polymarket resolution drift
What it checks
In the last 5 days before a polymarket resolves, the YES price often overshoots its all-time fair value. We mean-revert: short when overshot up, long when overshot down.
Mechanism
In the last ~5 trading days before a polymarket resolves, the YES price becomes noisy — microstructure flows, low liquidity, and asymmetric information distort the price. Extreme reads in the terminal window MEAN-REVERT to the lifetime volume-weighted fair value, NOT to the eventual outcome. The equity then follows the prediction-market direction.
Signal rule
Within 5d of resolution_date, if |z(yes − fair_value)/σ_yes| ≥ threshold → take position in -sign(z) direction. Hold 1/3/5d. T+1 lag.
Data dependencies
daily_pricesAdjusted-close OHLCV for every US-listed ticker; primary price feed.
polymarket_marketsWorker data table — see services/worker schema.
polymarket_prices_dailyWorker data table — see services/worker schema.
Expected edge
- Paper alpha
- untested — moonshot
- Paper window
- T-5 to T+5 around resolution
Untested — moonshot. Target 50-150 bps per qualifying terminal-window fire; ~10-30 fires per linked ticker per year.
Example tickers where this is likely to fire
Illustrative only — the signal fires based on the live data, not a fixed list.
Related families
polymarket event premiumEvent-drivenPrediction-market prices on company-specific binary events (FDA approvals, M&A close-by-date, CEO departures, earnings beats, bankruptcy, lawsuit outcomes) update faster and more accurately than equity options price the same outcome. When the polymarket yes_price deviates from the option-implied probability of the same outcome by more than threshold, the equity has not yet absorbed the news.
polymarket iv skew spreadEvent-drivenPrediction-market YES probability vs equity-option-implied probability of the same binary event. When |P_polymarket − P_option_implied| > threshold and the equity has an options chain, take the cheaper-side equity position. Distinct from #206 polymarket_event_premium which compares against realized vol (0.5 baseline); this is the true Moonshot-C cross-instrument arb.
polymarket ma close driftEvent-drivenFor merger/acquisition markets ("Will X merger close by date Y?"), the polymarket YES probability is a clean prior on deal-completion. When the polymarket says >75% close-prob but the equity is trading far below the deal price (wide deal-spread), the classical Mitchell-Pulvino risk-arb edge is amplified: there's a paid wait AND a prediction-market vote of confidence on completion. Symmetric short on <25% close-prob (break risk).
Explore polymarket resolution drift on alphactor.ai
See which tickers this family is currently firing on, with live signals and rankings.