Skip to main content
AI Learning Resources

Quick Overview

In this results review video, the creator behind the All About AI channel presents performance data from thirty days of automated bot trading across Kalshi and Polymarket. The video examines live metrics from three algorithmic strategies running on a virtual private server and concludes with a live demonstration of an autonomous order-execution script.

Key Points

  • 1.The one-cent Bitcoin lottery bot generated a modest net profit of thirteen dollars and twenty-four cents over three weeks, but experienced a significant maximum drawdown of sixty-three dollars and sixty-five cents.
  • 2.The QuantWFX automated weather trading model was the best-performing strategy, delivering a twenty-eight point eighty-seven percent return with a low maximum drawdown of ten dollars and forty-seven cents.
  • 3.A machine learning model targeting five-minute Bitcoin price dislocations achieved a seventy-one point eighty-eight percent win rate and one hundred and ten dollars in net profit across thirty-two trades.
  • 4.All three algorithmic strategies run autonomously on a virtual private server, requiring monitoring rather than manual execution.
  • 5.A live demonstration of the Kalshi weather bot showed automated placement of resting limit orders on temperature prediction markets, resulting in an immediate fill on a Miami weather contract.

Summary

The video provides a thirty-day performance review of three automated algorithmic trading strategies deployed across prediction market platforms, specifically Polymarket and Kalshi. The presenter explains that multiple bots have been running in the background on a virtual private server, with unprofitable systems shut down early while the three longest-running models continued operating autonomously.

The first strategy examined is the one-cent Bitcoin lottery approach, which places one-cent bids on both sides of five-minute binary contract windows in hopes of capturing outsized returns on low-probability fills. Across three weeks of live tracking, the model generated thirteen dollars and twenty-four cents in net profit, yielding a three point fifty-one percent return. However, the strategy suffered a steep sixty-three dollar and sixty-five cent maximum drawdown. The presenter notes that while the model stayed positive overall, its robustness declined sharply in the third week, making it relatively inefficient for capital growth.

The second model reviewed is QuantWFX, an automated weather trading strategy operating on event markets. Over twenty-five market days on an account of three to four hundred dollars, the system generated a net profit of one hundred and twenty-one dollars and eighty-five cents, representing a return of twenty-eight point eighty-seven percent. The bot posted a fifty-eight point eighty-two percent win rate and maintained a low maximum drawdown of ten dollars and forty-seven cents. Because of its steady equity curve and low downside risk, the presenter identifies QuantWFX as the strongest performing setup in the portfolio.

The third strategy evaluates a machine learning model built to forecast five-minute Bitcoin up and down markets and execute trades when pricing discrepancies arise. The model operates at a low frequency, averaging approximately one trade per day for a total of thirty-two completed trades over roughly a month. It delivered a net profit of one hundred and ten dollars and two cents, an eighteen point sixty-eight percent return, and a seventy-one point eighty-eight percent win rate. Although it experienced an eighty-dollar drawdown earlier in the test run, adjustments made to the model resulted in a sustained upward equity streak over the final two weeks.

The presenter also displays MaxxQuant, a micro hedge fund project designed for AI agents with investment increments starting at one-millionth of a dollar. To conclude the session, the presenter connects live to a terminal running an automated Kalshi weather script at market opening around four in the afternoon. The system submits automated limit orders across Miami and New York City temperature brackets, successfully filling one contract at fourteen cents and securing an immediate two-cent unrealized profit.

One-Cent Bitcoin Binary Window Strategy

The presenter reviews performance data for a strategy that places one-cent bids on both sides of five-minute Bitcoin binary prediction windows. Over roughly three weeks of runtime, the strategy achieved a net profit of thirteen dollars and twenty-four cents, representing a return of three point fifty-one percent. Despite remaining profitable overall, the strategy suffered a severe maximum drawdown of sixty-three dollars and sixty-five cents due to shifting market conditions and low robustness during the final week.

QuantWFX Weather Trading Strategy

The QuantWFX system automates trading on event prediction markets based on weather forecasts. Over twenty-five market days, the model produced one hundred and twenty-one dollars and eighty-five cents in net profit on a starting balance of three to four hundred dollars, achieving a twenty-eight point eighty-seven percent return. With a win rate of fifty-eight point eighty-two percent and a tight maximum drawdown of ten dollars and forty-seven cents, the presenter highlights this model as the most reliable and consistent of the group.

Machine Learning Model for Bitcoin Up-Down Markets

The third strategy uses a machine learning model to detect pricing dislocations in five-minute Bitcoin up and down prediction markets. Operating at a low frequency of approximately one trade per day, the bot executed thirty-two trades over roughly a month. It recorded a seventy-one point eighty-eight percent win rate and a net profit of one hundred and ten dollars and two cents, equal to an eighteen point sixty-eight percent return, despite an eighty-dollar drawdown earlier in the testing period.

Live Order Execution on Kalshi

The presenter connects to a virtual private server around four in the afternoon to demonstrate an autonomous bot placing resting orders on Kalshi temperature markets for Miami and New York City. The script calculates opening price expectations across temperature buckets and submits limit bids into the order book. While several resting bids remained open due to wide spreads, one Miami temperature contract filled at fourteen cents and immediately showed an unrealized gain of two cents.

The Bottom Line

The video demonstrates that autonomous trading bots can generate positive net returns on prediction platforms like Kalshi and Polymarket, with weather prediction and machine learning price dislocation models significantly outperforming naive lottery bidding strategies. While the QuantWFX and machine learning models achieved solid returns with respectable win rates, drawdowns in the Bitcoin models illustrate the vulnerability of these systems to shifting market regimes. The video leaves long-term scalability and future multi-strategy portfolio integration to subsequent testing.

FAQ

What is AI bot trading on Kalshi and Polymarket prediction markets?

It is the use of automated software algorithms and machine learning models running on servers to identify mispriced contracts, execute limit orders, and trade event outcomes autonomously on platforms like Kalshi and Polymarket.

How did the one-cent Bitcoin binary window strategy perform over the testing period?

The strategy produced a net profit of thirteen dollars and twenty-four cents, representing a return of three point fifty-one percent, but it experienced a large maximum drawdown of sixty-three dollars and sixty-five cents.

What were the thirty-day performance results for the QuantWFX weather trading strategy?

QuantWFX achieved a net profit of one hundred and twenty-one dollars and eighty-five cents, a twenty-eight point eighty-seven percent return, a fifty-eight point eighty-two percent win rate, and a maximum drawdown of ten dollars and forty-seven cents over twenty-five market days.

How many trades did the machine learning Bitcoin up-down trading model execute?

The machine learning Bitcoin model operated at a low frequency, executing thirty-two trades over roughly a month, averaging about one trade per day.

How does the automated Kalshi weather trading bot place orders during live execution?

The bot runs autonomously on a virtual private server, monitors new daily temperature markets when they open around four in the afternoon, and places resting limit bids across specific temperature ranges based on forecast data.

Worth watching for

Algorithmic traders, quantitative developers, and individuals interested in automated trading strategies on prediction markets such as Polymarket and Kalshi.

  • polymarket
  • kalshi
  • bot-trading
  • machine-learning
  • prediction-markets
  • algorithmic-trading
Watch on YouTubeMore videos from All About AI