Benjamin-Cup

Posted on Mar 16

Polymarket trading bot : Mountain of Dual-side arbitrage small profit

#polymarket #arbitrage #python #opensource

I recently built an automated volatility / probability arbitrage bot for prediction markets, and the results have been fascinating.

Instead of trying to guess the winner, the bot focuses on market inefficiencies. Prediction markets often move extremely fast — especially short-duration crypto markets — and small pricing imbalances appear constantly. The idea is simple: detect those mispricings faster than humans and capture small edges repeatedly.

Here are a few of the core strategies the bot uses:

1️⃣ Dual-side arbitrage (YES + NO < $1)
Sometimes when a new market opens — especially short 5-minute markets — liquidity is thin and prices drift apart.

Example:

YES = $0.48
NO = $0.47
Total = $0.95

Since one side must resolve to $1, buying both creates a locked $0.05 profit per set.

The bot continuously scans markets and places maker limit orders to capture these situations. Individually the edge is small, but with hundreds or thousands of trades per day, the compounding effect becomes meaningful.

2️⃣ Model vs. Market probability gaps
Another part of the system estimates a “true probability” using a lightweight model combining signals such as:

momentum in price movement
Bayesian updates based on new information
reaction to sudden news, tweets, or macro signals

Trades are only executed when there is a clear statistical edge:

EV = (model probability − market price) > threshold

Position sizing is controlled using fractional Kelly, allowing the system to scale exposure while avoiding excessive risk during volatile periods.

3️⃣ Spread capture / micro market-making
Short-term markets are extremely volatile.

Instead of predicting direction, the bot often quotes both sides of the order book, capturing:

bid-ask spread
maker rebates
micro mean-reversion during rapid price swings

This turns volatility into an opportunity rather than a threat.

4️⃣ Risk management and hedge protection

In my opinion, the real edge is not just the strategy — it’s risk management.

A few key principles I implemented:

• Position caps per market – prevents overexposure to a single event
• Liquidity-aware order sizing – avoids being trapped in thin markets
• Dynamic hedge logic – if probability shifts sharply, the bot hedges with the opposite side to neutralize risk
• Drawdown protection – the system automatically reduces position sizes after consecutive losses
• Inventory balancing – ensures the bot never becomes heavily biased toward YES or NO positions

This essentially allows the system to hedge losses quickly and keep overall exposure close to neutral, which is crucial when trading very short-term markets.

What surprised me the most

The biggest lesson wasn’t just the strategy — it was the importance of automation and speed.

Many opportunities exist for only a few seconds, especially when:

new markets open
sudden news drops
liquidity temporarily disappears

Humans simply can’t react fast enough.

The result:
Small statistical edges compounded across large trade volumes with strict risk control.

Building this system was a great experiment combining:

probability theory
automation
market microstructure
risk management
and quantitative execution

Prediction markets are still relatively inefficient compared to traditional financial markets, which makes them a really interesting playground for quantitative strategies.

I’m continuing to refine the models and execution logic — especially around latency, signal accuracy, and smarter hedging mechanisms.

Curious if anyone else here is experimenting with quant or arbitrage strategies in prediction markets. Always open to exchanging ideas. 🚀

Contributing

Contributions are welcome.

Submit ideas, pull requests, or issues on GitHub.

https://github.com/Gabagool2-2/polymarket-trading-bot-python

Continuous Updates & Development

This Polymarket trading bot is actively maintained and continuously updated to adapt to new Polymarket trading opportunities, crypto market conditions, and strategy improvements.

New features, optimizations, and trading strategy enhancements are released regularly to improve performance, stability, and profitability.

If you're interested in:

Polymarket trading automation

crypto trading strategies

prediction market bots

or contributing to the project

feel free to stay in touch.

If you'd like to see the system in action, I can arrange a live Google Meeting demonstration to showcase the bot running in real time and explain how the trading strategies operate.

I'm always happy to connect with developers, traders, and researchers working in the Polymarket and crypto ecosystem.