what we know about a quantity as a probability distribution and updates it as data arrives.
How we use it
Our team ratings are Bayesian: each team's offense and defense rating starts close to league average and moves away only as far as its games justify. Early in the season this keeps one blowout from producing an extreme rating, and it gives every rating an uncertainty range, which we show on the stats pages.
The simulations draw ratings from those distributions, so the uncertainty carries through to win probabilities.
Where to see it
Related terms
- Survival Analysis
- A family of statistical methods for how long something lasts before an event happens. We apply it to drives: how many yards a drive survives before it ends.
- Drive Efficiency
- A team rating for how far drives get before they stall in a punt or field-goal attempt, adjusted for the opponents faced and the game situation. Zero is league average and higher is better, for offense and for defense.
- Monte Carlo Simulation
- Estimating the chance of an outcome by playing it out many times with random draws and counting how often it happens.