How each sport earns trust

Current status

The earlier CFB ridge projection remains retired after failing validation. The owner has separately authorized SR-CFB-CONTEXT-RESIDUAL-0.2 to release capped 0.25-unit provisional spread plays when both game-specific quarterbacks, current availability, a fresh exact paired price, and its frozen edge rule pass. Its retrospective margin error still trails the market and its prospective promotion gates have not passed. The older efficiency challenger continues as private research. Through Sunday, Oct. 25, the owner has also authorized capped 0.25-unit provisional NHL fitted-moneyline plays when both goalies are confirmed, a fresh exact paired quote and frozen research thresholds pass, and the model-market difference is not anomalously large. NHL goalie quality and lineup effects are not fitted, and prospective market validation has not passed. All provisional results are labeled and reported separately from model-qualified Official Plays. Every historical play retains its locked projection and release version.

1. Sport-specific research inputs

The current MLB projection uses official team scoring and runs allowed, probable-starter game logs and workload, confirmed-lineup OPS when available, bullpen workload for volatility, and fresh paired prices. Venue, roof, and weather are captured as release evidence, but fitted park effects, xERA, handedness and pitch mix, catcher, umpire, defense, and baserunning adjustments are not yet implemented in the run mean. Football requires its own contract: quarterback and roster availability, opponent-adjusted efficiency, explosiveness, finishing drives, trenches, coverage, special teams, pace, coaching, travel, rest, weather, and the full two-sided market. Missing critical inputs lower confidence rather than being silently guessed.

2. Probability before opinion

The model produces a fair win probability. The market price is converted to implied probability and adjusted for sportsbook margin. A recommendation exists only when the difference is large enough to survive uncertainty and the available odds remain within the published limit.

3. Validation and provisional experiments

Every sport is judged with time-aware tests that prevent future information from leaking backward. We monitor calibration, out-of-sample return, closing-line value, performance by edge bucket, and game context. Results from one sport cannot validate another. The owner-approved CFB and NHL trials are small, explicitly provisional, and never described as validated model-qualified releases; a backtest does not turn either into one.

4. Version every material change

Feature additions, weighting changes, data-source changes, and threshold changes receive a new model version. Each published recommendation stores the version that produced it so results from different approaches are never blended invisibly.

5. Publish, lock, and grade

A live recommendation stores its timestamp, selection, available price, maximum playable price, unit size, probability, estimated edge, rationale, and model version. After publication, the record is locked. The result and closing price are added later; the original research is not edited away.

6. Measure closing-line value separately

Every graded release is evaluated against a directly captured closing market when the exact paired evidence is available. CLV is reported separately from win/loss results and never changes the original grade. CFB shadows stay in the research ledger; any owner-approved provisional play enters the released-play ledger under its own label and is excluded from model-qualified promotion statistics.

Read the current CLV report card →

6. Responsible use

Model outputs are estimates, not guarantees. Sports Rounders uses small fixed-unit sizing, discourages chasing, and may publish no play when the evidence or price is weak. Members remain responsible for their own decisions and local laws.