Skip to main content
PolyQuantLab continuously scans the Polymarket book for mispricings. Every signal we surface is logged to arb_audit_log (Postgres) at detection time, then reconciled to resolution outcome the instant the market closes. The aggregate is queryable via GET /v1/arb/audit/aggregate or browsable at the public audit page.

Four tiers

TierTriggerDirectionWhat it claims
logicalyes_ask + no_ask < $1.00BUY_BOTHMath-guaranteed profit at resolution.
endgamemodel says near-certain + dual-oracle agreement, in final 0–120 sBUY_YES or BUY_NOHigh empirical hit-rate from book-lag in the last seconds.
stableprobability-arb on the maker-zone book (fill ≥ $0.30)BUY_YES or BUY_NOWhere the maker bot is parked — usually well-calibrated.
staleprobability-arb on a thin / deep book (fill < $0.30)BUY_YES or BUY_NODeep mispricings, often already arbed by HFT.
The /v1/arb/live endpoint returns currently- firing opportunities across all four tiers in real time.

Audit record per detection

When the detector emits an opportunity, the following is logged before any user sees it:
ColumnTypeNotes
detected_attimestamptzWall-clock at detection.
market_idtextThe CLOB market hash.
ticker / event_typetextBTC / 5m, etc.
tiertextlogical, endgame, stable, stale.
directiontextBUY_YES, BUY_NO, BUY_BOTH.
fill_pricenumericReal ask we’d hit.
model_yes_probnumericOur log-normal model’s P(YES).
model_ev_per_sharenumericNet EV after entry fees.
est_fee_per_sharenumericPolymarket 2026 taker fee.
resolution_attimestamptzWhen the market will resolve.
On resolution the settler joins to events.resolution_outcome and fills in:
ColumnTypeNotes
resolved_attimestamptzReconciliation time.
resolved_outcometextyes_won / no_won.
realized_pnl_per_sharenumericpayoff − fill_price − est_fee, computed correctly for each direction (BUY_YES, BUY_NO, BUY_BOTH).
Nothing is deleted. The aggregate you read from /v1/arb/audit/aggregate is unfiltered count(*), avg(realized_pnl_per_share), etc.

Calibration

The aggregate returns both avg_model_ev and avg_realized per tier so you can compare. A model whose avg_realized ≈ avg_model_ev is well-calibrated; the gap is the engine’s honest blind-spot. For BTC/ETH/SOL Up/Down markets across our entire log, the probability arb (stable + stale) tiers have realised ~zero edge — Polymarket’s maker bots are well-calibrated for crypto. The logical tier is positive by construction, and the endgame tier has shown ~95 % empirical win rate from book lag in the final seconds. Read more on the public audit page — same numbers, no signin required.