There is a former Citibank quantitative trader who created a Polymarket bot, turning $1,200 into $797,000.

I tracked his 43,495 trades using Opus 4.7, then tested them against a dataset of 72 million trades to analyze his strategy.
The core of this bot is using a Markov chain to analyze sequences of price states, identifying entry opportunities most traders can't see.
Most traders see "95 cents" and place orders based on intuition—thinking it's cheap, expensive, or about the same, and they jump in with a single click.
But this bot doesn't rely on gut feeling. It looks at how prices bounce between different states over the past N windows and calculates the probability of transitioning from one state to the next.
To put it simply:
Within the same window, if the DOWN direction is between 94 and 98 cents, and the UP direction is between 3 and 5 cents, totaling less than a dollar, then this price difference is an arbitrage opportunity.
The Markov state model marks when there's over a 40% chance that a 3-cent UP will lead to settlement, and the bot automatically adjusts its position.
For example, if BTC rises by 3.4 cents, from a previous 7.5 cents, that's a 120% increase within one window.
And it runs simultaneously in both directions on BTC and ETH.
Here's the personal homepage of this trader:

If you want to copy trade on Polymarket, it’s recommended to use Kreo:

#Polymarket