Avoiding the big loss — the only rule that has to hold
Ask someone who blew up an account how it happened and you will rarely hear a story about a bad strategy. You will hear about one afternoon. The chart was obvious, the trade went the wrong way, and somewhere in the middle of it the stop order stopped being a stop order. Everything that came before — the plan, the entry work, the position-size arithmetic — was undone by a decision that took four seconds and felt, at the time, like patience. This lesson is about why that decision is the one place where the usual trade-off between discipline and flexibility does not apply, and what the bill actually comes to when you measure it.
The climb is made of many small decisions you controlled. The drop is one decision you stopped controlling. The dashed line is the bill — and it is longer than the climb that came before it.
KEY TAKEAWAYS
- A big loss is one unsized position, not a losing streak. Ten 1% losses in a row cost about 9.6% and every one of them was chosen in advance. A big loss is a number nobody chose.
- Other rules stretch; this one snaps. Breaking the entry rule on every trade for a whole year — 200 violations — costs 25.8% of capital. That is what one 26% loss costs in an afternoon.
- Recovery is convex, and it is paid in trades. A 25% loss needs 152 clean trades to undo; a 50% loss needs 367 — about 21 months at four trades a week.
- Cancelling a stop usually works, and that is the trap. In 200,000 simulated paths it escaped back to even 87.7% of the time at an average cost of 1.17% against 1.00% for taking the stop — but by the tenth use there is a 69.2% chance it has already blown up once.
- Shrinking the position does not make a moved stop legal. Discipline has three layers: capital, knowledge, consistency. Keeping the 1% while moving the stop breaks the other two.
- The one-question test: is this stop move coming from analysis or from hope? Analysis names a level on the chart. Hope names a distance.
What counts as a “big loss”, exactly?
A big loss is a loss from one position that is larger than the cap you set for that position before you opened it. The defining feature is not the size of the number — it is that the number was never chosen. A ten-loss streak at 1% a trade costs you about 9.6% and every single one of those losses was a decision you made in advance. A single trade that ends at −30% was not decided by anyone.
That distinction matters because the two feel identical while they are happening and behave nothing alike afterwards. Lesson 40 already showed that a system winning 40% of its trades produces a nine-loss streak as its median outcome — streaks are normal, priced in, and recoverable. A big loss is not a longer streak. It is a different object: a bounded risk that had its bound removed mid-flight.
Two tests tell them apart. Was the amount known before entry? If yes, it is a streak, however painful. Could it have been worse if price had kept going? If yes, it was unbounded — and unbounded is the only category this lesson is about.
Why is this called the only rule that has to hold?
Because the other rules are elastic. Break them and you lose a percentage of your edge; break this one and you can lose the account that the edge was supposed to compound. It is worth putting numbers on that, because “be disciplined” is advice nobody has ever acted on.
Take the system this site has used since Lesson 39: 40% of trades win, winners are worth twice the losers, 1% of capital risked per trade. Now imagine a trader who breaks the entry rule — takes the mediocre setups, buys at A instead of waiting for A′ as Lesson 51 describes — and does it on every trade for a full year. Say that drags the win rate from 40% down to 35%. Over 200 trades:
| One year, 200 trades | Ends with | Cost of the sloppiness |
|---|---|---|
| Entry rule kept (40% wins) | 1.46× starting capital | — |
| Entry rule broken every trade (35% wins) | 1.08× starting capital | 25.8% of capital |
Read that again, because it is the centre of the lesson. An entire year of doing the entry badly — two hundred separate violations — costs about what one 26% loss costs in a single afternoon. The entry rule stretches. You can abuse it for months and the damage arrives as a slope, slowly, with plenty of time to notice and correct. The stop rule does not stretch. It holds or it fails, and the failure arrives whole.
Four of the five bend. Break an elastic rule and you pay a slope you can see coming; break the rigid one and there is nothing to absorb it.
What does one big loss actually cost in work?
Percentages hide the answer because losses and gains are not symmetric. Lose 50% and you do not need 50% to get back — you need 100%, because you are now earning it on half the capital. The useful way to feel this is to convert the hole into the thing you actually spend: trades.
On the system above, each trade adds an average of 0.19% to capital in log terms. That is the exchange rate. Divide the hole by it and you get the number of flawless trades — not attempts, trades from a system running at its proper win rate — needed just to return to where you were standing before.
| Single-trade loss | Gain required to recover | Clean trades needed | At four trades a week |
|---|---|---|---|
| 10% | +11.1% | 56 | 3 months |
| 25% | +33.3% | 152 | 9 months |
| 40% | +66.7% | 270 | 16 months |
| 50% | +100% | 367 | 21 months |
| 75% | +300% | 733 | 3.5 years |
The column that matters is the last one. A 25% loss is not “a quarter of the account” — it is nine months of doing everything right, and doing everything right for nine months is not a thing most people manage even once. That is the real exchange rate between an unstopped trade and a career.
If it is obviously fatal, why does anyone remove a stop?
Here is the part that almost nobody explains, and it is the reason the habit is so hard to kill: removing the stop usually works. Not sometimes — usually. The trader who cancels a stop and waits is overwhelmingly likely to be rewarded for it, which is exactly how a behaviour gets learned.
We simulated it. Take a long position that is already 1% offside at the moment the stop is cancelled, and let price wander with no drift at all — a coin-flip walk of 0.4% a candle, 500 candles, 200,000 paths, seed 20260911. The trader closes flat if price ever returns to the entry, and capitulates if it reaches 10% against. Here is what happens:
| Outcome | Frequency | What it feels like |
|---|---|---|
| Price returns to entry, closed flat | 87.7% (median 15 candles) | “See? I was right to wait.” |
| Capitulates at −10% of price | 11.1% | At 5× notional, half the account |
| Still holding at candle 500 | 1.2% | Capital locked in a position, doing nothing |
Average cost of cancelling: a 1.17% adverse price move. Average cost of simply taking the stop: 1.00%. On average the rule-breaker is barely worse off. That is not a defence of the habit — it is the explanation for it. The feedback loop pays out nine times out of ten, and the average barely flinches, so nothing in ordinary experience ever tells you to stop.
The bill is in the shape, not the average. The expected number of times you can do this before the first failure is nine. By the tenth use, the chance it has already blown up once is 69.2%; by the twentieth, 90.5%. A habit that works 87.7% of the time and ends you 11.1% of the time is not a coin-flip you get to keep flipping.
Important limit on that 87.7%. The walk in this simulation has no trend — it is the friendliest market that can exist for someone holding a loser. Real markets trend, and a position offside in a trending market comes back less often than this. Treat 87.7% as the best case, not the forecast.
Is moving the stop different from cancelling it?
It is more dangerous, because it borrows the vocabulary of discipline. The trader who cancels a stop knows they broke a rule. The trader who moves it down and shrinks the position to keep total risk at 1% can say, with a clean conscience, that they are still following money management. The maths even agrees with them. And the account still dies.
Our course handles this by splitting discipline into three layers instead of one:
Layer one is the number: risk stays inside your fixed percentage. Layer two is the location: the stop sits where market structure says the idea is wrong, which is what Lesson 43 spends its whole length establishing. Layer three is time: the rule you are following at minute sixty is the rule you were following at minute one.
Shrinking the position to fund a wider stop keeps layer one and breaks layers two and three. The new stop is no longer at a level that means anything — it is at a level your wallet can tolerate, which the market has never heard of. And you have changed the rules of a trade while the trade was live, which means from now on your rules are negotiable and both you and the market know it.
There is a one-question test for this, and it is brutally effective because it is answerable honestly in about two seconds: is this move coming from analysis, or from hope? Analysis produces a specific level and a reason — a support band, a swing low, an invalidation price that has printed on the chart since you entered. Hope produces a distance. If the answer to “why there?” is a number of dollars or percent rather than a thing you can point at on the chart, the move is hope wearing a risk-management costume.
Where do big losses actually come from, structurally?
They cluster. They are not evenly sprinkled across the year — they concentrate in one recognisable market condition, and knowing the condition is worth more than any amount of willpower.
The condition is a move that runs in a straight line with no pullback, where every timeframe you look at is pointing the same way at the same time. Our course calls this the risk wave, and identifies it as the single most common cause of a destroyed account. The logic is precise: a trend that is genuinely tradeable arrives in sequence — the small frames turn first, then the next one up, then the one above that, so there is always a pullback where a stop has somewhere to live. When every frame moves together instead, the market has given you no pullback, which means no structure below your entry, which means nowhere to put a stop.
And that is exactly when it looks most obvious. A clean unbroken run is the most persuasive thing a chart can show you, and people respond to persuasion by adding leverage. So the two conditions arrive at the same moment: maximum confidence and zero stop location. The trades opened there are the ones that later get described as “the market did something insane”.
The rule that comes out of it is uncomfortable but short: if there is no pullback on any timeframe, there is no entry on any timeframe. Not a smaller position — no position. Lesson 21 gives you the mechanics for checking whether the frames are moving in sequence or all at once.
So what do you do instead when price goes against you?
You take the small stop, and then you look again. That second half is the part people skip, and skipping it is why the first half feels unbearable.
A stop being hit is not a verdict on the idea. It is a verdict on this attempt at the idea. If the higher-timeframe structure that made you interested is still intact after the stop, the correct response is to watch the level below, wait for the same conditions you always wait for, and take the trade again with a fresh stop. Two 1% losses and a winner is a perfectly normal, profitable sequence. One refusal to accept a 1% loss is not a sequence at all — it is an open-ended commitment.
Worth saying plainly: re-entering is not what Lesson 44 means by scaling in. Scaling in means adding while the trade is winning, after the stop has moved up to a new structural level, with total risk recalculated against that new stop. Buying more of a losing position does the opposite — it raises exposure while the evidence is getting worse, and it lowers the price at which you are wiped out. Same hand movement, opposite trade.
The account-level stop
Per-trade rules protect you from one trade. They do nothing about the afternoon where you take six trades in a row because the first three hurt. So write one sentence, before the session, in this form:
“If the account is down ___% today, I close the platform, whatever the chart is doing.”
Three per cent is a common answer for someone risking 1% a trade — it is three ordinary losses, which is an ordinary day, not a disaster. The point is not the number. The point is that the number exists, in writing, before the emotion that will argue against it exists. This is the one line of your trading plan that has no discretion in it.
PRACTICE CORNER
Everything above depends on one mechanical thing: the stop being a real resting order on the exchange from the same moment the position opens, not a level you are watching. An order sitting on the book does not negotiate with you at 2am. If you want to check that your platform lets you attach the stop to the entry ticket itself, these are the exchanges we use:
Affiliate links — we may earn a commission at no cost to you. Disclosure · Education only, not financial advice.
What people get wrong about the big loss
- Treating it as a discipline problem. It is usually a structure problem that became a discipline problem. Most unstopped trades were entered where no stop location existed — in the straight-line move with no pullback. Fix where you enter and most of the willpower question disappears.
- Believing the average. Cancelling a stop costs 1.17% on average against 1.00% for taking it. The average is the least informative number in the whole table, because you do not experience averages — you experience one path, and 11.1% of paths end the account.
- Counting a shrunken position as compliance. Keeping risk at 1% while moving the stop to a level with no structural meaning breaks two of the three layers. The percentage was never the point; the location was.
- Thinking a bigger account makes it safer. The recovery table is in percentages, so it is scale-free. A 50% loss costs 367 clean trades whether the account is $2,000 or $2,000,000.
- Confusing averaging down with scaling in. Adding to a winner after the stop has moved up is a recalculated risk. Adding to a loser is an uncapped one. Lesson 44 is about the first.
When this lesson is wrong
Three honest limits.
The simulation is a friendly market. A driftless random walk is the best possible world for someone holding a losing position. In a trending market the 87.7% escape rate would be lower and the tail fatter. The number is a floor on the danger, not a description of it.
Spot positions in assets you intend to hold for years are a different problem. If you bought with no leverage, no stop, and a multi-year thesis, then price going against you is not an unstopped trade — it is a drawdown inside an investment. The rule in this lesson governs trades: positions opened for a defined move, with a defined invalidation. Do not let the two blur into each other, because the blurring is itself a common way that a trade quietly becomes a “long-term investment” at exactly the wrong moment.
Stops can be gapped through. A stop is a promise to try, not a guarantee of price. In a fast market your stop can fill well below its level, which is a real argument for smaller size — and no argument at all for not having one. The alternative to an imperfect stop is not a perfect stop; it is an unlimited loss.
Frequently asked questions
Does a mental stop count?
Only if it has never once failed to execute, which is a claim almost nobody can make honestly. The whole difficulty this lesson describes is that the decision to exit gets made by a person who is currently losing money, and that person is not the one who wrote the plan. A resting order removes the decision from the moment it is hardest to make. If you use mental stops, at least test them: for twenty trades, write the level down in advance and then record whether you actually acted on it.
How is a big loss different from a drawdown?
A drawdown is the sum of many sized losses; a big loss is one unsized loss. They can produce the same number on the screen and they carry completely different information. A 12% drawdown from twelve 1% losses tells you the system is in a normal losing stretch. A 12% drawdown from one trade tells you nothing about the system and everything about the process, because that trade was never part of the system in the first place.
If removing the stop works 87.7% of the time, is there a size small enough to make it worth it?
The question contains its own answer: if the position is small enough that a 10% adverse move does not matter, then the 1% stop loss did not matter either, and there was nothing to escape. The habit only produces relief when the position is big enough for the loss to hurt, which is exactly the size at which the 11.1% tail is unaffordable. The two conditions cannot be satisfied at once.
What should I do if I am already holding a position that is far underwater with no stop?
Treat it as a position-sizing question rather than a prediction question. Decide the maximum further loss you are willing to take from here, in money, and place a real order at the level that produces it. That replaces an open-ended commitment with a bounded one, which is the only thing you actually control. Whether the position recovers is not something anyone can tell you, and a lesson that pretended otherwise would be lying to you.
Next: Setting goals in trading — daily and weekly targets, and the measures that quietly push you into the behaviour this lesson just warned you about.