Two months after the October 10 crypto market meltdown that noticed $19 billion in positions liquidated, Gauntlet CEO Tarun Chitra has claimed that the favored automated deleveraging (ADL) mechanism led to HyperLiquid's large losses.
In a prolonged publish on X, Chitra mentioned over $650 million was robotically deleveraged from worthwhile dealer positions. That quantity, he claims, was 28 instances the potential dangerous debt going through exchanges that used ADL.
This “bloodbath of harmless individuals” might be averted with a brand new ADL algorithm, the hooked up 95-page report explains.
learn extra: Did Binance’s USDe “Depeg” value the change thousands and thousands of {dollars}?
Computerized deleveraging with autopilot
Chitra describes ADL as a “final resort” that applies “haircuts” to merchants who’re making income to “cowl dangerous money owed in bancrupt positions”.
The last decade-old “queue” algorithm is extensively used on perpetual futures platforms corresponding to Binance, Hyperliquid, and Lighter.
Nevertheless, beneath excessive market situations, repeated activation of the ADL may causeGrasping queuing methods fail fully”
The technique allocates “haircuts” based mostly on income and leverage, which concentrates losses to the largest winners whereas exceeding the quantity wanted to liquidate, Chitra mentioned.
He proposes a “risk-aware proration” algorithm that allocates ADL based mostly on the leverage of every place.
This publish acknowledges that there isn’t any excellent (ADL) technique. Nevertheless, the so-called ADL trilema The brand new method seems to considerably outperform Queue when it comes to (solvency, fairness, and income) based mostly on Hyperliquid information as of October tenth.
Chitra concludes by calling for additional innovation within the design of algorithmic liquidation: “ADL was invented in 2015 as a Band-Support. We haven't even began exploring the design area but.”
learn extra: Aster vs Hyperliquid: Contained in the high-stakes DEX struggle
hyperlivid
In response to Chitra's publish, HyperLiquid's Jeff Yang quipped, “Those that can do it, do it. Those that can't, rattling it.”
Nevertheless, relatively than straight responding to claims of inefficient automated deleveraging, he disputes the outline of ADL's relationship with HyperLiquid's HLP Insurance coverage Fund.
He accused Chitra of “spreading lies cloaked in fancy ML jargon to sound good”.
Different Hyperliquid supporters chimed in, stating apparent inaccuracies and bias from investing in opponents.
learn extra: ZKasino Exploiter liquidated $27 million in hyperliquid buying and selling
Within the wake of the October 10 crash, Yang argued: “ADL generated a whole lot of thousands and thousands of {dollars} in web income for its customers.” Shut worthwhile brief positions at favorable costs”
He emphasised that the platform's ADL queue incorporates “each used P&L and unrealized P&L” and thanked customers for his or her suggestions. He additionally talked about finding out “whether or not important enhancements might be made which might be price including extra complexity.”
