Skip to content

Variance

villekuosmanen edited this page Feb 2, 2019 · 1 revision

Variance describes how much the qualifying results of a particular entity (driver, engine...) vary. It doesn't affect the entity's Power Ranking and is only used in the Monte Carlo simulation. In the simulation, higher variances lead to a bigger spread on the entrant's predicted qualifying positions, but does not usually affect their average qualifying position.

Average variance is calculated using an Exponential Moving Average (EMA) model, where results from the most recent to the oldest have an exponentially decreasing effect on the average, but their effect never reaches zero.

Variance differs from Deviation, which describes the spread of all qualifying times in a qualifying, and is used at normalising them to a single, uniform scale.

Clone this wiki locally