polymarket-bot-lab

polymarket-bot-lab is a research-oriented repository that systematically evaluates 11 different trading strategies on Polymarket historical data. The strategies range from simple (MA crossover, RSI mean reversion, volume breakout) to moderately complex (Random Forest classifier on market features, LSTM on price sequences). Each strategy is implemented as a standalone Python class with a unified interface: `train()`, `predict()`, `backtest()`. The companion 'polymarket-canary-tape' dataset provides cleaned historical CLOB snapshots and trade data. The key differentiator is the 'honest results' promise: the author claims to publish all backtest results regardless of performance, including losing strategies. This is refreshing in a space full of overfit backtests. The code quality is decent for a research framework: Jupyter notebooks for exploration, Python scripts for backtesting, and basic visualization. However, the repo is brand new (created today, July 22, 2026) and has zero stars, so code review is limited. There are no deployment scripts or live trading modules—it's purely for research. Malware risk is minimal: no binaries, standard Python data science stack (pandas, scikit-learn, numpy, matplotlib). The framework is ideal for anyone wanting to understand which factors actually predict Polymarket price movements without the hype.