Where a stop sits on the chart

The two questions that produce a price 

Two traders open the same position on the same instrument at the same moment, and each sets a stop. The first reads the chart, finds the low from which the most recent advance began, and puts the line a little beneath it. The second reads the ticket, decides the position should not cost more than a stated amount of money, divides that amount by what the position is worth per unit of movement, and puts the line at whatever distance the division returns. Both now hold a stop, at different prices, in answer to different questions.

The first line answers a question about the market: at what price is the arrangement of past highs and lows the position was opened against no longer intact. The second answers a question about the account: at what price does a position of this size cost the intended amount. Nothing prevents the two from returning the same number, and nothing arranges for it either. Where they differ, a price arrived at by division describes the ticket rather than anything that happened on the chart. Most of what follows is about the first question, and the second returns below.

Placing relative to a swing point 

The charts module defined a swing low as a bar whose low sits beneath the lows of a stated number of bars on either side, and a swing high as the same comparison on highs. A widely stated convention places the stop for a long position beneath the most recent confirmed swing low, and for a short position above the most recent confirmed swing high. The reasoning offered is that a swing point marks where the last push in the opposite direction was turned back, so price beyond it has done something the position was not opened in expectation of.

Two limits travel with the convention from the lesson that defined the swing. A swing point is a function of the confirmation setting, so the most recent confirmed swing low under a setting of three bars either side is frequently not the price it is under a setting of ten, and the phrase names a different level for two readers who chose differently, neither of whom has made an error. Confirmation also looks forward, so the newest bars carry no confirmed swing at all, and the nearest confirmed one can sit a long way from where price is now.

Placement at exactly the swing price is uncommon. Conventions add a buffer beneath the low, or above the high, on the stated grounds that the swing price is one a great many chart readers can identify from the same chart, so an instruction sitting there sits where the largest number of other instructions is likely to be resting. Some conventions state a fixed increment, some a proportion of the instrument's recent range, and the next lesson in this module treats distances derived from a measure of volatility. Nothing settles which of the three the buffer should be.

Placing relative to a level 

The other structural reference is a support or resistance level, defined in the same module as an area where price has previously changed direction more than once. The convention has the same shape: the stop for a long sits beneath a level price has repeatedly turned up from, the stop for a short above one it has repeatedly turned down from, and the reasoning offered is the same reasoning. The difference is that a level is not a price.

A level is a zone with a measurable width, because the turns that defined it did not agree with each other exactly. That width is not a nuisance to be averaged away before the line is drawn. It is the placement decision. A stop just beyond the near edge sits inside a region price has already crossed in both directions, so it stands in the path of ordinary movement within the zone itself. A stop beyond the far edge stands outside that region, at the cost of a wider distance and, with the risk amount held constant, of a smaller position. Both conventions are held, and the zone cannot arbitrate, because the zone is the thing that is uncertain.

Why a round number is not a structural reference 

Stops are frequently placed at whole figures: at the round number just beneath an entry, or at the next one down. It looks structural and is not. A round number is a property of the numbering system the instrument is quoted in. It sits at the same price on every timeframe and on every chart, whether or not price has ever been near it, and it would be in the same place on a blank chart. A swing low exists because a particular sequence of transactions happened. Nothing about the round figure beneath it is a record of anything.

A long standing description in trading literature holds that resting instructions cluster at round numbers, on the reasoning that people reach for round numbers when asked to name a price. It is a description rather than a measurement, and no figure for it appears on this page. The traditions that cite it draw opposite conclusions: one treats a round number as a reference many participants are watching and therefore a level worth using, and another treats the same observation as the reason to put the line elsewhere, since a price where many instructions are resting takes comparatively little movement to reach. Two opposite conclusions from one observation is a fair sign the observation does not settle the matter.

A narrower point stands whatever anyone concludes about clustering. Placement against a swing or a level produces a distance that follows the chart. Placement at a round figure produces a distance that follows only where the entry happened to fall, so an entry a little above a round number receives a tight stop and an entry a little below the same round number a wide one, with the chart identical in both cases. The distance has been set by the arithmetic of the quote rather than by anything the instrument did.

What changes when the amount is chosen first 

Earlier in this module the calculation ran in one direction. A risk amount and a stop distance were given, the cost of reaching the stop on one unit of size followed from the distance, and the risk amount was divided by that cost. Three inputs, one unknown, and the unknown was the size.

Choosing a stop so that a position of an already decided size costs a preferred amount solves the same equation with a different unknown. The size is given now, the risk amount is given, and the only term left free is the distance. The equation resolves and returns a number of pips or points, which is added to or subtracted from the entry, and a line is drawn at the result. The line is a correct arithmetic answer and a description of nothing.

All three arrangements cost the same amount if their stop is reached, which is the counterintuitive part. Inverting the calculation does not make a trade more expensive. It changes what the level means. In the first the chart nominated the level and the size followed from it. In the other two the size was nominated first and the level landed where the division put it, in one case beyond the structure and in the other well short of it, inside ground the instrument had already covered.

Placement and size are one decision 

With a risk amount fixed, placement and sizing are not two decisions taken in sequence. The distance from the entry to the level enters the denominator of the size calculation, so the moment the level is settled the size is settled, whether or not anybody performs the division. Two consequences of that bind in practice.

The first consequence is in the contract value rows. Two positions costing the same amount at their stops carry contract values ten times apart, and the margin requirement is assessed on the contract value, not on the stop. A structural level close to the entry produces a large position, and a large position can meet the margin requirement as a binding constraint long before it meets anything on the chart. Where the two constraints disagree the smaller size binds.

The second consequence is a floor. Instruments are dealt in stated increments, and a division can return a size beneath the smallest one permitted. Where the nearest structural level is far enough away, the size that expresses the intended risk amount is not a size the instrument can be dealt in at all. Rounding up to the smallest permitted size spends more than the intended amount, which is why the sizing lesson's fourth step exists: it turns the overshoot into a number somebody has looked at.

Why structural placement produces uneven distances 

Structure is where it is. A level is the residue of transactions that have already happened, so the distance between an entry and the nearest one is whatever the chart's history has left there, and it differs between instruments, between timeframes, and between one week and the next on a single chart. A rule that places the stop a fixed distance from the entry produces distances that are all the same and levels that mean something different every time. A rule that places it against structure produces the reverse. Only one of the two can be held constant, and a placement method is a choice about which.

The unevenness propagates. Position size varies inversely with the distance, so a sequence of trades sized this way shows tickets of very different sizes while costing a constant amount at their stops. Reward to risk, described earlier in this module as one distance compared with another, has a denominator that changes on every trade. None of that is a defect. It is what it means for the chart rather than the account to have nominated the level.

What placement does not settle 

A stop is an instruction to close a position once price reaches a specified level. It is not a guarantee of the level at which the closing trade is done, and no care over placement changes that. Where a market gaps over the level, where a session opens away from where it closed, or where price moves through a thin book quickly, the position closes at the first price available beyond it, and the realised cost is larger than the intended one by however far that price sat from the stop. The mechanics are set out in the guide to slippage and gapping, and a later lesson in this module deals with the standard and guaranteed varieties of the order and what each one undertakes.

A second detail is specific to placement. A chart is conventionally drawn from the bid side of the quote. A long position is closed at the bid as well, so its stop is assessed against the series the candles show. A short position is closed at the ask, which sits above the bid by the width of the spread, so its stop is assessed against a series that is not drawn on the screen. A level that looks untouched on the chart can therefore have been reached, and a buffer measured to the last decimal against drawn candles is measured against the wrong series in one of the two directions.

Placement also says nothing about what will happen. A stop beneath a swing low is not a claim that price will not go there, a wider stop is not a safer one in any sense the arithmetic recognises, and no figure appears here for how often levels of any kind are reached, because any such figure depends entirely on the definitions chosen before the counting begins.

Where practitioners disagree 

The largest disagreement is whether structure should govern the level at all. The tradition described here holds that it should, on the grounds that a level read from the chart means the same thing on every trade, so a record of many trades is a record of one method rather than of many improvisations. A well argued opposing tradition holds that the prices a great many chart readers can identify from one chart are precisely the prices at which closing instructions accumulate, and prefers a distance derived from a measure of the instrument's own movement. The next lesson sets out how those distances are computed. Neither tradition can produce evidence that settles the question.

The buffer is a second and narrower dispute. A fixed increment beneath the swing is simple, reproducible and the same on a quiet day as on a violent one, which is the whole of the objection to it. A buffer scaled to a measured range adapts to conditions, at the cost of a second setting whose value changes the answer and which nobody derives from first principles. A third convention treats the swing price itself as the level, on the grounds that a buffer is an admission that the level chosen was not the right one.

Two further arguments are worth naming. The first is whether the exit needs to be a price at all, since a tradition exists of closing a position after a stated number of bars rather than at a level. The objection is direct: an exit with no price attached has no defined cost, so the sizing calculation has no denominator. The second is whether the instruction should rest in the market or be held in mind. Those who hold it in mind argue that an instruction resting at an identifiable level is itself information available to others. Those who rest it argue that an instruction not in the market is an intention rather than an instruction, and that the moment it is most needed is the moment it is hardest to act on, which is the subject of the last lesson in this module.