Does a fixed 2% stop and 4% target bracket make crypto trades profitable?
Short answer: On BTC and ETH from 2023 to 2026, a 2% stop and 4% target bracket hit the target about one trade in three, the share chance predicts, for random entries and trend entries alike. Every entry rule lost between 0.16% and 0.37% per trade after costs. The bracket shapes each trade; it does not create an edge.
Two sources pointed the same way. A trader on a popular forum swears by a fixed 1:2 bracket: risk 1, aim for 2, no discretion. And a published study of 3,505 AI trading agents trading real money found that a fixed 2% stop and 4% target beat every exit the agents chose for themselves. So we tested the bracket on the two most traded coins.
How we tested
- The rule was written and timestamped before any data was pulled.
- Every trade: stop at -2%, target at +4%, time exit after 7 days. Walked on 15-minute candles; when one candle touched both levels, the stop counts.
- Three entries, none tuned: random (the null), random only when the daily trend was up, and a 20-hour breakout only when the daily trend was up.
- Cost 0.30% per round trip. Long only, one trade at a time.
- Verdict on 2023 to September 2026, with 2019 to 2022 reported alongside.
- Pass rule: for the breakout entry on both coins, a positive average per trade that stays positive at the 95% lower bound, and better than a plain 24-hour exit on the same entries.
Results, 2023 to 2026
| Entry | Trades | Average per trade | Won | Hit target / hit stop | Compounded |
|---|---|---|---|---|---|
| BTC, random | 468 | -0.16% | 37% | 34% / 62% | -61% |
| BTC, trend plus random | 324 | -0.16% | 37% | 35% / 63% | -47% |
| BTC, trend breakout | 302 | -0.24% | 34% | 33% / 65% | -57% |
| ETH, random | 616 | -0.28% | 34% | 33% / 66% | -86% |
| ETH, trend plus random | 373 | -0.37% | 33% | 32% / 67% | -78% |
| ETH, trend breakout | 303 | -0.29% | 34% | 33% / 66% | -63% |
| Holding BTC / ETH | +406% / +125% |
2019 to 2022 looked the same: every entry lost between 0.15% and 0.36% per trade.
Why one in three
If a price has no edge, the chance of moving up 4% before moving down 2% is about one in three. Every entry we tried, including the trend breakout, hit the target 32% to 37% of the time. That is the bracket’s geometry and carries no signal. A fair bet minus 0.30% in costs per trade is a slow, steady loss, and over hundreds of trades it compounded to between -47% and -86%.
The study of AI agents was a relative finding: the bracket was the least bad of the exits they measured, and it still lost. Our data agrees. On ETH the bracket did beat a plain 24-hour exit (-0.29% against -0.40% per trade). Both lost.
Limits
- 15-minute candles cannot show the order of moves inside a candle. Counting the stop first when both levels are touched biases results slightly down.
- Three simple entries. A bracket could still help an entry that has a real edge; it cannot supply one.
Verdict
FAIL. A fixed 2% stop and 4% target won one trade in three on every entry, exactly what chance predicts, and lost money after costs on both coins. Not financial advice.