Learn · tested on history
Timeframe and risk per trade: what common trading mistakes cost
Experienced traders most often criticise beginners for two things: trading too small a timeframe and risking too much per trade. We measured what that costs in money. The rule is always the same — buy when EMA 20 crosses above EMA 50, exit on the cross back, 3 ATR stop. Only one thing changes: the timeframe or the risk.
How: 2021-01-01 – 2026-10-31, BTC/USDT, ETH/USDT and SOL/USDT on Binance 1-minute candles; 0.1% fee per side and 0.05% slippage on every market order; no leverage. The period is cut by time: train 60%, validation 20%, out-of-sample the last 20%. "Held up" means profitable after costs on both validation and out-of-sample. Every setting was fixed in advance; nothing was tuned to the result.
1. Timeframe: the same strategy on 5m, 15m, 1h and 4h
On 5-minute candles the strategy looks busy: many trades, many signals. But every trade pays the fee and the slippage, and a move of a few minutes is usually smaller than those costs. Look at the "Costs" column: fees plus slippage as a share of the starting capital.
| Coin | Timeframe | Check | Whole period | Out-of-sample | Max drawdown | Trades | Costs | Report |
|---|---|---|---|---|---|---|---|---|
| BTC | 5m | failed | −100% | −100.0% | −100% | 6 542 | 132% | Open |
| BTC | 15m | failed | −99% | −65.4% | −99% | 2 043 | 149% | Open |
| BTC | 1h | failed | −39% | −22.6% | −62% | 480 | 103% | Open |
| BTC | 4h | held up | +75% | +4.1% | −27% | 117 | 23% | Open |
| ETH | 5m | failed | −100% | −99.0% | −100% | 6 386 | 186% | Open |
| ETH | 15m | failed | −97% | −79.6% | −99% | 2 048 | 344% | Open |
| ETH | 1h | failed | +27% | −30.8% | −45% | 487 | 152% | Open |
| ETH | 4h | fragile | +88% | +2.3% | −21% | 119 | 19% | Open |
| SOL | 5m | failed | −100% | −98.3% | −100% | 6 149 | 461% | Open |
| SOL | 15m | failed | −92% | −76.8% | −98% | 1 969 | 439% | Open |
| SOL | 1h | failed | +586% | −15.3% | −48% | 457 | 419% | Open |
| SOL | 4h | held up | +525% | +15.4% | −26% | 115 | 37% | Open |
Takeaway: the smaller the timeframe, the more trades and the larger the share of the gain eaten by costs. On 4h the same logic trades rarely and costs stay a small part of the result.
2. Risk per trade: 1%, 2% and 5%
Risk per trade is the share of the account lost if the stop is hit. The position is sized from it: the higher the risk, the larger the position. In good years a high risk gives big numbers, but drawdowns grow with it, and one bad streak throws the account far back.
| Coin | Risk | Check | Whole period | Out-of-sample | Max drawdown | Trades | Costs | Report |
|---|---|---|---|---|---|---|---|---|
| BTC | 1% | held up | +43% | +3.8% | −15% | 117 | 10% | Open |
| BTC | 2% | held up | +75% | +4.1% | −27% | 117 | 23% | Open |
| BTC | 5% | failed | +60% | −6.9% | −52% | 117 | 36% | Open |
| ETH | 1% | fragile | +43% | +1.8% | −11% | 119 | 8% | Open |
| ETH | 2% | fragile | +88% | +2.3% | −21% | 119 | 19% | Open |
| ETH | 5% | failed | +69% | −16.5% | −43% | 119 | 41% | Open |
| SOL | 1% | held up | +174% | +8.2% | −16% | 115 | 10% | Open |
| SOL | 2% | held up | +525% | +15.4% | −26% | 115 | 37% | Open |
| SOL | 5% | held up | +2 923% | +7.4% | −46% | 115 | 341% | Open |
Takeaway: return and drawdown grow together. At 5% drawdowns are several times deeper, and a strategy that held up at 1–2% can fail: the losing streak lands on large positions.
Check your own idea
Describe a strategy in your own words and we test it the same way: on history, with costs, on data it has not seen.