How the mechanics work, written out in full.

Traders routinely compare pricing by looking at one number, usually the advertised spread, and that single number is the least comparable figure in the whole schedule. A trade carries three distinct costs, quoted in three different units, charged at three different moments. Until all three are expressed in the same unit — your account currency, per position, over your actual holding period — you are not comparing anything.

The first line is the spread: the distance between the bid and the offer at the instant you transact. It is quoted in price units, which means it costs you nothing until you convert it. The conversion is spread multiplied by pip value multiplied by position size. Pip value itself depends on the quote currency and the size of one contract, which is why the same nominal spread on two different instruments is not the same cost. A one-pip spread on a currency pair quoted in the account currency and a one-pip spread on a cross with a third currency in the denominator will not produce the same debit, and no advertised comparison table will tell you that.

The second line is commission. Where a broker quotes raw spreads it usually charges a flat commission per lot, per side. This is the easiest line to reason about because it is already in currency units and it is fixed per unit of volume: it does not widen with volatility and it does not vary with the instrument beyond the published schedule. It is also the line most often left out of informal comparisons, because it does not appear on the chart. A raw-spread model with commission and a zero-commission model with a marked-up spread can be exactly equivalent, cheaper, or more expensive — and which of the three depends entirely on your average spread over your actual trading hours, not on the headline minimum.

The third line is swap, or overnight financing. It applies only if the position is still open at the daily rollover, and unlike the first two it can be a credit rather than a debit depending on the direction you hold and the interest differential between the two currencies involved. Swap is charged per night and it accumulates, which means it is irrelevant to a position closed within the session and dominant for a position held for weeks. Any cost comparison that does not state a holding period has silently assumed one.

The practical method is to build one number and rebuild it whenever your style changes. Take a representative position size for your account. Multiply the typical spread — not the advertised minimum, the spread you actually see during the hours you trade — by pip value by size. Add commission for both sides of the trade, because you will pay it twice. Multiply the swap rate for your direction by the number of nights you typically hold. The sum is your round-trip cost in account currency. Divide it by the position size and you have a cost per unit that you can carry between brokers, instruments and account types without translation errors.

Two consequences fall out of that arithmetic immediately. The first is that cost scales with turnover, not with account size. A strategy that trades twenty times a week pays twenty round trips, and a small change in per-trade cost compounds into a large change in annual expense; a strategy that trades twice a month is dominated by financing instead. The second is that the same headline schedule produces different effective costs for different traders, so any statement that one pricing model is cheaper than another is incomplete unless it names the trading pattern it is cheaper for.

Finally, note when each line is charged. Spread is paid on entry, embedded in the price you receive, which is why a position typically opens showing a small negative. Commission is debited at entry and again at exit. Swap is applied once a day at rollover. Because they arrive at different moments, a statement showing only a net figure hides which of the three is consuming your returns.

The most common sizing error is not aggression, it is using the wrong input. A trader opens the platform, sees available margin, and sizes the position against it — half the free margin, or a fixed number of lots that historically felt comfortable. That method answers the question "what am I permitted to open?" when the question that determines survival is "what does this position cost me if I am wrong?"

Those are different questions with different arithmetic. Margin requirement is a function of leverage and notional value: it is a deposit the broker holds while the position is open, and it tells you nothing about the outcome of the trade. Loss on a stopped-out position is a function of the distance between entry and stop, multiplied by pip value, multiplied by size. Leverage does not appear in that expression at all. Doubling permitted leverage doubles the position you may open; it does not alter what a fifty-pip adverse move costs on a position you already hold.

The correct sequence inverts the usual one. First decide the risk budget: the amount of account equity you are willing to lose on this single position, stated in account currency and fixed as a policy rather than chosen per trade. Second, locate the stop level from the market structure — the price at which the reason for the position no longer holds. Third, measure the distance from entry to that level in pips. Only then compute size: risk budget divided by the product of stop distance and pip value. Position size is the output of the process, never the input.

This ordering has a consequence people resist. If the stop must sit further away because the structure demands it, the position must get smaller. The instinct is the reverse — to keep the familiar size and simply accept more room — but that keeps size constant while risk expands linearly with distance. A stop at eighty pips instead of forty on the same position size is not a more patient trade, it is double the loss.

It also makes volatility a sizing variable rather than a surprise. In a period of elevated ranges, structurally valid stops sit further from entry, so the same risk budget necessarily buys a smaller position. Traders who size from margin headroom experience this backwards: they hold constant size into wider conditions and discover that their per-trade losses have quietly doubled while their process appeared unchanged. Sizing from stop distance adapts automatically, because the input that widened is the input the formula divides by.

Two refinements matter in practice. The first is that pip value is not constant across instruments. The same stop distance in pips on two different symbols produces different account-currency losses, because contract size and quote currency differ. Sizing per instrument, using that instrument’s pip value, is the only way a fixed risk budget stays fixed across a portfolio. The second is that concurrent positions aggregate. Six open positions each risking a small share of equity is a large aggregate exposure if they are correlated — and several major currency pairs sharing the same base or quote currency are effectively one exposure wearing six labels. The risk budget belongs to the book, not only to the ticket.

None of this guarantees the stop fills at the stop price. Gaps and fast markets can execute beyond the level, which is a separate mechanism covered elsewhere in these notes. Stop-distance sizing bounds the loss under normal conditions and makes the abnormal case the exception you plan for, rather than the outcome you discover. That is the whole claim — not that risk is eliminated, but that it becomes a number you chose in advance instead of one the market chose for you.

A raw spread is not a fee schedule. It is the price a liquidity provider charges for standing ready to take the other side of your trade and carry that inventory until it can be offset. The width of that price is compensation for a specific risk: that the market reprices between the moment the provider fills you and the moment it can hedge. Everything about pre-release widening follows from that one sentence.

When a high-impact number is due at a known minute, the probability of a large instantaneous repricing inside that window rises enormously. The provider faces the same schedule you do. Quoting the same size at the same width across the release would mean accepting a materially worse expected outcome on every fill, so providers respond in the two ways available to them: they reduce the size they are willing to show at the best price, and they widen the distance between bid and offer to be compensated for the risk they still carry.

On an aggregated book this compounds. The best bid and best offer come from whichever providers are quoting most aggressively at that moment. As individual providers step back, the aggregate top of book is assembled from progressively less aggressive quotes, so the visible spread widens faster than any single provider’s own quote. The book also becomes thinner in depth, which matters separately: a size that previously filled at one price now walks through several levels.

Two features of the timing surprise people. First, widening typically begins before the release rather than at it — often several minutes ahead, as providers reduce exposure into the event. A trader who plans to enter thirty seconds before a print has usually already lost the tight market they were relying on. Second, normalisation after the release is not instant. The quote may print aggressively within a second, but depth returns over a longer horizon as providers re-establish inventory and confidence in the new level. The visible spread can look normal while the size behind it does not.

That mechanism makes some responses useful and others futile. Futile: expecting pre-release spreads to persist through the window, or treating widening as evidence of manipulation. It is the observable consequence of inventory risk being priced correctly, and it happens across venues simultaneously. Useful: deciding in advance whether you are in the market for the release at all, which is a position-sizing decision made in the calm hours before, not a reaction in the noisy seconds after.

If you do intend to hold through, several mechanical points follow directly. Reduce size so that the wider cost of entry and any adverse fill are proportionate to a risk budget you set deliberately. Recognise that a stop resting inside the widening zone can be triggered by the spread alone — on a long position the bid is what touches your stop, and the bid moves further in a widening market even if the mid does not. Understand that a market order executes against whatever depth exists at that instant, which in a thin window may be several levels away from the price you saw. And note that margin requirements on affected instruments may be raised ahead of known events, so free margin available today is not a commitment about free margin during a rate decision.

The general principle is that a scheduled release changes conditions, not just price. Cost of transacting, reliability of resting orders and capital consumed by the position all move together in the same few seconds. Reading the schedule in advance is what converts that from something that happens to you into something you priced.

A leveraged position is a financing arrangement wearing the costume of a single trade. To hold a notional value substantially larger than your deposit, capital is effectively borrowed for the duration, and holding it past the daily cutoff makes that borrowing explicit as a debit or credit on the account. That line is the swap, sometimes called the rollover or overnight financing charge.

The economics differ slightly by asset class but the shape is constant. In foreign exchange, every position is simultaneously long one currency and short another, so holding it overnight means earning the interest rate of the currency you are long and paying the rate of the currency you are short. The net of those two, adjusted by a broker markup, is the swap. Because the differential can point either way, swap is genuinely two-sided: on one direction of a given pair it is a debit and on the other it can be a credit. On instruments without a currency pair underlying them, the same logic applies against a reference financing rate rather than a second currency.

The timing is mechanical and worth internalising. Swap is applied at a fixed rollover moment once per day, not continuously and not at position close. A position opened and closed within the same session never touches it. A position opened five minutes before rollover and closed ten minutes after is charged a full day. Because the charge is per night rather than per hour, the marginal financing cost of holding is lumpy — an intraday trader can ignore it entirely and a swing trader cannot, and there is no smooth gradient between them.

The triple-swap day is where the calendar intrudes. Spot foreign exchange conventionally settles two business days after the trade date. When a position rolls on a Wednesday, the new value date moves from Friday to the following Monday, because Saturday and Sunday are not settlement days. That single roll therefore carries three days of financing rather than one, and the entire weekend is charged in one debit or credit on Wednesday. For most FX instruments that is the Wednesday roll; instruments with different settlement conventions can carry their triple charge on another weekday, which is precisely why the specification for each instrument states its own rule rather than relying on a general one.

Two practical consequences follow. First, a Wednesday roll on a position you were already ambivalent about is a materially larger event than any other night of the week, and traders holding negative-swap positions frequently discover this after the fact rather than before. Second, holiday calendars shift settlement in ways that add days to a roll unpredictably, so the assumption that the charge is a smooth daily line is wrong at exactly the moments when it is largest.

Swap rates are not constants either. They are derived from prevailing interest rates and funding conditions, which change, and they are typically republished on a rolling basis rather than fixed for the life of a position. A carry that was positive when the position was opened can turn negative during a policy cycle, and a position sized on the assumption of a credit can quietly become one paying a debit. Checking the current rate before entering, rather than relying on the rate you remember, is the entire discipline here.

Where interest-based charging is unacceptable for religious reasons, a swap-free treatment replaces the standard financing line under separately published terms — an option applied to an account rather than a different product. In every case the arithmetic to run before opening a multi-day position is the same: current swap rate for your direction, multiplied by your size, multiplied by expected nights held, plus an extra two nights if the holding period crosses the triple-charge day. That total belongs in the trade’s cost estimate from the beginning, not as a surprise discovered on the statement.

A stop order is widely described as a way to limit a loss to a chosen amount. That description is close enough to be useful and wrong in exactly the situation that matters. A stop is a conditional instruction: when the market trades at or through your level, submit an order. What that order is worth depends on the state of the book at the instant it arrives, and no order type can promise a price that no one is offering.

Under ordinary conditions the distinction is invisible. The book is deep, prices move in small increments, and a triggered stop fills within a tick or so of the level. The distinction becomes visible in exactly three situations: a scheduled high-impact release, an unscheduled shock, and the reopen after a market close. In all three, price does not walk through your level — it steps over it. There is no trade at your price to fill against, so execution happens at the first price where liquidity exists, which may be materially beyond where you asked.

This is why the language matters. Slippage is the difference between the price that triggered your order and the price at which it executed. It is a consequence of the market moving faster than the order can be matched, and it is symmetric in principle: an order can fill better than requested when the move runs in your favour during the same interval. Traders remember negative slippage more vividly, and one reason is selection — the moments people most often submit market orders are the moments the book is thinnest and moving against the crowd.

Weekend gaps are the clearest illustration. A currency market that closes on Friday and reopens on Sunday evening has had two days in which information accumulated with no venue to price it. The reopening quote reflects all of it at once. A stop resting inside that gap did not fail; it was triggered on the first available quote and filled there, which may be far from the level. The same mechanism applies over public holidays and around any suspension in an underlying instrument.

There are two things worth knowing about order types here. A stop-loss triggers and then executes at the best available price, which caps nothing in a gap. A stop-limit adds a price boundary — it will not fill worse than your limit — but that protection has a cost: if the market gaps beyond the limit, the order simply does not execute and you remain in the position while it continues to move. Neither is strictly safer. One accepts an uncertain price to guarantee an exit; the other accepts an uncertain exit to guarantee a price. Choosing between them is choosing which failure mode you prefer.

The sizing consequence is the practical one. If your position size is set so that a stop filled exactly at the level produces an acceptable loss, and a fill fifty percent beyond the level produces an unacceptable one, then you are sized against the average case and exposed by the tail. Sizing with headroom — so that a worse-than-requested fill is survivable rather than merely unpleasant — is the only mechanism that actually addresses this, because no order type can supply a price the market is not offering.

Two related exposures belong in the same mental category. Holding a leveraged position across a weekend means accepting the reopening gap as a known, recurring feature rather than an accident. And negative balance protection, where offered, is a backstop against an extreme outcome rather than a substitute for sizing — it defines the worst case, it does not make the worst case comfortable. The honest summary is that a stop order is a discipline mechanism, not an insurance policy, and treating it as the latter is what turns a normal loss into an unrecoverable one.

Every instrument on a trading platform is a contract with published terms, and those terms determine the behaviour of a position before any market opinion is involved. The specification is not administrative fine print — it is the definition of what you are actually holding. A striking share of the outcomes traders describe as unfair turn out, on inspection, to be terms that were published all along and never read.

Start with contract size, because everything numerical descends from it. One lot means a specific quantity of the underlying, and that quantity differs by instrument and asset class. Contract size combined with the price quotation gives tick value: the account-currency result of the smallest price increment on one contract. Those two lines are what convert a chart move into money. Without them, "the price moved thirty points" is not yet a statement about your account. A trader who has internalised tick value on one instrument and assumes it transfers to another is the most common version of this error.

Read trading hours next, and read them as a constraint on your orders rather than a convenience note. An instrument that is closed cannot trigger your stop, which means an adverse move during a closed period is realised at the reopen rather than at your level. Many instruments also have a daily break or a session structure that differs from the underlying market’s reputation, and hours shift with daylight saving in the venue’s own jurisdiction. If the instrument does not trade continuously, then "my stop was at that level" describes an intention, not a mechanism.

On dated instruments — futures-style contracts on energy, indices and some commodities — expiry and rollover rules matter enormously, because they change your position without any action from you. The specification states the expiry date, how a position is treated at expiry, and whether it is rolled into the next contract and on what terms. A position that survives a roll is not the position you opened: the reference contract changed, and any price difference between the expiring and the new contract is dealt with by an adjustment mechanism the specification describes. Being surprised by this is entirely avoidable, and only by reading.

Then read the sizing constraints. Minimum trade size, maximum trade size and the step increment between them determine whether the position your risk arithmetic produced is actually available. If your stop-distance calculation returns a size below the minimum increment, the honest response is to skip the trade or widen the account, not to round up to the minimum and quietly accept more risk than your policy allows. Maximum position and aggregate exposure limits per instrument matter for the same reason at the other end.

Finally, read the margin and financing lines together. Margin requirement per lot tells you what capital the position consumes, and it can be tiered — larger positions attracting higher requirements — and it can be raised ahead of known events. The swap line tells you the daily financing consequence and which weekday carries the multi-day charge. Neither line is a market view; both are certainties you can compute before entering, and both are routinely discovered after the fact.

The habit worth building is simple and takes minutes. Before the first trade in any unfamiliar instrument, open its specification and write down six numbers: contract size, tick value in your account currency, margin per lot, trading hours in your own time zone, swap for both directions, and the minimum size increment. If any of those six is a surprise, you have just avoided learning it the expensive way. If none is, you have lost five minutes and gained the ability to size the position correctly on the first attempt.