A test runs your rules on past Binance prices and shows what would have happened. It can’t tell you what will happen next. Its job is to show how a rule behaves.
Run a test
Open Bots, then Historical testing, and answer three questions.
- Which strategy. Each card shows that strategy’s last result.
- Which period. Use a preset or pick dates on the timeline.
- How much. The starting balance. Fees and slippage are included and shown on one line.
Choose Run test. Results collect beside the form, and each one opens a full report.
Read the report in this order
- Compared with holding. The report draws your rule next to simply holding the coin. A rule can make money and still do worse than doing nothing.
- The worst drop. The furthest the balance fell from a high. Ask yourself whether you’d have kept the bot running through it.
- The number of trades. A handful of trades proves very little. Tradenos warns you when a sample is small.
- Trades won. A high share of wins can still lose money if the losses are bigger than the wins.
- Fees. Every trade pays them twice. A rule that trades many times a week needs a much bigger edge.
Three ways a test can fool you
- Too short a period. A rule tested only on a rising market has never seen a fall.
- Too much tuning. If you adjust the numbers until the past looks perfect, you’ve fitted the past. Round numbers that work on several coins are safer.
- Ignoring costs. A result without fees and slippage will look better than real trading.
What your plan includes
| Newcomer | Learner | Strategist | Trader | Operator | |
|---|---|---|---|---|---|
| Tests a month | 10 | 50 | 300 | 1,000 | 3,000 |
| History you can test on | 6 months | 1 year | 3 years | Full | Full |
When a test looks reasonable, run the bot in Practice before you use real money.