Every trader has been told that with a positive edge, results converge over time. That is true, and it is also the least useful true thing in trading, because it says nothing about the part that actually decides outcomes: the order the trades arrive in.
So we removed every other variable. One fixed set of 200 trades — 80 winners at +2.5R, 120 losers at −1R, 0.05R of cost on every one. Expectancy +0.35R, a genuinely strong system. Not simulated trades: the same 200 results, every time. The only thing that changes is the shuffle. 60,000 shuffles, 2% risk per trade, $10,000 to start.
Every ordering ends at exactly the same number
Start with the reassuring part, because it is real. Multiplication does not care about order, so a fixed-fractional account running the same trade set always finishes in the same place: $35,971.79, every single time, across all 60,000 shuffles. Not approximately. Exactly.
If you could guarantee you would take all 200 trades at the same risk fraction and never deviate, the sequence would be irrelevant and this article would end here.
Now look at how you get there
Same 200 trades, same risk, same destination. Only the order differs. These are the mildest, the median and the harshest orderings found in the 60,000 shuffles. Note what the harshest one looks like: the losses cluster early, the account is more than halved from its own high before trade 60, and it still finishes at the same number as the other two.
| Measure | Mildest ordering | Median ordering | Harshest ordering |
|---|---|---|---|
| Final equity | $35,972 | $35,972 | $35,972 |
| Maximum drawdown | 10.5% | 20.8% | 54.4% |
| Deepest point, as a fraction of the peak before it | −10.1% | −21.2% | −51.4% |
Across all 50,000 orderings the 5th percentile of maximum drawdown is 15.2% and the 95th is 31.7%; the two extremes above are the single mildest and single harshest orderings found.
The account that ended at $35,972 having never been more than 10.5% underwater and the account that ended at $35,972 after being cut in half held identical positions and identical results. They differ only in what order the envelope was opened.
The number that matters: would you still be there?
Final equity is a fact about the trade set. Drawdown is a fact about the human being. Almost every trader has an explicit or implicit line — a prop firm limit, a personal rule, a spouse, or just the point at which conviction fails — below which they stop.
One trade set with +0.35R expectancy, 2% risk per trade, 200 trades.
Read that as a survival table. If your stop-trading line is −25%, then 23.7% of the orderings of a genuinely profitable system take you out — not because the system failed, but because the losses happened to cluster early. At −20%, more than half of them do. And 95.8% of orderings visit a 15% drawdown at some point, which means a trader who panics at 15% is not really trading this system at all.
The same edge, the same trades, the same risk. Somewhere between one in four and one in two versions of you does not finish the experiment.
The statistic nobody quotes
Here is the one that explains why profitable trading feels so bad. In these simulations the account spends a median 78% of its life below its own previous high. Not 20%. Not half. More than three quarters of every trading day, a system compounding to 3.6× is showing you a number worse than one you have already seen.
The longest single stretch below the previous peak, in the median ordering, was 38 consecutive trades. At the 95th percentile it was 76 trades, and the worst run in sixty thousand shuffles was 141 out of 200 — nearly three quarters of the entire track record spent below a high-water mark that was eventually beaten by a mile.
Being below your peak is not the exception in a winning system. It is the normal state, and new equity highs are the exception. If your emotional model of a good system is one that mostly sits at new highs, you are measuring your real system against something that has never existed.
| Checkpoint | Median | 5th percentile | 95th percentile |
|---|---|---|---|
| Equity at trade 50 | $13,772 | $9,750 | $19,452 |
| Equity at trade 100 | $18,966 | $12,532 | $28,704 |
| Lowest point ever visited | $9,584 | $8,175 | — |
Note the 5th percentile at trade 50: below where you started, fifty trades into a system with +0.35R expectancy. And 78.7% of all orderings dipped below the starting equity at some point — four books out of five in a system that triples the account by trade 200.
What you can actually do about it
Sequence risk cannot be removed. It can be made survivable, and the levers are unglamorous.
Size for the unlucky ordering, not the median one. Halving the risk to 1% roughly halves every drawdown number above, and the destination becomes $19,537 instead of $35,972 — slower, and reachable from far more starting sequences. The question is never “what does this system return?” It is “what risk level makes the 95th-percentile path something I will actually sit through?”
Set the stop-trading line before you need it, and set it wide. A rule at −30% eliminates 7.4% of orderings. A rule at −20% eliminates 58%. If you must have a line — and a prop account gives you one whether you want it or not — it has to sit outside the ordinary noise of your own system, which means you need to know what that noise looks like first.
Judge the process, not the position of the curve. Two traders running the identical system, starting three months apart, will have completely different opinions of it. Neither opinion is information. What is information: did you take the trades your rules defined, at the size your rules defined? That is the only question whose answer is not being generated by a shuffle.
What the lab says
- The destination is fixed; the journey is not. Every ordering of the same trades ends at the same equity and produces drawdowns between 10.1% and 51.4%.
- Drawdown is where sequence risk gets paid. A −25% stop-trading line ends the run in 23.7% of orderings of a profitable system.
- Below the peak is the normal state. The median account spent 78% of its life under its own previous high, with runs of 38 trades typical and 141 possible.
- Size is the only real defence. Halving risk halves the drawdowns and turns a large fraction of unsurvivable sequences into survivable ones.
- A losing first fifty trades proves nothing. About one ordering in fifteen was below its starting equity at trade 50 — with a +0.35R edge running perfectly.