Claude Built a Trading Bot That Made $518K on Polymarket

What happened

Someone asked Claude to write a trading script for Polymarket’s hourly crypto Up/Down markets. The script was deployed, left running — and never stopped.

Six weeks later, it’s the single most profitable crypto strategy on the platform.

• $518,909 total profit
• $51.27M total volume
• 17,700+ markets traded
• Fully autonomous — zero human intervention since launch

The bot trades BTC, ETH, SOL, and XRP hourly binary markets around the clock. It buys both Up and Down on the same market with asymmetric sizing, capturing the spread between the two sides. When Up + Down prices sum to less than $1.00, both sides lock in guaranteed profit on resolution.

No edge in prediction. Pure execution speed and math.

The numbers

Period	Profit	Volume
All Time	$518,909	$51.27M
Month	$70,382	$7.6M
Week	$65,005	$308K
Today	$710	$49K

The strategy generates roughly $2,300/day on average. Some days more, some days less — but the equity curve barely dips. It compounds silently while you sleep.

Why this matters

This wasn’t built by a quant fund. It wasn’t designed by a team of PhDs. It was one prompt to Claude that produced a working spread-capture algorithm. The script runs on commodity infrastructure, costs pennies per hour, and outperforms 99.9% of human traders on Polymarket.

The implications:

1. The barrier to algo-trading is now zero. You don’t need to code. You need to describe what you want.
2. Execution beats prediction. This bot has no opinion on whether Bitcoin goes up or down. It doesn’t need one.
3. Speed compounds. 17,700 trades in 6 weeks = ~420 trades/day. No human can operate at this frequency. No human needs to.

How to copy it

The strategy is live on Merlin. You can copy every trade in real-time through the Merlin bot — the same positions, the same sizing logic, scaled to your capital.

→ Copy this strategy on Merlin

Merlin’s take

We’re past the point of debating whether AI can trade. It’s trading. It’s profitable. And it just showed that a single conversation with Claude can produce a strategy generating $500K+ in weeks.

The question isn’t whether to copy it — it’s how long you wait before the edge gets arbitraged away. Hourly crypto markets are still inefficient enough for this to work. That window won’t stay open forever.