Look at the figure — the same chart, the same trade, two traders, two stops.
Trader one places a stop at ₹490 — a nice, round, comfortable-looking number, roughly 2% below entry. It feels disciplined. It feels safe.
Trader two places a stop just below the recent swing low — the structural point on the chart where, if price actually traded there, the entire reason for taking this trade would be proven false.
The stock dips. Ordinary, unremarkable noise — the kind every chart produces daily. Trader one's comfort stop, sitting inside that normal noise, gets clipped. Out, at a small loss, moments before the stock does exactly what the original trade expected. Trader two's structure stop, sitting beyond the noise, survives the wobble untouched — because the wobble never actually threatened the reason the trade was taken.
Same chart. Same entry. Same instrument. One stop was placed by feeling. One was placed by evidence. Only one of them got to find out if the trade was right.
This is the idea underneath every good stop-loss, and it's simpler than any indicator: a stop should sit at the exact price where your reason for being in the trade is proven wrong — never at the price that merely feels uncomfortable to lose.
Feel-based stops fail in one of two directions, and both cost real money:
Too tight, and you place it inside normal noise — a round number, a fixed percentage chosen without reference to the chart, 'just past my entry'. The Technical Science school taught you that every level is a zone with a smear of activity around it, not a laser-precise line; a stop placed inside that smear gets clipped by the market simply breathing. You're not managing risk. You're donating small losses to ordinary volatility, over and over, on trades that would otherwise have worked.
Too loose, and you're not really managing risk at all — 'I'll give it room', with no defined point, hoping instead of planning. This isn't structure either. It's the absence of a decision, dressed up as patience.
The professional answer sits between them, and it comes straight from the Technical Science school you may already know: place the stop just beyond the nearest real structure — below the last swing low for a long trade, above the last swing high for a short, beyond the zone where crowd memory actually lives. Not because that specific price is magic, but because it's the point where the story of the trade — 'this level holds', 'this trend continues' — is actually falsified by the market's own evidence, rather than by your nerves.
Now, the mechanics — because a well-placed stop is worthless if it isn't defended properly, and this is where execution enters risk management for the first time.
The mental stop is not a stop. "I'll sell if it hits ₹480" with no order actually placed is a promise made to a version of you that won't exist in the moment — the same version Chapter 6 of the Psychology school already warned you about. In a fast-moving market, at the exact price where fear peaks, mental stops evaporate. They become 'let me just wait for a bounce.' A stop that lives only in your head defends nothing.
The SL order (stop-loss, limit) triggers a limit order once your stop price is touched. It guarantees the price, if filled — but in a fast, gapping market, it can fail to fill at all, leaving you holding a falling position with an order sitting uselessly behind the price.
The SL-M order (stop-loss, market) triggers a market order once touched — it guarantees the exit, not the price. In volatile stocks, thin option strikes, or anything that can gap hard, SL-M is usually the safer choice: a worse fill beats no fill. Know which one you're using, and know why — 'I always use SL' is not a strategy, it's a habit that hasn't met a bad day yet.
The trailing stop is structure's stop, kept moving: as a trade proves itself right, the stop follows — locking in progress without capping how far a genuine trend can run. It's the mechanical answer to the greed-versus-fear tug-of-war the Psychology school described: you don't have to decide in the moment whether to hold or sell. The rule already decided for you, in advance.
One last, uncomfortable truth to close the chapter: a stop being hit is not proof you were wrong to take the trade. A well-placed stop, hit occasionally, is the cost of admission for every strategy that has real edge — Chapter 4 will show you exactly why a system can lose most of its trades and still be excellent. The goal was never a stop that's never hit. The goal is a stop that only gets hit when the evidence genuinely changed — which is, not coincidentally, the exact number Chapter 2's formula needs to calculate your size correctly. The two chapters are one calculation, split in half.
Key Takeaway
A stop belongs at the price where your reason for the trade is proven wrong — not at a round number, not at 'wherever feels safe'. Place it just beyond real chart structure. A mental stop is not a stop; use SL for guaranteed price, SL-M for guaranteed exit in volatile instruments; trailing stops mechanise the hold-vs-sell decision in advance. A stop being hit occasionally is the cost of having a real edge, not proof you were wrong.
Think About It
Look back at your last stop-out. Was the stop placed at a point where the market actually told you something — a structural level breaking — or was it placed at a number that simply felt like the right amount to lose?
Risk Lab — Re-Stop Your Last Five Losers
Pull up your last five losing trades. For each, mark on the chart where you ACTUALLY placed your stop, and where the nearest real structure (swing low/high, support/resistance zone) actually was.
Note the gap between the two. Were your stops mostly comfort stops (round numbers, fixed %, inside the noise) or structure stops?
For each trade, ask: if the stop had been placed at the structural point instead, would the trade have survived the noise and gone on to work — or would it have failed anyway, just later and for more money?
Write one rule for yourself: 'From now on, my stop goes at ___, and I defend it with a [SL / SL-M] order because ___.' That sentence is Chapter 2's missing input, filled in properly.