Cross margin

What you will learn

The formula

Margin = lots × contract size × price ÷ leverage. In XAUUSD one lot is 100 troy ounces, so at $4,600 and 1:500 a single lot ties up $920 while controlling $460,000 of gold.

Why the confusion is expensive

If you think you are risking $920, you will size as if you were risking $920. But a $10 move in gold is $1,000 on that lot: you have already lost more than the margin held. Margin limits how much you can OPEN, never how much you can lose.

The four figures

Balance is realised money. Equity is balance plus floating P&L. Used margin is what open positions tie up. Free margin is equity minus used margin, and it is the only thing that decides whether the next order fills.

The two thresholds

Margin call usually sits near 100% and stop-out between 20% and 50%, depending on the broker; the EU margin close-out rule fixes it at 50% of the minimum required margin. The first is a warning; the second closes positions by itself, starting with the worst loser. In a fast market you can cross both without seeing the warning.

How each one works

Under isolated margin each position has its own collateral and its liquidation does not touch the others. Under cross margin all the equity backs everything, so one losing position eats the cushion of the rest.

When cross is the right choice

With correlated or hedged exposure, moderate leverage and trades that form a single idea. There it avoids liquidations caused by a temporary wick. With aggressive pyramiding it does the opposite: it shortens the distance between the margin call and liquidation, because everything shares collateral.

The two distances, measured

Five gold lots bought at 4,328.15 with 1:500. Isolated, the position loses its margin $8.66 away — price ÷ leverage, independent of size. Cross, with a $5,000 balance and a 50% stop-out, the close-out lands at $5.68: at 4,322.47. The account with MORE collateral behind it liquidates first, because the broker's threshold is not losing the margin but dropping below half of it.

And margin moves as well

In CFDs margin is recalculated at market price, not frozen at entry: as a long's price falls, the required margin shrinks. The effect on the cushion is small — here it goes from $5.672 to $5.678 — and it works in your favour, so that is not where the danger sits. It is worth knowing for the opposite reason: whoever cites it as cross margin's big problem has not done the arithmetic. The problem is the threshold.

Worked case

A $5,000 account, 1:500, five gold lots at 4,328.15. Used margin 4,328.15 and free margin 671.85. Price rises five cents and you want to add another five lots: it needs 4,328.20 and there is 696.80. Order rejected; only 0.80 lots fit. The whole plan rested on a step that does not happen.

What price unlocks it

The second rung does not fill until 4,335.49, and there it leaves free margin at zero. In other words: for a five-cent ladder to work, price has to travel seven dollars first — a hundred times the spacing you planned. With $10 spacing the same plan fills completely, because floating profit grows faster than margin. That is the difference between pyramiding and stacking.

The check

The cost of a rung is not "lots × unit margin": it is the difference between used margin before and after. Only that way is it right when the rung goes the other way and reduces net exposure instead of increasing it.

The three rules

Add only on confirmation, make each rung smaller than the last, and raise the stop with every addition so total risk never exceeds what was defined on the first entry. Drop any one of the three and the technique turns into its opposite.

The alarm signal

If the cushion to the stop-out SHRINKS as you add rungs, you are not pyramiding. In the simulator's ladder with $10 spacing in your favour, the cushion goes from $5.68 after the first rung to $17.65 once the fifth is in: the trade is working and equity rises faster than margin. Run that same ladder against you and the second rung no longer fills.

What does happen

Under the net margin model, five longs and five shorts give zero exposure and zero margin: the margin level goes from 116% to undefined and the order fills even if you had almost no free margin.

What does not happen

If you use that margin to add five longs, you are net five lots again and margin returns to the original amount. Same starting point, minus the spread on four trades. And when you unwind the lock the full exposure reappears at once: closing the protective leg with the margin level at 100% triggers the stop-out on the next tick.

The model matters, a lot

Not every broker nets. Under the model that charges both legs in full, those same five hedged lots go from $4,328 of margin to $8,656 and the margin level falls from 116% to 58%: eight points from a 50% stop-out. The hedge, intended as protection, is what triggers the liquidation. Check it in the instrument sheet before, not after.

The numbers

With $5,000, gold at 4,328.15 and a requirement to survive a $70 adverse move without liquidation: at 1:200 you fit 0.620 lots; at 1:500, 0.673; at 1:1000, 0.693; at 1:2000, 0.704. Going from 1:500 to 1:1000 buys you 3% more defensible size.

The ceiling

The formula converges to balance ÷ (contract × cushion). With $5,000 and a $70 cushion the ceiling is 0.714 lots, and no leverage beats it. When an account cannot carry a position, the problem is the capital, not the leverage.

The nuance almost nobody mentions

For a position that is ALREADY sized, more leverage does give more cushion: five lots survive $5.68 at 1:500 and $7.84 at 1:1000, because the stop-out threshold is a percentage of a smaller margin. Leverage was never the danger. The danger is the size it tempts you to open.

Order of magnitude

A $0.25 spread in gold across five lots is $125 per trade. Opening and closing a lock is four trades: $500. Against the $671.85 of free margin in the worked case, 74% goes on execution alone.

What the chart does not show

Spread widens around macro releases and into the weekly close. Swap is charged every night and on a locked position it is usually charged on both legs. Sunday's opening gap can jump your entire stop-out: liquidation happens at the first available price, not at the one you computed.

The ruin formula

With no drift, the probability of reaching +a before losing −b is b ÷ (a + b). With a $7.66 cushion and a $260 target that is 2.9%. Adding a strong uptrend — $20 a day with $60 of daily volatility — lifts it to 8.6%: still a plan that fails nine times out of ten.

The trade-off

Cutting the entry size raises the cushion and the odds, at the cost of requiring more travel for the same profit in money. That trade-off is the whole decision, and it is taken on the first order: after that you cannot change it without closing.

The consistency check

If a plan requires tripling the account before it starts, and you know how to do that, the plan is redundant: repeating that same thing gets further without ever going through a margin call. Any plan that leans on having already solved a harder problem than itself is circular.

The caps

Major currency pairs 30:1; non-major pairs, major indices and gold 20:1; other commodities and non-major indices 10:1; single equities 5:1; crypto 2:1. On top of that, mandatory position close-out at 50% of the minimum required margin and negative balance protection per account.

What you give up outside

At 20:1, five gold lots at 4,330 require $108,250 of margin instead of $4,330. High leverage only appears at firms in jurisdictions such as Belize, Seychelles or Mauritius, where negative balance protection is not guaranteed by rule and the complaints route is different. Before trading on leverage, check in writing whether you have it.