An exchange being unreachable during a sharp move is not a minor inconvenience. Positions cannot be closed, collateral cannot be added, and liquidations proceed regardless.

Load arrives exactly when it hurts

Trading systems are sized for expected volume, and a violent move produces order rates far above anything a normal day generates.

Matching engines, risk systems and the interfaces users depend on all face that surge simultaneously, and the weakest component determines what fails.

The correlation is what makes this dangerous. Outages concentrate in the moments when the ability to act has the most value.

Liquidation does not wait for access

Risk engines usually run independently of the user-facing systems, so a position can be closed for insufficient margin while the owner cannot log in.

Adding collateral requires the same unavailable interface, which removes the one action that would have prevented the outcome.

The asymmetry is structural rather than malicious. Automated risk management is built to keep running when everything else degrades.

Withdrawals are a separate failure

Trading can be operational while withdrawals are paused, often for reasons involving the underlying chain or the venue's own hot wallet management.

A paused withdrawal traps assets at one venue, which is a serious problem for anyone whose strategy depends on moving collateral between places quickly.

It also breaks the arbitrage that holds prices together, letting a venue's quoted price drift from the wider market until transfers resume.

Where the loss actually falls

Terms of service at most venues disclaim liability for losses arising from downtime, and enforcing a claim against an offshore operator is difficult for an American user.

Some venues have compensated affected users voluntarily after severe incidents, which is a discretionary gesture rather than an obligation anyone can rely on.

The practical position is that outage risk sits with the trader, priced into nothing and recoverable through no defined process.

How the risk is normally managed

Spreading positions across venues reduces concentration but multiplies the number of counterparties and the operational work of monitoring them.

Lower leverage is the more direct response, since a position with a distant liquidation level survives an hour of inaccessibility that a tightly margined one does not.

Placing protective orders on the venue rather than intending to act manually helps only if the venue's own systems remain functional, which is the assumption in question.