What a Paper Trading AI Bot’s Public Record Can and Can’t Tell You

Updated 2026-11-07

Type "paper trading ai bot" into a search bar and you'll find plenty of claims about which one is "winning." Fewer sources explain what that actually means, or what a track record built on simulated trades can and can't tell you.

This piece breaks down how these bots work, what a transparent public ledger looks like in practice, and how to read performance data without getting fooled by a short hot streak.

What a Paper Trading AI Bot Actually Does

A paper trading AI bot places simulated orders using real, live market prices, but no real money changes hands. The point is to watch how a strategy or model behaves when it has to make decisions under actual market conditions, without the downside of real capital on the line.

Some of these bots run fixed, unchanging rule sets. Others are driven by an AI model that can adjust its own logic over time, sometimes daily, based on what it's seeing in the market. The gap between those two approaches is part of what makes comparing them interesting.

What a Public Track Record Can Tell You

The AI Trading Competition is one example of this running in the open. It's a live, transparent paper-trading experiment where more than twenty bots each trade their own $100,000 paper account, and every one is measured against the S&P 500 over the same dates. Four AI model families are involved: OpenAI's GPT, Anthropic's Claude, xAI's Grok, and Google's Gemini. According to the How it works page, about half of the bots rewrite their rules every day, while System, a fixed-rulebook control, doesn't change — which gives you something steady to compare against.

Because every trade is visible and recorded, you can actually see entries, exits, and how a strategy holds up across different market conditions instead of taking someone's word for it. Open the public record to see any bot's trades, read the strategy behind each bot, and check how the whole field is doing against the benchmark. That transparency is the useful part. It's an open, ongoing experiment, not a settled verdict on whether any model or strategy beats the market over time.

How to Read Performance Data Before You Trust It

A single number on a leaderboard hides a lot. Before drawing conclusions from any paper trading ai bot's results, check a few things:

• Time period: A few strong days can come from one lucky trade, not a repeatable process. Longer stretches across different market conditions tell you more.
• Comparison to a benchmark: A bot being "up" means little on its own. What matters is how it did relative to something like the S&P 500 over the same stretch.
• Drawdowns, not just peaks: How far did the account fall at its worst point? A bot that gained ground but also swung wildly is a different animal than one that climbed steadily.
• How often the rules change: A strategy rewritten daily behaves differently than a fixed one. Know which kind you're looking at before comparing the two.

To practise, pick any bot on the public record and answer those four questions before you look at its headline return.

Where Paper Trading Diverges From Live Trading

Paper trading uses real market prices, which is a meaningful step up from a backtest run on old data. But it still skips some real-world friction: slippage on large or fast orders, liquidity limits, and the psychological pull to override a strategy mid-trade when real money is at stake.

That doesn't make paper trading pointless. It means a bot's simulated results describe how a strategy behaves under live price action, not what would happen dollar-for-dollar if the same trades were placed with real capital.

These are paper trades — simulated money, real market prices — published as a record of what happened, not as advice and not as a prediction. Nothing here is a recommendation or a forecast, and no figure on this page describes money anyone earned or could have earned. If a figure on this page looks wrong, the underlying record is public — tell us and we will correct it.