Money Basics

Backtest Truth, Part 5: Risk Per Trade and Surviving a Drawdown

At 2% risk, one test made about +23% in a year with an 18% drawdown. The same family all-in made about +11%, and a strong setup still saw a 65% drawdown on older data.

On this page
  1. All-in versus 2% risk on the same rules
  2. Smaller risk does not repair a bad rule
  3. Even the stronger test got hurt
  4. Letting winners run
  5. A last look at the whole series
  6. Key takeaways
  7. FAQ

This is education only and not financial advice. Trading involves risk, and you can lose money.

Risk per trade is how much of your account you are prepared to lose if the stop is hit. A common beginner range is 1–2% per trade. That is not the size of the position. It is the loss if the stop is hit: stop distance times position size, as a share of the account. A wider stop needs a smaller position. A tighter stop can be larger, but if it is too tight, normal noise can hit it.

All-in versus 2% risk on the same rules

“All in” looks simpler, but it is rougher. The position is the whole account, so a drawdown in the price is a drawdown in the account, one for one. The same 4-hour trend rules were compared at full size versus 2% risk.

At 2% risk the 1-year result was about +23% with an 18% drawdown, and the 2-year result was about +26% with a 24% drawdown. The all-in version made only about +11% in both windows, with deeper drawdowns. After a loss, 2% risk still had more of the account left to compound. Compounding means the next trade is sized off what remains. If a loss took a huge share, the next win has less money to work with. Keeping the planned loss near 2% of the account left more of the account standing.

This is not a promise that 2% risk will make about +23% or +26% in your own future. Those figures are what this comparison printed on these rules, in these windows. The lesson you can carry is the shape: the capped-risk version gained more in the test and took a shallower path than the all-in version of the same rules.

Smaller risk does not repair a bad rule

A 2% cap also reduced pain on a weaker rule. The 2% Renko flip lost money over 2 years either way, but the drawdown shrank from about 51% all-in to about 29% at 2% risk. Smaller risk does not turn a bad rule into a good one. It can keep a bad stretch from becoming a collapse.

Part 4 already showed that this flip rule lost about 27% over 2 years with a 51% drawdown in the all-in picture of that window. Here the new comparison is the depth of the drop when risk is capped at 2%: about 29% instead of about 51%. The rule was still a loser over those 2 years. Surviving a loser with more of the account left is not the same as finding a winner.

Even the stronger test got hurt

Even good systems get hurt. The stronger pullback setup, at 2% risk, had about a 24–25% drawdown on recent 1-year and 2-year windows. On 2020–2024 data the drawdown was about 65%. That is why a slightly smaller risk, such as 1.5%, is a safer starting point. A 25–65% drawdown can happen in a rule that still finishes profitable. If you cannot live through that drop, the paper return does not matter.

A paper return is the ending percentage in a test. It does not pay rent during the months when the account is down 25%, or 65%, from a high. An OFW account often has a job to get back to, family who expect a remittance, and no spare year to “wait for the backtest to recover.” If a drop of that size would force you to stop, or to break the rules, the ending percentage is a number you will never personally reach. The starting risk has to be small enough that the ugly middle is still a life you can pay for.

Letting winners run

The other half of survival is letting winners run. When a fixed 2-to-1 target was used instead of a trailing stop, the last year went almost flat, about −1%, because the target cut off the runners. A 2-to-1 target means the planned gain is twice the planned loss, and the trade is closed when that gain is hit. The trailing version of the same idea stayed green. A trailing stop moves with the trade, so a large move is not cut off at a small target.

In several of the better tests, only about 1 in 3 trades won. The plan worked when the winners were larger than the losers, and when no single loss could sink the account. A win rate near one in three sounds poor if you expect most trades to work. It can still fit a plan when the winners pay for more than the losers cost, and when each loser is a small, planned slice of the account. That is the pair: room for winners, and a cap on each loss. Neither half replaces the other. The fixed target made the last year about −1% by cutting the runners. The missing stop, in Part 3, let one hold become a long problem. You need both limits.

A last look at the whole series

Part 1 asked whether the fill was a price you could have got. Part 2 asked whether the result survived a realistic fee and a slightly worse fee. Part 3 asked whether every trade had a stop. Part 4 asked whether the good window was the only window. Part 5 asks whether the risk per trade leaves you able to sit through a drawdown of the size these tests actually printed, from about 18% and 24% on one comparison, through about 24–25% and about 29%, up to about 51% and about 65% in the harder windows.

None of those figures is a forecast. They are records of tests on Bitcoin rules. They do not say what to buy or sell tomorrow. They say what an honest test had to survive before anyone should treat it as a plan.

Key takeaways

  • Risking about 1–2% per trade means the loss at the stop, not the full size of the position.
  • Drawdowns of 25–65% can happen even in profitable tests, so the risk must stay small enough to survive them.
  • A plan also needs room for winners to run. Cutting every trade at a small target can erase the edge.

FAQ

What does 1–2% risk per trade mean?

It is the loss if the stop is hit, as a share of the account. It is not the size of the whole position. A wider stop needs a smaller position. A tighter stop can be larger, but normal noise can hit a stop that is too tight.

How did 2% risk compare with all-in on the 4-hour rules?

At 2% risk the 1-year result was about +23% with an 18% drawdown, and the 2-year result was about +26% with a 24% drawdown. The all-in version made only about +11% in both windows, with deeper drawdowns.

Did smaller risk fix the Renko flip rule?

No. It lost money over 2 years either way. The drawdown shrank from about 51% all-in to about 29% at 2% risk. Smaller risk does not turn a bad rule into a good one.

Why did a fixed 2-to-1 target hurt?

The last year went almost flat, about −1%, because the target cut off the runners. The trailing version of the same idea stayed green. In several of the better tests only about 1 in 3 trades won, and the plan needed those winners to be larger than the losers.

This is education only and not financial advice. Trading involves risk, and you can lose money.