Pattern 1 — UMA Exit-Liquidity Harvester

Live BLR bot state · 2-minute refresh ·
service
kill-switch
auto-exec
uptime
last placement

Paper PnL — what auto-exec WOULD have made (since pivot 2026-04-26)

Trades captured
Realized paper PnL
Open / disputed
Win rate
By classification
class
n
win%
pnl
By category
category
n
win%
pnl
⚠ Realistic edge — size-aware fee model
Polymarket has zero trading fees; cost is gas (~$0.05 round-trip on Polygon, fixed) + ~0.18% slippage. Cost % shrinks as size grows. Below shows what each closed trade WOULD have made at different position sizes, applied to the same gross % return the market actually delivered.
High-spread cluster (entry ≤ 0.97) — REAL EDGE
closed trades · avg gross
sizefee%net PnL$/tradenet%
Snap-spread cluster (entry > 0.97) — BLOCKED by guardrail
closed trades · avg gross · negative-EV at every size
sizefee%net PnL$/tradenet%
Disputes seen in paper: 0. Strategy thesis depends on dispute rate <5% — sample too small to validate yet.
All paper trades (newest first)

Live positions baseline (real BLR ledger, 30d closed)

Settlements
Paper PnL
Win rate
Dispute rate
By classification
class
n
win%
pnl
By category
category
n
win%
pnl

Filter funnel — last 24h ( total decisions)

Active candidates (passing all filters in latest scan)

Alerts (24h) — markets that surfaced past the pipeline

Open positions

Recent settlements (30d) — closed positions, paper PnL