Per-sport CLV review
A descriptive view of terminal in-play CLV by sport. Runtime live/paused state is deliberately omitted because this analytics view is not its source of truth. These observations do not authorize scaling, restriction, reactivation, or a claim of model skill.
Updated 2026-10-31 · runtime state intentionally not inferred here
NCAAMB
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
NCAAWB
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
WNBA
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
LOL
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
NFL
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
SOCCER
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
WTA
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
NHL
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
MLB
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
CS2
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
ATP
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
CFB
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
NBA
Do not scale, pause, or reactivate from terminal in-play CLV alone. Review fixed-horizon markouts, exchange fills and slippage, calibration, sample uncertainty clustered by game, and the sport's independent risk gates.
Why we publish this. The current aggregate can be unfavorable, and showing it creates an audit trail. It is one descriptive execution measurement, not a verdict on a model or sport.
Investigation tags. Negative-CLV rows receive deterministic heuristic tags such as bad_model, bad_blend, bad_timing, or bad_clv_label. These names are hypotheses for follow-up, not proven causes; the counts cannot identify causality by themselves.
Operational decisions. Scaling, pausing, and reactivation require separate evidence: fixed-horizon markouts, execution quality, calibrated out-of-sample forecasts, clustered uncertainty, and risk controls. This page does not drive those decisions.
Data & methodology. CLV is computed weekly via backfill_clv.py from Polymarket price history. Failure-mode tags are deterministic from clv_failure_tagger.py. Both are reproducible from the public /clv.json + /clv/failures.json endpoints.