Take a method that wins 50 percent of the time. Over 200 trades, a run of eight consecutive losses is not unusual. It is close to expected. At two percent risk per trade that run costs roughly 15 percent of the account. At five percent it costs about 34 percent, and on a funded account with a 10 percent overall limit it ends the account somewhere around trade five of the eight.
Nothing in that paragraph required a rare event. It is arithmetic on an ordinary streak. A drawdown calculator exists so that the arithmetic happens before the trading rather than after it.
What the tools actually compute
There are three distinct calculations sold under the same name, and mixing them up is the usual source of a misleading answer.
The first is a streak probability. Given a win rate and a number of trades, how likely is a losing run of a given length? For independent trades with loss probability p, the chance of avoiding a run of length n gets small quickly as the sample grows. This is the calculation that tells a 45 percent win-rate trader that a run of nine or ten losses across a year of trading is not bad luck but a scheduled event.
The second is compounded account impact. Apply that streak to a starting balance at a fixed percentage risk. Percentage risk shrinks the loss per trade as the account falls, which is why fixed-fractional sizing cannot mathematically reach zero, while fixed lot sizing can. The output is a peak-to-trough figure you can compare against a rule limit.
The third is Monte Carlo simulation. Take the actual distribution of results from a backtest or a trade history, shuffle the order thousands of times, and record the worst drawdown in each run. What comes back is a distribution rather than a single number: the median worst drawdown, the 95th percentile, and how often the sequence would have breached a given limit. This is by far the most useful of the three, because it uses your real result sizes rather than assuming every win and loss is the same magnitude.
Why the historical maximum understates the risk
A backtest reports one number for maximum drawdown, produced by one specific ordering of trades. Reshuffle the same trades and that number moves, often a lot. Many strategies whose backtest shows a 12 percent maximum will produce runs above 20 percent in a meaningful share of simulations, purely because the losses cluster differently.
Treating the historical maximum as a ceiling is the most common modelling error in retail trading. The realistic planning figure is closer to the 90th or 95th percentile of the simulated distribution, and even that assumes the future distribution resembles the past one. It will not, particularly across a volatility regime change, which is one reason overfitted backtests are dangerous rather than merely useless.
If your planned risk per trade only survives the median simulated drawdown, you are sized for the average future and not the plausible one. Size so the 95th percentile run leaves you trading.
Feeding a calculator numbers that mean something
Garbage in is the real limitation. Four inputs decide the output, and each has a way of being wrong.
Win rate has to come from a sample large enough to be stable. Thirty trades tells you almost nothing; a win rate measured over 30 trades can be off by ten percentage points either way. Average win and average loss matter more than win rate does, since a method winning 35 percent of the time with a three to one payoff has a completely different profile from a 60 percent method scalping for one to one, a relationship covered in expectancy.
Risk per trade must be the realised figure, not the intended one. Slippage, gaps and stops that get widened in the moment mean the realised average loss is usually larger than the plan. Pull it from the trade history rather than the rulebook.
Correlation breaks the independence assumption that streak maths relies on. Three open positions in EURUSD, GBPUSD and AUDUSD against the dollar are close to one position at triple size, and a calculator counting them as three independent trades will understate the drawdown badly. Either model them as a single risk unit or cap total exposure per correlated group.
Using the output on a funded account
Evaluation and funded programmes make this concrete, because the limit is contractual rather than a matter of taste. The question is simple to state: at my risk per trade, what is the probability that a normal losing sequence takes me through the overall limit, and through the daily limit, before I reach the profit target?
Run the simulation against both anchors. A daily loss limit is breached by a cluster inside one session, so the daily figure depends on how many trades you take per day as well as on risk per trade. A trader taking six trades a day at two percent each has a plausible bad day of eight to ten percent, which will breach most daily limits regardless of how good the strategy is over a month. The answer is fewer trades or smaller risk, and the calculator tells you which combination survives. Our guide to prop firm drawdown rules covers how the anchors differ between programmes, since the same risk number passes one rulebook and fails another.
What a calculator cannot do
It cannot tell you whether the strategy has an edge. Every one of these tools assumes the input distribution is real, and a method with negative expectancy produces perfectly precise numbers describing a slow loss. Expectancy comes first; drawdown modelling comes second.
It also assumes you follow the plan. The largest drawdowns in most trading records are not sequences of planned two percent losses. They are one abandoned stop, or a doubled position after a losing morning. No model captures that, which is why the number a calculator gives you is a floor on the risk rather than a ceiling. Trading carries a high risk of loss, and the sizing that survives the simulation is the sizing that gives you a chance of surviving the reality.
Pair the model with a live view of the account so the plan and the present agree. Equity trackers handle the second half of that job.
"Traders pick risk per trade by feel, then discover the streak. Run the numbers first and the streak is an event you planned for instead of a crisis."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- A run of eight losses is ordinary at a 50 percent win rate over a few hundred trades, and the account impact is arithmetic.
- Monte Carlo reshuffling of real trade results beats streak formulas, because it uses actual win and loss sizes.
- A backtest's historical maximum drawdown is one ordering out of many; plan against the 95th percentile instead.
- Correlated open positions break the independence assumption, so model a correlated group as one risk unit.
Frequently Asked Questions
How much drawdown should I plan for?
Run a Monte Carlo simulation over your own trade history and take the 95th percentile worst run rather than the historical maximum. If that figure breaches the limit on your account, the answer is to reduce risk per trade or the number of simultaneous positions.
Does a drawdown calculator work with a small trade history?
Poorly. Win rate and average result size are unstable below roughly a hundred trades, so the simulation inherits that uncertainty. Use it as a rough guide at that stage, and re-run it as the sample grows.
Why do my three open trades count as more risk than the calculator says?
Because streak maths assumes trades are independent. Positions in correlated instruments, such as several pairs against the same currency, move together and behave like one larger position, so the real exposure is higher than the sum of three separate risk slots suggests.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.