Published as a TradingView Idea on BLUAIUSDT. View the chart on TradingView
Position sizing advice is close to universal. Decide what you are willing to lose, measure the distance to your stop, and divide. Risk two percent, place the stop where the idea is wrong, size so that those two agree.
We did exactly that. The trade still cost one and a half times the intended risk and every cent of margin behind it, and the stop order never executed once.
What the four prices were
Long at 0.013180. Stop at 0.012499, which is 5.17% below. Two percent of the book divided by that distance gave the size. Intended loss if wrong: 6.48.
The exchange has two prices of its own on the same position.
The liquidation trigger at 0.012631, which is 4.16% below entry. The bankruptcy price at 0.012126, where the margin posted is exactly used up.
Put them in the order price meets them on the way down and the problem is visible without any arithmetic at all: 0.012631 comes before 0.012499. The exchange’s exit is above ours.
Why the stop never fired
It could not. By the time price reached 0.012499 there was no position left to close. It had been liquidated at 0.012631, a hundred and thirty two ticks earlier.
This is worth sitting with, because a stop that never executes does not look like a failure. The position simply is not there any more, and if you are not watching you will assume the stop did its job.
What it actually cost, which is the surprising part
We assumed, right up until we derived it, that being liquidated at 0.012631 would cost roughly what stopping out at 0.012631 would cost. About 5.16 on this size.
It cost 10.32.
Liquidation does not close you at the liquidation price. That price is only the trigger. The close happens at the bankruptcy price, the level where your posted margin reaches zero, and the margin goes with it.
| amount | |
|---|---|
| intended loss, at the stop | 6.48 |
| the price move to the liquidation trigger | 5.16 |
| margin posted | 9.90 |
| actual loss | 10.32 |
Margin plus round trip fees. Five dollars of adverse price became a ten dollar loss, and the extra five was not price at all.
But surely the exchange would not allow that
It is the first thing anyone asks, and it is the right question. An exchange will not usually let you place a stop beyond your own liquidation price.
It did not have to. When the stop was placed, it was not beyond it.
These positions run on cross margin, and in cross margin your liquidation price is not a property of that position. It is a property of your entire account’s equity. It moves while you sit there, and the thing that moves it is your other trades.
The stop never moved. The liquidation price walked up to meet it.
So the check is not a formula you run once
In isolated margin you can size it at entry. Liquidation sits roughly 1/leverage below your entry for a long, pulled closer by the maintenance margin rate, so the ceiling on leverage is
L < 1 / (stop% + MMR)
At 12.5x that is 8%, comfortably outside a 5.17% stop. Add maintenance margin and the trigger moves to 4.16%, inside it.
In cross margin that arithmetic is necessary and not sufficient, because the right hand side is not yours alone. You can pass it at entry and still be liquidated first, because a position you were not thinking about moved your liquidation price.
What protects you is duller than a formula. Read the liquidation price and the stop off every open position and compare them, repeatedly.
For a long, the stop must sit above the liquidation price. For a short, below it. When that ordering inverts, the stop is decoration.
Here is the same book today:
| symbol | side | stop | liquidation | apart |
|---|---|---|---|---|
| PORTAL | long | 0.015086 | 0.014154 | 5.42% |
| HYPE | short | 88.21 | 90.31 | 2.52% |
| CL | long | 86.94 | 85.37 | 1.72% |
| AAVE | long | 121.85 | 118.20 | 2.78% |
| CHIP | short | 0.05886 | 0.06274 | 7.88% |
All five correctly ordered. The tightest is 1.72% from inverting, and that gap changes every time anything else in the account moves.
The term everybody gets wrong
The maintenance margin rate is tiered. It depends on how big the position is, and on this symbol it ranges from roughly 4% to 60%.
We had carried a remembered rule of thumb. The real figure was several times larger, which is exactly the direction that hurts, and it is why the fix was not a better memory but reading the number from the exchange’s risk limit table at the tier the notional lands in.
Two funding payments also took 0.25 out of a 10.03 margin while the position was open. Funding walks the liquidation toward your stop before price has done anything at all.
What we changed
Every card the system now produces derives a maximum leverage from the live risk limit table and refuses a setup it cannot size with the liquidation safely outside the stop, with a buffer for the residual and for funding.
Checked against the real event: predicted liquidation 0.012653, actual 0.012631. A difference of 0.17%.
The honest scope
One position. The mechanism is arithmetic and reproduces on any exchange with tiered margin. The loss is a single receipt, not a study.
And none of this is an argument against leverage or a claim about position sizing, which was correct here. It is one check most people never run: whether the stop you sized from is on the reachable side of the exchange’s own exit.
Published as a TradingView Idea on BLUAIUSDT. View the chart on TradingView
Past chart behaviour and measurement, described after the fact. Nothing here is a recommendation or an indication of future results.