Reducing Leverage Without Closing: A Partial Close Strategy

What a partial close actually changes

Here's the thing most guides skip: your liquidation risk is not set by your leverage number. It is set by the relationship between two moving quantities — your margin balance and your maintenance margin requirement.

Maintenance margin is the minimum collateral your exchange demands to keep a position open. It scales with notional value: roughly notional × maintenance margin rate (MMR). Margin balance is your collateral plus unrealized PnL, minus funding paid. Liquidation fires when margin balance falls to the maintenance requirement.

That gives you exactly two levers. You can raise the margin balance (add margin), or you can lower the maintenance requirement (reduce notional). A partial close pulls the second lever: you close part of the position, the notional shrinks, and the maintenance requirement shrinks with it.

The subtle part — and where people get burned — is what happens to the margin that was backing the closed portion. In isolated margin, closing part of a position releases a proportional slice of margin back to your wallet. If you leave that slice in the position, your remaining notional is backed by more margin per unit, and your margin ratio improves. If you withdraw it, you have booked the loss and shrunk the position without improving your cushion at all.

The arithmetic: a worked example

Let's be real about the numbers, because "cut your position" means nothing until you see the margin ratio move.

Say you are long 1 BTC at $80,000 on 10x leverage in isolated mode. Your position margin is $8,000. The exchange's maintenance margin rate for that notional tier is 0.5%. Price falls to $78,000.

Now you close half — 0.5 BTC — at $78,000, booking a $1,000 realized loss, and you keep the freed margin inside the position. The remaining 0.5 BTC has a notional of $39,000 and a maintenance requirement of $195. Your margin balance stays at $6,000 because you did not withdraw anything.

The margin ratio halved. In margin-ratio terms, you went from "one bad candle from trouble" to a materially larger buffer — and you did it without predicting the next move. You did not need to be right about direction. You just needed to stop the maintenance requirement from scaling with a position that was working against you.

The trade-off is equally real: you have locked in half the loss, and if price snaps back, you only recover on half the original size. A partial close is not free — it is an insurance premium paid in forgone upside.

Partial close vs adding margin vs full close

Most guides get this wrong by treating the three responses as interchangeable. They are not — each one moves a different variable and carries a different cost.

There is no universally correct answer. The honest framework is this: if your original reason for the position is intact and you have spare collateral, adding margin keeps the trade alive at a cost of capital. If your reason is intact but you have no spare collateral, a partial close is the compromise. If the reason is gone, the size question is irrelevant — close it.

One mechanical detail worth knowing: on tiered-margin venues, adding margin can backfire. If your added collateral pushes the position into a higher notional tier, the maintenance margin rate for that tier rises, and part of your extra buffer gets eaten by the higher requirement. Check the tier table before you top up a large position. For the full mechanics, see initial vs maintenance margin.

Execution: reduce-only orders and slippage

The single most common execution mistake is sending a normal market order to "close half." On many venues a plain order can be interpreted as a new order, and if it over-fills you can accidentally flip your position or leave an unwanted residual. Use the order type built for this:

• Reduce-only. This flag guarantees the order can only shrink your position, never increase or reverse it. It is the correct tool for any partial close. If you send a reduce-only order larger than your position, the exchange fills only what is needed and cancels the rest.
• Limit or IOC over pure market. A market order in a fast move pays the full spread and any slippage on the way down — exactly when spreads are widest. A limit order at or near the mark, or an IOC that fills what it can and cancels the rest, usually gets a better average fill.
• Size to the tier boundary, not a round number. If cutting your notional drops you back into a lower maintenance tier, the benefit compounds. Check whether a slightly larger close crosses a tier line.

Here's the deal on timing: partial closes work because they are boring and mechanical. If you find yourself agonizing over the exact entry price for the reduction, you have already lost the plot — the point is to reduce the maintenance requirement before the next candle, not to time a perfect exit.