“Where do I put my stop?” is the most-asked practical question in trading, and it is almost always asked backwards. The usual version is how much am I willing to lose on this trade? — and that question has nothing to do with the chart, so any answer it produces is a number invented by the trader and imposed on a market that has never heard of it.
The stop is not a budget. It is the price at which the reason you took the trade is no longer true. Everything else — how many contracts, how much money — follows from it, in that order.
The order of operations, which is the whole discipline
There is a correct sequence and it is short:
- Find where the idea is wrong. A level, a structure, a condition. This comes from the chart and has nothing to do with your account.
- Measure the distance from your intended entry to that point.
- Decide what you are willing to risk as a share of the account — a decision made once, in advance, for all trades.
- Divide. The contract count is what comes out.
Run it the other way — pick a size, then place a stop where the loss feels tolerable — and you have set your invalidation level by reference to your own comfort. The market will find that level with no difficulty at all, because it is not a level; it is an arbitrary distance from an arbitrary entry.
The arithmetic of step four is what the position size calculator exists for. Step one and two are what the rest of this guide is about.
Three families of stop, and only two of them are defensible
Structural. Behind the thing that would have to break for the idea to be dead: the far side of a level, the origin of a move, the boundary of a range. This is the strongest kind because it is falsifiable — there is an event that disproves the trade, and the stop sits behind it.
Volatility-based. A multiple of a measure of normal movement — most commonly Average True Range. This does not identify where the idea is wrong; it identifies where the market’s ordinary noise ends. Useful on its own in systematic trading, and useful alongside structure as a sanity check: if your structural stop sits inside one ATR, it is inside the noise and will be hit by nothing in particular.
Fixed-money. “I risk $200 a trade, so my stop is $200 away.” This is not a stop placement method. It is a position size masquerading as one, and it is the single most common reason traders conclude that “stops always get hit.”
The stop is hit because of where it is, not because stops are a bad idea
Illustrative. The distance between the amber line and the teal line is not extra risk — at a fixed risk budget it is fewer contracts.
Every trader who has watched price wick through their stop by two ticks and then run to target has drawn the same conclusion: the market is hunting me. The mechanism is duller and more useful than that.
Large orders need someone on the other side. Resting stop orders are, by construction, a pool of guaranteed counterparties sitting at a known price. When size needs to fill, the obvious level is where the filling is easiest. Nobody is looking at your account. They are looking at liquidity, and you put yours where everyone else put theirs.
Which gives a rule with no mysticism in it: if a level is obvious enough that you can see it instantly, it is obvious enough that the stops are already there. The stop belongs behind that pool, not in it. That extra distance is not extra risk — at a fixed risk budget it is simply fewer contracts, and the stop distance calculator will do that trade-off for you in both directions.
How wide is wide enough?
The honest test is volatility. If your stop distance is smaller than the instrument’s normal movement over your holding period, it will be hit by movement that carries no information at all.
A serviceable starting frame, to be calibrated per instrument rather than adopted as gospel: under 1× ATR is inside the noise; 1–2× is normal territory for an intraday structural stop; beyond about 3× you are usually no longer trading the idea you thought you were. ATR is a measure of range, not a prediction, and it changes with regime — so recompute it rather than remembering it.
What this frame cannot do is tell you the trade is good. A stop outside the noise on a bad idea is still a bad trade, taken more patiently.
The order type matters more than people expect
- A stop order is a market order in waiting — and on CME Globex, a bounded one. When touched it becomes a market order capped by protection points, fills at every level inside that band, and any unfilled remainder rests in the book as a limit order at the protection price. So it is neither “whatever is available” nor a guarantee: it usually gets you out, usually a little worse than the stop price, and in genuinely violent conditions it can leave a residual working.
- A stop-limit order can leave you in the trade. It converts to a limit order, so if price runs past your limit it simply does not fill, and you still hold the position with no protection. Traders use these to avoid slippage and occasionally discover they have avoided the exit instead.
- A bracket (OCO) submitted at entry puts both exits in the market at the moment you are calm, and fills one while cancelling the other. It removes one discretionary decision from the moment you are least equipped to make it.
- A mental stop is not a stop. It is an intention, held by the person least equipped to execute it at the moment it becomes necessary.
When a stop cannot protect you
Being honest about this matters more than the placement rules.
Gaps. A stop is an instruction to trade at a price. If the market reopens beyond it, the instruction executes at the first available price, which can be far away. Weekends, scheduled releases and overnight sessions are all gap risk, and no stop placement removes it — only size does.
Limit-locked markets. When a contract is limit up or down, trading is restricted at the limit price and your stop becomes an order in a queue that may not clear. In March 1980 COMEX held the back-month silver futures inside a $1-a-day price limit while the spot month fell unrestrained — by 27 March the two prices were $13.26 apart. A stop resting in a limit-locked contract was an instruction nobody could fill.
Liquidity holes. Overnight, in thin sessions, or in an illiquid contract, the book between your stop and the next resting bid can be nearly empty. The fill is not the price you saw.
All three point the same way: the stop defines your risk under normal conditions, and position size defines it under abnormal ones. Sizing so that a two-stop-width gap is survivable is what keeps a gap from being an account event rather than a bad day.
Moving a stop
The rule taught in Block A is that after entry, the stop moves only in the trade’s favour — and that the absence of exceptions is the feature, because widening a stop mid-trade converts a defined risk into an undefined one at the exact moment judgment is least reliable. It is a rule worth automating rather than remembering.
Moving it toward the trade is legitimate but not free. Trailing a stop, or moving it to breakeven, is a real change to your expectancy and should be priced rather than assumed. We tested the popular version: moving to breakeven at +1R cut losing trades from 70% to 47% and cut expectancy by 44%, while improving the median drawdown by barely a point. The full working, and the condition under which the rule does pay, is in the lab experiment.
How the Conflux Method uses this
This is the first principle of Block A, and it is stated more bluntly there than anywhere else on this site: it does not take a trade without a stop, and the entry is secondary to where the stop is. The structure comes first — the Reaction Level, a zone where an actual buyer–seller battle happened — and the stop goes behind it. Only then is there a question about entry.
Two further ideas from the method bear directly on this page. Visible levels are the ones where the crowd’s stops sit, and they are named precisely so they can be avoided as stop locations. And from Block C, Balance — the options-derived daily equilibrium built from the straddle half-cost — is used explicitly as something to hide a stop behind, rather than as a target.
Watch the Block A preview lessonWhere to go next
The tool that turns volatility into a stop distance and then into a contract count is the stop distance calculator. For what a stop costs you when you move it, the breakeven experiment. For why widening one is so expensive, the wider-stops myth. And for reading whether a level is actually being defended before you commit, order flow.
The short version
- The stop is where the idea is wrong, not where the loss becomes uncomfortable. One of those is on the chart.
- Size follows the stop. Find invalidation, measure the distance, apply your fixed risk, divide. Never the reverse.
- The obvious level is where the stops already are. Behind it, not in it — and that distance costs contracts, not risk.
- Inside one ATR is inside the noise. A stop there is hit by movement that means nothing.
- Know what your order type actually does. A stop-limit can leave you holding an unprotected position.
- Moving a stop toward the trade is a real change to your expectancy, not a free safety measure — and widening one mid-trade converts a defined risk into an undefined one.