Worked scenario: two positions and a margin call

The account before the second position 

The starting state is built entirely from fields this module has already described one at a time. A balance sits in the account. One position is open against it, and the margin required for that position has moved out of free margin and into used margin, where it stays for as long as the position is open. The position has moved against the account, so an unrealised loss is running: it is subtracted from balance to give equity, and it is not deducted from balance itself, because nothing has been realised. Margin level then divides equity by used margin and prints the answer as a percentage. The block below does nothing new. It performs those familiar steps in sequence, for one account, at one moment.

Two features of that state are worth naming before anything is added to it. The numerator has already moved and the denominator has not: the unrealised loss has taken equity down, while used margin has held still, because neither the size of the position nor the margin requirement written against it has changed. And margin level is a ratio, so either part can move it alone. A fall in equity moves it. A rise in used margin moves it. Neither requires the other, and the next section changes only the second.

What the second position changes before any price moves 

Opening a position does not spend money. It reserves it. The margin required for the new position is moved from free margin into used margin, and both totals change by the same amount in opposite directions. Balance is untouched, because nothing has been realised. Equity is untouched at that instant too, because the new position opens at the price it opens at and has not yet moved. So the numerator of the fraction is exactly what it was a moment earlier, and the denominator is larger. Margin level falls, and it falls before the second position has done anything at all.

The halving in that block is not a loss. No price moved, no money left the account, and the balance is the figure it was before. What changed is the quantity margin level is measured against. That is why two accounts can show identical equity and very different margin levels, and why the field reads as a relationship rather than as a score: the second number is set by what is open, not by what the market has done.

Two positions are not automatically two separate exposures 

The account panel adds used margin up without asking what the positions are written on. Nothing in the arithmetic distinguishes two positions that respond to entirely different forces from two positions that respond to the same one. The market frequently does distinguish them. Instruments that share a currency, an index, a sector, a producing region or a sensitivity to the same interest rate decision tend to move together, and when they do, a single event registers in both unrealised figures at once. Both losses land in the same equity total, because there is one equity figure for the account and not one per position.

The statistical name for that tendency is correlation, and the word carries more precision than it is usually given. Correlation is a measurement taken over a chosen window of past prices, describing how two series moved during that window. It is not a property of the instruments, it is not fixed, and it is not a forecast. Practitioners who use it routinely disagree about how much weight it can carry: correlations measured in calm conditions have shifted during stressed ones, and instruments that had appeared to move independently have moved together in the same episode. A number taken from one window can therefore describe the next window poorly, which is a limit of the measurement rather than a fault in a particular figure.

Both positions move against the account 

From the state above, the market moves. The first position, which was already showing a loss, shows a larger one. The second position, opened at the price it was opened at, now shows a loss of its own. Neither has been closed, so both losses remain unrealised, and both are subtracted from balance in the same step. The denominator has not moved at all: used margin is a function of position size and the margin requirement, and neither has changed. So the whole of the fall in margin level comes from the numerator, and the numerator is falling at the combined rate of both positions.

The adverse column is the compounding this lesson is named for, and it is worth being precise about what compounded. The individual losses did not grow faster because there were two of them. Each position lost exactly what its own size and its own price move produced. What changed is that the account absorbed both of them out of one equity figure, while measuring that figure against a used margin total that both positions had already enlarged. A denominator raised by the second position and a numerator falling from both of them is the whole of the mechanism, and there is nothing else in it.

It also has to be said plainly, and in the adverse direction first, that the arithmetic in that block does not stop at the bottom of the account. Profit and loss on each position are calculated on the full value of the contract rather than on the margin held against it, so an adverse move is measured against the whole of both contracts, the combined loss is not bounded by the combined margin, and a loss is not limited to the amount deposited. A favourable move is measured on exactly the same basis and to exactly the same degree, which is what the favourable column of the block shows.

The level the arithmetic reaches 

The margin call in the adverse column is not a separate event that happens to the account. It is the last row of the block. A margin call level is a stated percentage, margin level is a percentage, and the comparison between them is made continuously by the platform against whatever the current figures are. Nothing is calculated at that moment which was not being calculated a moment earlier, and no decision was taken by anybody. The reading arrived at a threshold that had been published before either position was opened, and the notification reports that it did.

Below the margin call level sits the close out level, which is the point at which the platform begins closing positions on its own initiative rather than reporting the state of the account. It is likewise a published term rather than a market variable, and it is likewise compared against the same single fraction. The stop out level at YAL is 50%. The blocks in this lesson use a different figure for their close out threshold, chosen only so that its arithmetic stays legible, and each one states that it is an assumption rather than a term.

The same account holding one position 

The comparison that makes the scenario mean something is the one where everything is held constant except the second position. Same balance, same first position, same margin requirement, same price path for the instrument that first position is written on. The block below runs that account through the same adverse move and the same favourable move, and then measures both versions against the same assumed close out level, so the two endings can be read side by side rather than described.

The two adverse readings come from the same market and the same first position, and they are a long way apart. Reading the rows in order shows why, and it is two effects rather than one. The second position raised used margin, which lowered every margin level the account would ever print from that moment onward, including the ones it printed before the market moved again. It then contributed a loss of its own to a single shared equity figure. Either effect alone would have shortened the distance to the close out level. The block shows them arriving together, because in an account they arrive together.

The favourable rows are the same statement with the sign reversed, and they are not a smaller claim than the adverse ones. The version holding both positions ends the favourable move with more equity than the version holding one, for exactly the reason it ended the adverse move with less. Nothing in the structure of an account panel favours one direction over the other. What the comparison isolates is not a direction. It is that position count moves the denominator permanently and the numerator whichever way the market goes.

What the comparison does not show 

Three limits sit on everything above, and stating them is part of teaching the arithmetic rather than a qualification bolted onto it. The price path was held fixed by assumption so that one variable could be isolated, which is a property of a worked example and not of a market. The two positions were assumed to move against the account together; positions taken in opposite directions on instruments that move together can offset each other in the equity figure while still adding to used margin in full, and firms differ in how they treat that case. And none of this is a statement about how many positions an account holds well.

That third limit is the important one. This page knows nothing about any reader's circumstances, holdings or intentions, and it puts forward no figure for a number of positions, a share of free margin or a distance from a close out level. What it does is show where each field comes from when more than one position is open, so the panel reads as a calculation with visible inputs rather than as readings that change for reasons the account holder cannot reconstruct.

Where practitioners disagree 

The first disagreement is about whether used margin is a meaningful measure of what an account is carrying. One tradition treats it as the working measure, on the grounds that it is the figure the close out arithmetic actually uses, it is published per instrument, and it requires no estimation. Another argues that it describes what the firm is holding rather than what the account is exposed to, because two sets of positions with identical used margin can have very different sensitivity to a single event, and margin requirements are set instrument by instrument rather than across a portfolio. Both descriptions are accurate about different things, which is why the argument does not resolve. Margin level uses used margin because that is the quantity the account has posted, not because it is the better measure of risk.

The second disagreement concerns correlation itself. Some practitioners estimate it explicitly and treat instruments that have moved together as a single exposure for sizing purposes. Others hold that the estimate is unstable enough to be misleading, and point out that it is measured precisely in the calm periods when it matters least and shifts in the stressed ones when it matters most. A third position holds that the correlation question is beside the point at the level of an account panel, since used margin adds up regardless of what any two instruments have done historically and the close out comparison never consults a correlation figure at all. None of the three is refuted by the others, and this page takes none of them.