Every Major Crypto Crash, Ranked by How Fast It Happened

Crashes in this market fall into two types. Slow declines reflect changing expectations. Fast ones reflect a mechanism failing. The speed tells you which.

Ranked by time from peak to trough

1. TerraUSD and Luna, May 2022. About five days.

An algorithmic stablecoin lost its peg and the mechanism designed to restore it destroyed the value of the paired token instead. Tens of billions of dollars of nominal value went to effectively zero inside a week.

This was the fastest because it was a design failure rather than a repricing. The exit incentive ran backwards under stress: the more people doubted the peg, the more of the paired token was minted, which pushed its price down and deepened the doubt.

2. FTX, November 2022. About ten days.

A balance sheet report in the press prompted a withdrawal run, the exchange could not meet it, and a firm valued in the tens of billions filed for bankruptcy inside two weeks.

This was a custody failure, not a market one. Customer assets were not there. The market decline that followed was contagion through counterparties.

3. March 2020. One day.

Bitcoin fell roughly half in twenty-four hours as global markets sold everything liquid at the start of the pandemic. Crypto was one of the few markets open, which made it a source of cash.

Notable as the cleanest example of crypto trading as a risk asset rather than a hedge.

4. May 2021. About three weeks.

A combination of Chinese mining restrictions, leverage unwinding and a sharp sentiment reversal cut the market roughly in half. No single mechanism broke. This was a crowded position clearing out.

5. The 2018 decline. About twelve months.

The token issuance boom of 2017 unwound over a full year. Most of the assets that fell in this period never recovered, because there was nothing behind them.

Slow, and therefore repricing rather than failure.

6. The 2014 collapse. About fourteen months.

Mt. Gox failed in February 2014, and the market took more than a year to find a floor. The largest exchange of its era proved to have been insolvent for a long time.

The pattern

Speed	Usual cause	What it damages
Days	Design or custody failure	Confidence in a specific structure
Weeks	Leverage unwind	Positioning, not fundamentals
Months	Repricing of expectations	Valuations across the sector

The fast ones are more frightening and more survivable, provided the failure was in something you did not hold. The slow ones take less from you in any given week and more in total.

What each one changed

Mt. Gox produced the idea that exchanges should be audited, and the reserve attestations now published by venues such as an OTC crypto desk with a named counterparty are its direct descendant. FTX produced proof-of-reserves reporting. Terra ended the algorithmic stablecoin category. March 2020 ended the argument that Bitcoin was uncorrelated.

Each time, the specific failure was patched and the general lesson was forgotten within about two years. That is the part of the pattern most worth remembering.