The stock gapped down on a Monday on news he had not seen coming. By early afternoon the broker's risk system had flagged the account, and at 2.40 the message came: bring funds or the position is squared off. He called his brother-in-law, moved ₹2,80,000 by 2.55, and the position stayed open. Over the next three weeks the stock came back and he closed it at a small profit. When he tells the story now — and he tells it often, because it is a good story — it is about nerve: everybody else was panicking, he held, and he was right. It is a completely honest account of what happened and it draws precisely the wrong lesson from it, because the thing that decided the outcome was not his nerve. It was that his brother-in-law answered the phone. What that Monday actually produced was a measurement of how close the account was standing to a wall, delivered free of charge, and it has been filed under the heading of a win.
You take the same wet turn near the market every day. One morning the back wheel goes out, you catch it with a foot down, and you ride on. By evening it has become a story about your reflexes. What actually happened is that the road told you something about itself — the camber, the diesel on the surface, the speed at which that corner stops being fine — and you filed it as information about you. Tomorrow morning you take the turn at the same speed.
A near miss is a reading of the exposure, not a report on the operator. The instinct is to store the ending, because the ending is what you experienced. But the ending was the part decided by things outside the decision, and the exposure is the part that will still be there next Monday.
Why survival makes the risk feel smaller
The inversion at the centre of this lesson is the whole reason it needs a lesson. A near miss is direct evidence that an exposure exists and can be reached. Yet it reliably lowers the felt probability of the very thing it demonstrated, because the mind updates on outcomes and the outcome was fine. Each survived episode makes the next one feel more routine, and the standard drifts — a process students of engineering failure call normalisation of deviance. Nobody ever decides to take a large risk. They take a slightly larger one than last time, it is fine, and the new position becomes the baseline from which the next slightly larger one is measured.
| What happened | How it gets filed | What it actually measured |
|---|---|---|
| A margin call met by borrowing from family within fifteen minutes | "I held my nerve and I was right" | That the account can be reached by an ordinary bad day, and that solvency on that day depended on a third party being free to answer a phone |
| A concentrated holding fell 44% and recovered over eight months | "Conviction. This is why you do not panic-sell" | What a 44% fall in that position does to the household total, and whether anything in those eight months could have forced a sale at the bottom |
| A caller claiming to be from the broker asked for an OTP; you nearly read it out, then stopped | "Almost got me. Anyway" | That the household's defence against account takeover is a moment of alertness during a phone call, which is not a defence |
| An options position expired worthless in the buyer's hands after moving hard against you intraday | "It came back, as they usually do" | The size of the intraday adverse move the position can produce, which is the number the next one should be sized against |
| A large payment reached the wrong account and was recovered after four days of calls | "Sorted it out" | That the household's own process for large transfers has no check between typing the account number and the money leaving, and that getting it back depended on the other side agreeing to send it |
Four quiet years and a dangerous position are the same picture
The reason survival is such weak evidence is arithmetical, and one calculation makes it concrete. Suppose an approach carries a one-in-twenty chance each year of an outcome the household cannot come back from — a wipe-out, a forced sale of the house, a debt that cannot be serviced. That is a genuinely dangerous exposure by any standard. Now ask what it looks like from inside.
How to file one properly, in about four minutes
- Write down the reachable worst state, not the outcome. Not "the stock recovered" but "at 2.40 on Monday, a further 6% down and with nobody answering the phone, the position is squared off at the low and the loss is ₹X". That sentence is the finding.
- Name what actually saved it. Almost always it is something outside the decision: a relative was free, an exchange holiday intervened, a result came out on the right day, you happened to be looking at the screen. If the saving factor is not repeatable and not controlled by you, it is not part of your process — it is weather.
- Ask what the same event costs at your current size. Positions grow. An episode survived at ₹3 lakh is a different event at ₹18 lakh, and the version that nearly happened is the one that scales.
- Check whether the standard moved. What was the rule before this position — a size cap, a no-leverage line, a not-in-a-derivative line — and is the rule you are running now the one you wrote, or the one the last few exceptions left behind?
- Do exactly one thing that would have changed the reachable worst state. Not a resolution to be careful. A change to the arrangement: a smaller size, a reserve held against the position, an alert set at a level, the transfer limit lowered, the mandate that lets a second person act. A near miss that produces only a story has been paid for and not collected.
It came back, and now you are sizing the next one
The position that produced the margin call closed at a small profit three weeks later. The account is intact, the borrowed ₹2,80,000 has been returned, and a similar opportunity has appeared. Nothing has been written down about the Monday.
An approach carries roughly a one-in-twenty chance each year of an outcome the household could not recover from. Someone has used it for four years without incident. What does that record establish?
Somvaar ko stock gap down, 2:40 pe margin call, 2:55 tak jeeju se ₹2,80,000 aa gaye, position bach gayi, teen hafte mein bhaav wapas. Ab kahani yeh hai ki "maine himmat rakhi". Sach yeh hai ki jeeju ne phone utha liya tha. Us din jo naapa gaya woh aapki himmat nahi thi — deewar kitni paas khadi hai, yeh tha. Aur hisaab dekho: agar kisi tareeke mein har saal bees mein se ek baar sab doob jaane ka chance ho, tab bhi chaar saal bina kuch hue nikal jaana sabse aam nateeja hai — kareeb 81%. Isliye chaar shaant saal kuch sabit nahi karte. Jis din bacha, usi din likh lo: "wahaan se sabse bura kya ho sakta tha aur kitne ka" — aur intezaam mein ek cheez badlo, sirf saavdhaani ka vaada mat karo.
- A near miss measures the exposure; the ending measures the weather. Only the first is still true next month.
- Surviving lowers the felt probability of the event it just demonstrated, which is exactly backwards.
- Normalisation of deviance: each tolerated exception becomes the baseline the next one is measured from.
- A 5% annual chance of ruin survives four years about 81% of the time — quiet years cannot tell you anything.
- File a near miss by writing the reachable worst state and changing one thing about the arrangement, not by resolving to be careful.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- normalisation of deviance meaning
- Normalisation of deviance is the process by which a tolerated exception quietly becomes the new standard, so each fresh risk is measured against the last one rather than against the rule you originally wrote. Nobody decides to take a dangerous position; they take one slightly larger than last time, it works out, and that becomes the baseline. The term comes from the study of engineering failures and describes an account that drifts into trouble without any single reckless decision.
- what does it mean to take a risk you cannot come back from
- Risk of ruin is the chance that an approach produces an outcome you cannot come back from — an account wiped out, an asset force-sold, a debt that cannot be serviced — rather than merely a bad year. It behaves differently from ordinary volatility because it is absorbing: once it happens there is no position left to recover with, so no future average repairs it. It is what a near miss reports on, and what a run of quiet years cannot measure at all.
- if there is a 5 percent chance of ruin each year what is the chance of surviving four years
- About 81.5%, which is 0.95 raised to the fourth power. That arithmetic is why a clean record proves so little: four untouched years is the single most likely result of a genuinely dangerous exposure, arriving roughly four times out of five, and by then the approach has usually been described to friends and increased in size. Over ten years the survival chance falls to about 60%, so the unrecoverable outcome shows up somewhere in the decade around 40% of the time.
- a near miss in the market is best treated as a measurement of
- The exposure, not the skill of the person who came through it. A margin call met by borrowing from a relative within fifteen minutes measures how close the account was standing to being squared off, and shows that solvency that day rested on somebody else being free to answer a phone. The ending was decided largely by things outside the decision; the exposure is the part that will still be there next Monday.
- how do I record a near miss in my trading account
- Write down the worst state that was reachable from where you stood, in rupees, instead of what actually happened — not that the stock recovered, but that a further move against you with nobody answering the phone squares the position at the low for a loss of a stated size. Then name what genuinely saved it, ask what the same event would cost at your current position size, and change one thing about the arrangement rather than resolving to be careful. Do it the same day, because within a week the memory has been tidied into a story whose ending looks inevitable.