Same Trade, Different Liquidations: How Exchange Mechanics Decided Bitcoin’s 67-Day Funding Squeeze

For short sellers in BTC perpetuals, that meant roughly 201 separate funding debits — three settlements per day across 67 days — chipping away at margin while the price ground sideways to higher.

When the squeeze finally arrived and short positions started getting forced out, something unusual showed up in the post-mortem. Traders running identical short positions — same leverage, same entry, same size — did not all get liquidated at the same price. Some got partial closes, while others got wiped. CoinGlass liquidation data showed volumes distributed unevenly across major exchanges in the 48 hours that followed.

Why 67 Days of Negative Funding Punishes Short Positions

Perpetual futures use funding payments to keep contract prices anchored to spot. The mechanism is straightforward: when funding is positive, longs pay shorts. When it is negative, shorts pay longs. Most major venues settle these payments every eight hours, which over 67 days compounds to 201 settlement points.

For a short seller, each negative-funding settlement is a direct debit from margin. None of it shows up on the price chart. The position can be flat or even moderately profitable on a P&L basis while the underlying margin balance bleeds — slowly during quiet intervals, more aggressively during periods when funding turns deeply negative.

The 2026 streak hit that profile. Funding stayed in red ink while BTC traded between roughly $74,000 and $83,000, drifting higher without breaking. The price chart looked stable. The margin accounts of short sellers did not. By the time the squeeze caught up — coinciding with broader risk-off flows tied to renewed US-Iran tensions — many shorts were already operating with substantially less cushion than their notional position size suggested. Roughly $590 million in positions were unwound in the 24 hours leading into early European trading on 18 May, according to CoinGlass data. That part of the story has been well covered. The part that has not is why the same setup produced different outcomes on different venues.

Maintenance Margin: The Number That Decides Your Liquidation Price

Maintenance margin is the minimum equity an exchange requires you to hold against an open position before it force-closes the trade. The number sounds dull. It is the single most important variable separating one venue from another at the moment of liquidation.

Binance currently runs a 0.5% maintenance margin on standard BTC perpetual positions in its lowest tier, with the floor rising as position size increases. Other major exchanges run tiered structures with different starting points and different escalation curves. The differences look trivial in isolation. They are not.

On a 20x leveraged BTC short, a 0.1% difference in maintenance margin shifts the liquidation price by roughly 2% of the underlying. For a position opened around $80,000, that is the difference between a liquidation trigger near $83,200 and one near $84,800. In a market that moved close to 8% over a compressed window, that is the difference between getting force-closed and surviving the wick.

These tiers are public. They sit in exchange documentation, often several clicks deep. The Leverage.Trading breakdown of crypto futures liquidation mechanics compares them across the major venues. Most retail traders never look.

Partial vs Full Liquidation: The Engine Design Most Traders Ignore

The second mechanical variable is how the liquidation engine itself behaves once the trigger is hit. There are two broad designs in use across major venues, and they produce very different outcomes during fast moves.

Binance and Bybit use partial liquidation. When a position breaches maintenance margin, the engine reduces position size in increments — closing a percentage of the trade, allowing the remaining position to absorb the next price tick, and only escalating to full closure if margin keeps deteriorating. The trader takes damage but may retain part of the position if price stabilises.

Several other major perpetuals venues run full-position liquidation. There is no incremental step. The trigger hits and the entire position closes at the next available market price.

In the conditions that prevailed during the May squeeze, with BTC moving sharply across short windows and order books thinning, the design choice was decisive. Identical positions at identical entries on partial-liquidation venues experienced partial closes. The same trades on full-close venues were wiped in one move. The trader had no input in the outcome beyond their choice of exchange weeks or months earlier.

This is not a rare or esoteric scenario. It is the standard distribution of outcomes during any volatility event large enough to trigger cascading liquidations.

Funding Rate Caps and the Compounding Effect

The third variable is less visible but matters most over long funding regimes like the one just ended. Per-interval funding rate caps determine the maximum amount that can be debited or credited in a single 8-hour settlement.

Caps are not uniform across exchanges. Most major venues impose ceilings on how negative funding can go in any single interval, but the ceiling values differ — sometimes by factors that look small per settlement and large in aggregate. Over 201 sequential settlements during the 2026 streak, those differences compounded.

The result is that a short position of identical notional value, opened simultaneously across two venues, can arrive at the moment of squeeze with materially different margin balances. The trader who chose the venue with tighter funding caps walks into the liquidation event with more cushion. The trader who chose the venue with looser caps walks in already weakened.

“Three Numbers That Matter More Than the Trade”

Anton Palovaara, founder at Leverage.Trading, sees the same blind spot repeatedly in his work with retail and prop traders:

Why Exchange Selection Is a Risk Question, Not a Fees Question

The conventional approach to choosing a derivatives venue weighs maker-taker fees, spread quality, listed product range, and withdrawal frictions. These matter. They are also the wrong primary criteria for any trader who plans to use leverage.

What the 67-day funding regime and its aftermath showed — and what the exchange-level CoinGlass liquidation breakdown makes legible — is that the same position opened at the same time across two venues will not produce the same outcome during a volatility event. Maintenance margin tiers, mark price methodology, liquidation engine architecture, and funding cap structures all sit upstream of any trade decision. Once a position is open, none of these are negotiable. They are the ground conditions a trader accepts the moment they fund the account.

Most coverage of the squeeze framed it as a positioning story — too many shorts caught on the wrong side of a market that refused to break lower. That framing is correct as far as it goes. It does not explain why the distribution of pain was uneven across venues that all saw the same price action.

The exchange-level data tells a more useful story. The same trade, placed at the same time, produced different outcomes on different platforms. Not because of market judgement, but because of how each venue holds margin, caps funding, and executes forced closes.

For traders preparing for the next leveraged cycle — and historical analysis across six comparable negative funding regimes since 2018 shows all six delivered positive forward returns at 90 days, with win rates of 83% to 96%, per K33’s data — the lesson is upstream of any specific trade. Read the exchange before placing the position. Compare maintenance margin tiers at the size you intend to trade. Confirm whether the liquidation engine reduces or closes. Check the funding cap structure. The Leverage Trading crypto futures liquidation guide lays out a framework for running that comparison before capital is committed, not while a position is being force-closed.

The 67-day streak is now part of the historical record. The next one is already being set up by positioning that none of us can see yet. The only variable a trader can fully control between now and then is which venue they are on when it arrives.