# WEPA and adjusted metrics

Weighted Expected Points Added (WEPA) is CFBD's family of opponent-adjusted
efficiency metrics. It is intended to answer a practical question: how strong
was a team's or player's performance after accounting for the quality of the
opposition?

WEPA is a season-level adjustment, not a play-level value and not a win
probability. Use [PPA](./ppa) when the unit of analysis is an individual play.

## How the adjustment works

The model evaluates offense and defense together across the FBS schedule. It
uses regularization to separate the performance of the unit being measured
from opponent strength and home-field context while limiting extreme estimates
from small or unbalanced samples.

That process is applied to several metric families, including PPA, success
rate, rushing-line measures, and explosiveness. Player passing and rushing
outputs use the same general opponent-adjustment approach for players with a
sufficient qualifying sample.

WEPA is proprietary. Exact fitted coefficients, regularization settings,
sample thresholds, and training artifacts are not published.

## Reading team values

The team response includes adjusted offense and defense-allowed views:

- `epa` and `epaAllowed` include total, passing, and rushing efficiency.
- `successRate` and `successRateAllowed` include overall, standard-down, and
  passing-down splits.
- `rushing` and `rushingAllowed` include line, second-level, open-field, and
  highlight yards.
- `explosiveness` and `explosivenessAllowed` measure value on successful plays.

Higher offensive values generally indicate better performance. Lower allowed
values indicate that the defense conceded less adjusted production. Compare
like fields within the same season rather than treating different metric
families as one common scale.

## Player WEPA

The passing and rushing endpoints return a player's adjusted `wepa` value and
the number of qualifying plays behind it. The number is intended for
season-level comparison among players in the same role and metric family.

Opponent adjustment reduces one major source of schedule bias, but it does not
isolate a player from scheme, teammates, game state, or every selection effect.
Sample size still matters.

## Kicker PAAR

Points Added Above Replacement (PAAR) evaluates field-goal results against the
expected value of the attempts. It is published with the WEPA endpoints but is
a related, distinct metric rather than an opponent-adjusted rushing or passing
measure.

Positive PAAR indicates that a kicker produced more value than the comparison
baseline across the included attempts. `attempts` provides the sample behind
the rating.

## API access

- [Adjusted team season metrics](/api/adjustedMetrics#getadjustedteamseasonstats)
- [Adjusted player passing](/api/adjustedMetrics#getadjustedplayerpassingstats)
- [Adjusted player rushing](/api/adjustedMetrics#getadjustedplayerrushingstats)
- [Kicker PAAR](/api/adjustedMetrics#getkickerpaar)

See [Metrics and definitions](./metrics-and-definitions) for the component
metrics used in the response.