Polytoolhub logoPolytoolhubSearch tools
← Backtesting

SDK

prediction-market-backtesting

Build Polymarket strategy replays on NautilusTrader with historical vendor data, joint portfolios, parameter experiments and detailed charts.

3.5/ 5Editorial rating0 user reviews
View repositorygithub.com
Available on
Python, Rust, Jupyter
Pricing
The repository is publicly available with component-specific license notes

Screenshots and interface

Product preview1 / 2

Framework overview and release capabilities.

What is prediction-market-backtesting?

This framework adds prediction-market adapters and research runners around NautilusTrader. It supports historical book replay, cached data loading, notebook workflows and multi-market reporting. The public runnable scope centers on Polymarket; other venue work is at different stages.

Who it is for

  • Python researchers developing reusable replay experiments
  • Teams comparing parameters across multiple markets
  • Users who need inspectable execution assumptions and rich reports

Features and coverage

Available on
Python, Rust, Jupyter
Markets & venues
Polymarket
  • Nautilus integration

    Run strategies through the underlying event-driven simulation framework.

  • Book replay

    Combine supported order-book deltas and trade ticks.

  • Vendor loaders

    Use the documented PMXT and Telonex paths.

  • Cached staging

    Reuse local and materialized data instead of repeatedly acquiring it.

  • Joint portfolios

    Replay multiple markets under shared portfolio assumptions.

  • Optimization runners

    Explore parameter configurations with supported samplers.

  • Reports and charts

    Inspect equity, drawdown, fills, allocation, returns and scoring outputs.

  • BTC sandbox plumbing

    Experiment with the separate live-data sandbox runner.

Workflows with prediction-market-backtesting

Create a reproducible experiment

Choose the vendor and time window, define the runner and strategy parameters and preserve the execution assumptions with the result.

Compare parameters

Warm the data cache, run the supported sampler and inspect outlying fills and drawdowns rather than selecting solely by final return.

Test a portfolio idea

Use the joint runner so capital allocation across markets is modeled together.

Pricing and total cost

The repository is publicly available with component-specific license notes. Data vendors, compute and any separately enabled live operations have their own costs.

Getting started

Before you start: Kalshi components are research and fee-model plumbing, not a public runnable backtest path.

  1. Follow the current setup and runner-contract documentation.
  2. Choose a supported Polymarket data source.
  3. Acquire or point to a small local data window.
  4. Run the example experiment and inspect its report.
  5. Specify fees, latency and passive-order assumptions.
  6. Expand to a joint portfolio or optimization run after the single-market replay works.

Useful links

    What to check

    • Kalshi components are research and fee-model plumbing, not a public runnable backtest path.
    • Limitless and Opinion are planned rather than current replay integrations.
    • Vendor schemas and local file layouts need to match the selected loader.
    • Passive queue position and latency are model assumptions.
    • Sandbox and live runners are separate from historical replay and can have different account requirements.

    A closer look

    Ratings & reviews

    Write a review
    Editorial rating
    3.5/ 5
    0 user reviews
    Polytoolhub editorial

    Our take on prediction-market-backtesting

    Vendor loaders and explicit simulation assumptions provide a substantial research framework. Data dependencies, component licenses and unfinished venue paths make this a developer-led project.

    Based on documented features, setup and access terms. How we rate

    User reviews 0

    The main score switches to user ratings after 5 published reviews.

    Be the first to share your experience

    A specific example helps the next person choose the right tool.

    Write a review

    How was your experience with prediction-market-backtesting?

    Share what you used, what worked and what could be better. We confirm your email and review the submission before publishing.

    To submit a review, email info@polytradingbot.com with the tool name, your rating and your experience. Enable JavaScript to prepare a draft here.

    Related tools

    Browse category
    Marketlens Python SDK: Python installation and Strategy quickstart.
    Marketlens Python SDK
    SDK4.0Editorial

    Retrieve historical Polymarket books and run either tick-level execution simulations or faster bar-based signal backtests in Python.

    • Language Python

    An alternative data-plus-replay SDK.

    NautilusTrader: Polymarket adapter scope and integration-guide entry point.
    NautilusTrader
    SDK4.0Editorial

    Use a Rust-native Polymarket adapter inside NautilusTrader for live feeds, signed execution and strategy research.

    • Language Python · Rust

    Underlying simulation and trading framework.

    pmxt archive: Public mirror index identifying PMXT V1 and V2 capture windows.
    pmxt archive
    Dataset3.5Editorial

    Access the earlier PMXT Polymarket capture eras through a public mirror, with version-specific formats and documented coverage gaps.

    • Access Parquet, Python, DuckDB

    One historical capture source.

    Frequently asked questions about prediction-market-backtesting

    Can I run a Kalshi backtest immediately?

    The README explicitly separates its Kalshi research plumbing from a runnable public path.

    Does it support notebooks?

    Notebook runners are part of the documented workflow.

    Can several markets share a portfolio?

    Yes. Joint multi-market runners are supported.

    Why do two vendors produce different results?

    Coverage, event ordering, depth and loader assumptions can differ.