Best Polymarket Backtesting Data

The ranking rewards whoever does the specific job best, not whoever has the most features. The full matrix that underpins it covers 10 providers across historical prices, historical L2, resolution, free tiers, and bulk export. Facts below were verified against public pricing and docs pages on September 24, 2026, and recorded in our provider matrix.

• 1. PolyOrderbooks — for replay-grade depth: 250ms full ladders that keep the ask behind a reported mid, with resolved markets queryable and an AI backtest add-on.
• 2. Telonex — for tick-level trades and event-driven books in Parquet when the research workflow is bulk and multi-exchange.
• 3. Predexon — for free unlimited historical trades and order books across venues for cross-market studies.
• 4. PolymarketData — for backtesting at 1-minute L2 with broad market coverage and SQL-friendly access.
• 5. PolyTest — for Up/Down-specific backtests with a documented 8-level snapshot model and timestamp lookup.
• 6. PolyBackTest — for BTC/ETH Up/Down full-depth backtests on a self-reported up to 150-day accumulating window.
• 7. Dune Analytics (Dune API) — for fill-based backtests over on-chain executions when you can live with quote-to-fill divergence.

Measured coverage is the part a marketing page cannot answer. Our archive captures order books every 250ms and serves every plan, including free Starter, at that same 250ms query resolution. Competitors in the matrix cap documented L2 detail at 1-minute (PolymarketData), at 8 levels per snapshot (PolyTest), or rely on self-reported sub-second claims that conflict across their own pages (PolyTest, PolyBackTest). Resolution is not the only axis. Read the matrix row for interval-sampled versus event-driven capture (Telonex, DepthFeed), for on-chain-only limits that exclude the off-chain book entirely (Dune), and for archives that stop at coverage end or endpoint shutdown (polyReplay, Dome). Each listed product is the best answer for a specific job, and none is the best answer for every job.

Free tiers matter when the honest answer is the official API — which it usually is for live prices. Every provider in the matrix lists its free tier; where the free tier is only a sample or a trial, the row says so. If the job is historical depth, expect to pay for it, because depth history is the expensive thing to keep.

A backtest on this venue only means something if the reprice is modeled, since that is where the venue strays from most idealised simulators. The reference-aligned, empty-side event moves the executable price a full band in one second and produces the booked fills your strategy must suffer.

The honest cycle is: pull the 250ms archive, replay your order logic frame by frame, record the realized fills against the recorded floor, and compare against the venue's actual one-sided share. Providers that serve deep archives are the only ones that make that cycle possible at scale.

The outcome metric that matters is the realized spread after replay, not the headline win rate; a strategy that looks profitable on mids usually becomes a breakeven trade on floors.

A backtest ledger should record three numbers per event: the quoted floor at the decision second, the realized fill, and the reference on that second. A provider whose exports make those three numbers joinable is the one worth replaying against.

Require archive access with the real cadence, because a backtest on summarized data backtests the summariser.

Replay at least two repricing hours and one resolution hour before believing any result, since those are the windows where fills actually get booked.

Store the raw frames for every backtest, so any reported result can be reproduced — the archive is both the data and the receipt.

• Full provider matrix
• Order book data
• Price history
• Data download