15 versions
Uses one age-, position- and team-aware prior for unknown players and releases performance evidence continuously by cumulative senior minutes across rebuilds, expansions and performance updates.
Refreshes profile nationality and physical measurements, applies physical attribute deltas only to changed measurements, introduces goalkeeper Discipline, applies age-aware cold/stale context and reconciles sparse single-season history with current-team context.
Adds a versioned primary position from the current Sportmonks player profile and evidence-backed secondary positions from registered lineup starts. Ratings and broad model roles are unchanged.
Updates current-season performance and current squad membership from one complete frozen provider acquisition batch. Transfers, arrivals and departures are included; historically known arrivals reuse their frozen evidence and community impact remains disabled.
Adds Chance Liga, Greek Super League, NB I, Romanian Superliga and Niké Liga to the published population. All five use five historical seasons, a refreshed 2026-27 roster and the approved 27.9 Penalties and Discipline contracts. Community impact is disabled.
Retains the reviewed 27.8 cold-start, youth-evidence and stale-return cohort, restores all unrelated rating values to the approved 27.7 reference, and restores the approved Spearman Penalties inference. No new football or community evidence is used.
Retains the evidence-responsive under-20 release curve and replaces the stale-return heuristic with a rolling-origin backtested, age- and evidence-aware team-context model. Community impact is disabled.
Adds Austrian Bundesliga, HNL, Serie B, Ekstraklasa, Russian Premier League and 1. Lig. Liga MX and Saudi Pro League remain excluded pending independent cross-league calibration. Community impact is disabled.
Removes Saudi Pro League pending independent cross-league calibration and adds Ligue 2, La Liga 2, League One, League Two, Danish Superliga, Swiss Super League and Super Lig to the prior twelve-league candidate population. Existing and new players are rebuilt together from five completed seasons plus the frozen 2026-27 current-season input. Community impact is disabled.
Penalty taking is recalculated over five seasons with a volume-aware conversion model and Spearman-stable technical inference for limited or absent attempts. Discipline is recalculated retrospectively across all five historical seasons and the frozen current-season snapshot using the revised foul, card-severity and dismissal hierarchy.
Restores the required five-season penalty evidence that was unavailable during the 27.3 baseline build, correcting the compressed penalty-taking distribution. This release also introduces cumulative regulation penalty updates: scored and failed penalties update PENALTIES, while penalties faced and provider-recorded saves update PENALTY_SAVING. Evidence is strongly shrunk for small samples, penalty shootouts are excluded, and all other evidence uses the same 27.3 cutoff. Community evidence is not active.
Adds the Bundesliga and Liga Portugal as the seventh and eighth tracked leagues. Historical domestic-league evidence from 2021-22 through 2025-26 was rebuilt across all eight leagues and current squads were frozen on 20 August 2026. Cumulative 2026-27 competitive evidence is applied only where a domestic season has begun: Liga Portugal is active, while the Bundesliga starts on 28 August and receives no active-season freshness or performance adjustment in this release. Community evidence is not active.
Adds Serie A as the sixth tracked league. Historical domestic-league evidence from 2021-22 through 2025-26 was rebuilt across all six leagues, current squads were frozen on 20 August 2026, and cumulative 2026-27 evidence was applied where a domestic season had begun. Serie A starts on 22 August, so it receives no active-season freshness or performance adjustment in this release. Community evidence is not active.
Review candidate. Overall and attributes use frozen cumulative domestic-league player-season totals with full-payload change detection. Freshness uses canonical league-season start/end windows, so pre-season time contributes zero; role fits are recalculated from updated attributes.
Initial five-league ratings release using Overall V1.3, Attributes V1.1 and Role Fit V1.2.2.