Hyper-Manipulation. On October 10, 2025, the crypto market exploded in full force. In just a few minutes, thousands of traders saw their positions forcefully closed on Hyperliquid, sometimes at the best time, often at the worst. One word ignited the community: ADL (Auto-Deleveraging).
This analysis is brought to you by 21M ⭕, the community of crypto investors behind the Club 25%.
Context: the crash of October 10, 2025
On October 10, 2025, a sudden drop in the crypto market led to a spike in liquidations on the perpetual trading platform Hyperliquid. To avoid insolvency, the protocol triggered a mechanism called Auto-Deleveraging (ADL) : an automatic procedure which closes the most profitable positions to absorb losses from liquidated accounts. In other words: you are making a profit, everything is going well… and your position is cut. In absolute terms, nothing serious, but it becomes serious when you deploy a so-called strategy “delta neutral” where your position is linked to another. Example: long on Hyperliquid; shorts on Paradex. If your position is cut on Hyperliquid and the other is maintained, you find yourself completely exposed to the market. Hyperliquid recorded:
35 000 ADL in a few minutes
20,000 users affected
161 tokens affectedmostly BTC, ETH, SOL, HYPE
This system caused controversy: some traders accused Hyperliquid of having “stolen” their profit by closing their shorts just before the total crash. Three major voices then emerged, which you will find on Twitter: @notnotstorm, @seanlippel et @fiddybps1.
@notnotstorm's thesis: “the ADL protected traders and improved their PnL”
@notnotstorm defends Hyperliquid.
He claims that the dominant narrative is false : according to on-chain data, the majority of shorts were closed at the bottom pointactually maximizing profits.
Ses arguments :
99% of ADLs occurred during a 5-minute volatility peak (around 9:16 p.m. UTC). Prices then rebounded, so those who remained short would have lost part of their earnings. In other words, the ADL would have locked an optimal PnL for the majority of traders. Moreover, sans ADLHyperliquid risked insolvencywhich would have wiped out all of the positions.
The conclusion of @notnotstorm : the ADL system fulfilled its role: protecting the platform and, paradoxically, save profits of the majority of short traders.
@seanlippel's response: “It's a partial and misleading vision”
@seanlippel, a professional trader operating on several platforms (Drift, dYdX, Hyperliquid), responds to @notnotstorm and accuses him of propaganda.
His main criticisms:
@notnotstorm's analysis only covers large caps (BTC, ETH, SOL)while the majority of losses came from altcoins plus volatiles (ATOM, STX, APT, FET…). Them ADL on these secondary assets have been triggered earlierbefore the rebound phase, eliminating winning positions prematurely. For a trader in balanced long/shortADLs have disrupted portfolio balancesoften leaving the long ones open — which led to cascading liquidations. Finally, Sean denounces structural flaws : None transparency on the ADL queue
No protection flash crash
Oracles “one-touch” (liquidation on a single price tick)
No insurance funds significant And therefore a systemic risk for any institutional trader.
The conclusion of @seanlippel : Hyperliquid is not not reliable for complex strategies. The ADL system, opaque and rigid, has destroys healthy positions without offering the guarantees expected of an exchange of this size.
@fiddybps1's analysis: “ADL is a relic of the past”
@fiddybps1 has one structural reflection on exchange models. He goes further in calling into question the very concept of ADLjudge obsolete and unsuitable for modern platforms. The ADL system has just BitMEX (2017)a centralized model where each market operated in isolated marginwithout an overview of the user account. This system is incompatible with cross-marginwhere positions in multiple assets offset each other. The consequence is that a long BTC / short ETH trader can be forced to close his winning BTC via ADL to save Hyperliquid's insurance fund, while keeping his losing ETH shorts open → result: net loss despite an initially neutral position. The consequences: Breakdown of hedges and “balanced” positions (delta neutral)
Amplification of the volatility due to unnecessary forced sales
Injustice between traders according to the ADL queue
Proposed alternative: Socialized Loss (SL)
@fiddybps1 offers another model: socialized losses (Socialized Loss; SL).
Instead of closing positions individually, the SL distributes the losses on all users only if the deficit persists. It is :
Conditional : activated only if necessary
Reversible : canceled if the market recovers
Predictable and fair : same ratio for all
More compatible with blockchain : simple calculation, no ADL queue
21M's opinion ⭕
21M ⭕ is the community of crypto investors behind the Club 25%a strategy designed to generate a stable return of 25% per year, for 10 years, without active trading.
The incident of October 10, 2025 on Hyperliquid has become a case study of risk management on crypto exchanges : It shows the limits of an ADL system designed for centralized, simple and isolated markets. It illustrates the need for transparent cross-margin architecturescapable of maintaining the consistency of a complete portfolio. And it fuels the debate between protocol protection (@notnotstorm), user experience (@seanlippel,) et systemic integrity (@fiddybps1). This episode perfectly illustrates the limits of directional and leveraged trading : Even experienced traders, operating with sophisticated long/short models, have found themselves trapped in an opaque and arbitrary mechanism.
Chez 21M⭕we defend another vision of performance: one that does not rely nor on speculation, nor on directional tradingbut on measurable, transparent and decentralized performance mechanisms. We do not promise instant profits, but stable, rational and controlled growth of capitalmonth after month.
Click here to find out for free what we do
It is this philosophy that guides all our decisions: zero trading, zero stress, 100% real and liquid return.