PolyQuantLab —the Polymarketquant lab thataudits itself.
Sub-second L2 depth across Polymarket, Binance and Bybit. Strategy Builder, walk-the-book backtest, paper trading. Every arb signal we surface is tracked against resolution — model EV vs realised PnL, public.
Three tiers of signal. We show you which two pay.
Marketing claims are cheap. These numbers come straight from our audit log — every arb we surface, reconciled the instant its market resolves.
Data scale — measured, not marketed.
Six tools. One audited workbench.
Honest audit
Every signal the engine surfaces is logged on detection and reconciled at resolution. Model EV vs realised PnL — public, unedited, the basis of every claim on this site.
Strategy Builder DSL
Compose entries / TPs / SLs from typed conditions with AND/OR groups, σ-aware operators, drift term, dual-oracle gates. Python codegen tab if you want to extend.
Walk-the-book backtest
Real partial fills against L2 depth, Polymarket 2026 fees, maker queue position, cancel-rate aware. Realised PnL, profit factor, Calmar, equity curve.
Paper trading
Save a strategy. We run it on every new snapshot. Real-time virtual fills, live equity curve, baseline backtest overlay. Validate before deploying capital.
Parameter sweep
2D grid over your strategy's free parameters. Plateau detection highlights robust regions, not overfit spikes. One click to re-run with the best params.
Live mispricing monitor
Live arb signals — logical (math-guaranteed) when they appear, plus the audit-proven endgame setups in the final seconds. Every row tagged with track record.
Compose strategies visually. Or in code.
Drag conditions into AND / OR groups. σ-distance thresholds, drift terms, dual-oracle gates — all typed, all sweepable. The builder generates Python you can copy out and extend.
- Typed condition primitives (τ, σ-distance, token price, spread, drift, etc.)
- Nested AND / OR groups with one-click toggle
- Sweep any leaf parameter as a 2-D grid
- Python codegen — copy, run, extend
1# Strategy compiled by PolyQuantLab Strategy Builder2# Edit visually at https://polyquantlab.com/dashboard/strategy-builder34import asyncio, datetime, math, os5import httpx67# ─── Strategy parameters ─────────────────────────────────8POSITION_SIZE = 7.59TICKER = "BTC"10MARKET_TYPE = "5m"11MARKET_LIMIT = 5012FILL_MODE = "walk_book"13MAX_FILL_PRICE = 0.98514TAKER_FEE_RATE = 0.0721516API_BASE = "https://api.polyquantlab.com"17API_KEY = os.environ["POLYQUANTLAB_API_KEY"]1819# ─── Condition closures (translated from Strategy Builder) ──20def check_entry(snap, market_open, history, resolution_at, prev_state):21 return (22 (time_to_resolution_s(snap, resolution_at) <= 120)23 and (snap["mid_no"] <= 0.985)24 and (coin_move_since_open_pct(snap, market_open) <= -0.15)25 )2627# ─── Main loop ───────────────────────────────────────────28async def fetch_snapshots(client, market_id):29 r = await client.get(30 f"{API_BASE}/v1/snapshots",31 params={"market_id": market_id, "limit": 100000},32 headers={"Authorization": f"Bearer {API_KEY}"},33 )34 r.raise_for_status()35 return r.json().get("snapshots", [])Where each tool stops.
| PolyBackTest | PolyQuantLab | |
|---|---|---|
| Sub-second L2 data | ||
| Walk-the-book backtest | ||
| Strategy Builder DSL | basic | |
| Paper trading (live) | ||
| Public audit of own signals |
Audit-first. Then backtest. Then paper. Then deploy.
Open the live track record. See which tiers historically pay and which don't. Don't build against signals proven to fail.
Strategy Builder DSL — typed conditions, AND/OR groups, σ-aware operators. No code required.
Walk-the-book backtest + parameter sweep. Find the plateau, not the spike.
Run live for 1-4 weeks. Real-time virtual fills, baseline overlay. Only then deploy.
Things you're probably wondering.
You claim you audit yourselves — how do I know that's real?▾
How is this different from PolyBackTest?▾
Do I need to write Python?▾
Is this affiliated with Polymarket?▾
Can my bot consume the API?▾
What if I just want to try it before paying?▾
Free tier. Pro from $19.90/month.
Same data feed across every tier. Higher tiers unlock more throughput and concurrent backtests. Cancel anytime.