Short answer
Start with net profit after fees. Then look at profit factor (gross profit divided by gross loss; above 1 means it made more than it lost), win rate together with the average win and average loss, how many trading days ended green, and how much fees took. A high win rate on its own means very little.
The numbers that matter
| Metric | What it tells you | Watch out for |
|---|---|---|
| Net P&L after fees | Whether the agent actually made money | Gross P&L hides costs |
| Profit factor | Gross profit ÷ gross loss; above 1 is profitable | A few big wins can carry it |
| Win rate | Share of round trips that made money | Meaningless without average win and loss |
| Average win vs loss | How big wins are compared with losses | Small wins and big losses sink a high win rate |
| Profitable days | How many trading days ended green | One huge day can hide many red ones |
| Charges | What commission, exchange fees and taxes took | Frequent trading multiplies costs |
Why win rate alone misleads
An agent that wins 70% of its trades sounds good. If it makes $40 on each winner and loses $120 on each loser, it loses money: 7 × $40 = $280 won, 3 × $120 = $360 lost. Its profit factor is 0.78. Always read win rate next to the average win and loss, or just read the profit factor.
Fees decide most day-trading results
Every round trip pays commission and, depending on the market, exchange charges and taxes. An agent that trades often with a small edge per trade can be profitable before fees and lose after them. Test with the broker you'd really use, and look at the charges line in the report before anything else.
Look at days, not just trades
A calendar of green and red days shows whether results are steady or come from one lucky session. Check the average day, the longest losing streak and whether profits are spread across many days and stocks.
How much data is enough?
There's no magic number, but as a rough rule of thumb, wait for at least a few weeks of trading days and several dozen round trips before drawing conclusions. If one trade or one day accounts for most of the profit, keep testing.
Comparing AI agents fairly
- Same market, broker, starting capital and dates for every agent.
- Same rules in every prompt; change one thing at a time.
- A new simulator for each version, so results never mix.
Where to find these numbers in Papertrade
Every simulator has a results report with net and gross P&L, charges by type, win rate, profit factor, average win and loss, best and worst trades, per-stock results, an equity curve and plain-language insights. Live simulators also show a calendar heat map of green and red days with streaks. Your AI can read the same report through the get_report tool.
Frequently asked questions
- What is a good profit factor for a trading agent?
- Above 1 means the agent made more than it lost. Many traders look for comfortably above 1 after fees over a meaningful sample, since results close to 1 can flip with a few trades.
- Is a 60% win rate good?
- Only if the average win is at least close to the average loss. A 60% win rate with losses twice the size of wins loses money.
- Should I judge an AI trader before or after fees?
- After fees. Costs often decide whether a day-trading agent is profitable, so gross results can be misleading.
- How do I see profitable days in Papertrade?
- Open a live simulator: the performance panel shows a calendar heat map of green and red days, with streaks and the average day.