forex-scalping.com

Raising risk from 1% to 2% on good setups: simulated drawdown over 200 trades

Many traders risk 1% on a normal setup and double it when a trade looks perfect. Before doing that it helps to see what size does to the drawdown, using the same trade statistics throughout.

I simulated 20,000 runs of 200 trades with a 45% win rate and 1.5R winners, an expectancy of 0.125R per trade. Each trade risks a fixed percentage of current equity, so a win adds 1.5 times the risk and a loss removes the risk.

At 1% risk, the median maximum drawdown was 11.8%, the 95th percentile was 21.2%, and the median finish was a gain of 26.4%. The chance that the drawdown reached 20% at some point was 7.0%.

At 1.5% risk the median drawdown was 17.3% and the 95th percentile 30.7%, the median gain 40.5%, and 33.6% of runs hit a 20% drawdown. At 2% the median drawdown was 22.6%, the 95th percentile 38.7%, and the median gain 55.0%, with 64.8% of runs hitting 20%. At 3% the median drawdown was 32.2%, the median gain 84.2% and 95.0% of runs went through 20%.

So doubling the risk raised the median gain by about 2 times, from 26% to 55%, while the chance of a 20% drawdown went from 7% to 65%. That is a heavy price for a small improvement in typical results. The gain from risk scales roughly with the risk, but the drawdown grows faster at the tail.

If you size up on special setups, this table is what you're agreeing to on an average month.

A few caveats on the numbers. The simulation uses independent trades with a fixed 45% and 1.5R, and real results have a spread of outcomes, not two fixed values. It also assumes that you apply the same risk percentage to every trade, whereas a trader who raises size only on some trades has a mix. The mix is easy to model: the drawdown at an average risk of 1.5% is between the 1% and 2% rows, closer to the 1.5% row of 17.3% median. So if one trade in two is risked at 2% and the rest at 1%, you can use the 1.5% row as a first guide.

Another way to read the table is the ratio of median gain to median drawdown. At 1% it's 26.4 / 11.8 = 2.2, at 1.5% it's 40.5 / 17.3 = 2.3, at 2% it's 55.0 / 22.6 = 2.4, and at 3% it's 84.2 / 32.2 = 2.6. That ratio only moves from 2.2 to 2.6 across the table. So you aren't buying a better risk-adjusted result by sizing up, you're only scaling the gains and the pain by the same factor, and the thing that decides is how much drawdown you can accept in the account or the prop rules.

If the account has a 10% maximum loss rule, none of these rows is safe: at 1% risk the median drawdown is already 11.8%. This is a point that gets lost when people plan from the expected result. I ran the same simulation at 0.5% risk. The median maximum drawdown was 5.9%, the 95th percentile 11.2%, and 8.9% of runs went through 10% at some point. So even half the usual size leaves about a one in eleven chance of hitting a 10% limit over 200 trades, and a prop account with that rule needs the risk lower again or a smaller number of trades per attempt.

The takeaway for position sizing is to pick the drawdown you can live with first, and then pick the risk from the table, not the other way round. If 20% is a deal breaker and a 7% chance of it is acceptable, 1% fits. If you'd accept a 1 in 3 chance, 1.5% fits. Doubling for good-looking setups is just the 2% row, with a 65% chance of a 20% drawdown over 200 trades. Whatever your own edge is, the sim takes ten lines of code, so you can rerun it with your real win rate and payoff.