Performance rating reference

This reference explains Player performance ratings. Roster Elo comes from match results and is covered separately in Ratings and Elo.

Scope and exposure

Ratings are scoped to a League, game, and supported category/mode. They are not a universal cross-League player ranking.

The current calculation uses recent matches (a target window of 30) and can extend further back to reach the configured eligibility floor. Eligible SKM and LTS rounds contribute one exposure unit each. HPT uses observed active-player seconds divided by a 90-second nominal round. Other objective modes use their recorded game duration, hold segments, or the conservative fallback when no duration is available. At the default floor of 50 exposure units, that is 50 eligible one-life rounds or 75 observed HPT minutes; mixed scopes combine their appropriate exposure.

Connected Server time is not a substitute for this participation exposure. AFK connection time and monitoring gaps do not become rated match participation.

Valid observations

Only qualifying, non-voided observations contribute. For SKM and LTS, eligible completed rounds supply both the stats and the participation denominator; the game total is not added again. In verified Mordhau SKM captures, a first partial round after a mid-map attach or handoff is not treated as a fully observed round. Chivalry's committed adapter closes its first observed round from the event feed, so do not assume that partial-round exclusion applies to every Chivalry capture.

Current Roster changes stop future participation where appropriate, while valid historical rows remain. Current team Elo does not rewrite past player kills, damage, participation, or Impact in the performance calculation.

Provisional ratings and references

When a comparison reference is built, Heir selects up to 30 approved, completed matches from the current rating era in this League:

1. Take the most recent League matches first.
2. Fill remaining places with the most recent scrims.
3. Fill any places still remaining with the most recent pugs.

For example, 20 League matches and enough scrims produce a sample of 20 League matches plus 10 scrims. This comparison sample is separate from each Player’s own recent-match performance window and lifetime history.

A scope needs enough eligible Players and observations before its reference can be established. Existing Leagues adopt this reference policy on their next automatic update; their previous published ratings remain visible until a valid replacement is ready. The reference then remains stable until a supported reference transition; it is not continually recentered on whoever is currently playing.

The displayed overall value combines reliability-weighted skill scores. Eligible Players can still be marked provisional below the larger of 160 exposure units and the configured eligibility floor, or while their scope lacks a mature reference. This is a display-confidence policy, not a statistical confidence interval or a promised future rating.

Enhanced-stat coverage

Coverage is explicitly unknown, partial, or complete. A parry event does not prove complete damage observation. Numeric damage fields alone do not establish trusted, fully measured exposure.

The current Mordhau mod protocol does not provide a completeness heartbeat. Its enhanced observations are therefore partial: raw enhanced values can be shown where available, but precision/teamplay and damage-blended rating inputs require verifiable measurement coverage. Do not describe every visible enhanced number as a fully trusted rating input.

Older HPT data without observed per-player active time can remain raw history without contributing to the current rating formula. Historical missing coverage is not backfilled from a guessed full-map clock.

Informational Precision and Defense

The attribute panel may show separate 0–100 informational comparisons from observed game counters. Precision uses headshot rate and normalized damage per hit; Defense uses chambers per observed exposure unit. They use robust median/MAD normalization and sample-size shrinkage against the eligible Players in the frozen comparison sample. The displayed disclaimer is Not used in overall rating: these additional grades are attached after the composite is calculated, never added to its weights, ranking, or radar. Partial coverage can affect their interpretation. A valid fully measured Precision score already used by the core formula keeps its original behavior; the informational fallback does not replace it.

These fields update through ordinary automatic rating processing. The manual /admin rating-rebase command is not available.

Which Players are updated

Routine match changes refresh the affected Players. Unchanged Players keep their existing published ratings. A changed reference, formula, or eligibility rule can affect more Players and require broader recalculation. A Player who becomes ineligible must not retain a stale rating.

Interpreting changes

A score can change after approved data, eligibility, a reference transition, or a formula update. Check the scope, window, approval, and confidence label before comparing it with an earlier number. A rating's number is not the same as its percentile rank.

Next: Mordhau capabilities, Chivalry capabilities, or missing-stat troubleshooting.