A thousand versions of your next 200 trades

Same edge, same trades, different order. The spread between best and worst is luck — and luck is what you must survive.

Imagine two traders with an identical system: the same win rate, the same reward ratio, the same two hundred trades. One finishes the year up 60%. The other is down 30% and has quit. Nothing separates them except the order in which the wins and losses arrived.

What a simulation actually does

It takes your edge — win rate, reward-to-risk, risk per trade — and plays out the next N trades a thousand times, shuffling the order randomly each run. The result is not one outcome but a distribution: what typically happens, what happens in the unlucky 5%, and how often the account is halved along the way.

The three numbers to read

1. Median outcome. The middle result. Ignore the average, which a few lucky runs distort upward.
2. The 5th percentile. Your realistic bad year. If this number is unbearable, your risk per trade is too high — regardless of how good the median looks.
3. Typical worst drawdown. The fall you should expect to live through, not the one you hope to avoid.

Streaks are normal, not evidence

At a 45% win rate, a run of eight consecutive losses is entirely ordinary across two hundred trades. It says nothing about your system. But it arrives feeling like proof that something is broken, which is precisely when people abandon a working method or double their size to recover — and the simulation shows exactly how expensive both responses are.

How to use it properly

Take your real numbers from the journal, not the ones you hope for. Twenty honest trades give a rough win rate and reward ratio; feed those in. Then adjust risk per trade until the 5th percentile is a result you could actually tolerate without changing your behaviour.

That is the whole exercise. Not predicting your return — sizing so that a bad but normal run cannot remove you from the game.

Reading the distribution instead of the average

The average outcome across a thousand simulations is misleading, because a handful of extraordinary runs pull it upward. What matters is the shape of the distribution.

Size your risk so the fifth percentile is survivable. If a normal bad run would make you abandon the strategy, the strategy is too large regardless of how good the median looks.

Why raising risk stops helping

Intuition says doubling risk per trade should roughly double returns. It does not, and the reason is compounding against a moving base.

A 50% loss requires a 100% gain to recover. As risk per trade rises, drawdowns deepen, and each deep drawdown means subsequent gains are calculated on a smaller balance. Past a certain point — usually far lower than people expect — additional risk reduces the median outcome while continuing to increase the frequency of ruin.

Losing streaks are ordinary

At a 45% win rate, a run of eight consecutive losses across two hundred trades is unremarkable. At 40%, longer runs appear regularly.

Knowing this in advance changes behaviour at the moment it matters. A streak arrives feeling like proof that something broke, and that feeling produces two expensive responses: abandoning a working method, or doubling size to recover. Both are attempts to escape ordinary variance.

The simulation lets you see the streaks before you live through them. Traders who have watched a thousand paths tend to react to the ninth loss with far less drama than those who have not.

Getting honest inputs

The output is only as good as what you feed it, and the temptation is to enter what you hope rather than what happened.

1. Take the win rate from your journal, not from memory. Memory systematically overweights wins.
2. Use your realised reward ratio, which is usually lower than your target because winners are often cut early.
3. Include costs. A 2R target becomes closer to 1.8R after fees, funding and slippage.
4. Use at least twenty trades before trusting the numbers, and re-run it as the sample grows.

If those honest inputs produce negative expectancy, no position size fixes it. The rule itself must change — and finding that out through simulation is enormously cheaper than finding it out through two hundred real trades.

What the model deliberately ignores

• Correlation between trades. Real losses cluster, because bad conditions produce several bad trades at once. Real drawdowns are therefore usually worse than the simulation suggests.
• Changing edge. The model assumes your win rate is constant. Edges decay as conditions change.
• Your behaviour. It assumes you follow the rules through the entire drawdown, which is the assumption most often violated.

Each of these means reality is somewhat harsher than the output. Treat the simulation as an optimistic floor, not a forecast — and size accordingly.

Common questions

How many simulations are enough?

A thousand is plenty for stable percentiles. More runs refine the extremes slightly and change the practical conclusion not at all.

My median outcome is negative — what now?

Then the edge is negative and position sizing cannot rescue it. Either the entry criteria, the reward ratio, or the market being traded has to change. This is uncomfortable and far cheaper to learn here than in the account.

Does this predict my actual return?

No, and treating it as a prediction is the main way it gets misused. It describes the range of outcomes a given edge and risk level can produce. Its value is in the spread, not the centre.