[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.] #  @tarunchitra Tarun Chitra Tarun Chitra posts on X about hyperliquid, exchanges, binance, decentralized the most. They currently have XXXXXX followers and XX posts still getting attention that total XXXXX engagements in the last XX hours. ### Engagements: XXXXX [#](/creator/twitter::53836928/interactions)  - X Week XXXXXXX +8,750% - X Month XXXXXXX +305% - X Months XXXXXXX -XX% - X Year XXXXXXXXX -XXXX% ### Mentions: XX [#](/creator/twitter::53836928/posts_active)  - X Week XX +400% - X Month XX +7.70% - X Months XXX -XX% - X Year XXX -XX% ### Followers: XXXXXX [#](/creator/twitter::53836928/followers)  - X Week XXXXXX +0.63% - X Month XXXXXX +1.20% - X Months XXXXXX +6.90% - X Year XXXXXX +18% ### CreatorRank: XXXXXXX [#](/creator/twitter::53836928/influencer_rank)  ### Social Influence **Social category influence** [exchanges](/list/exchanges) #919 [finance](/list/finance) XX% [cryptocurrencies](/list/cryptocurrencies) XX% [technology brands](/list/technology-brands) X% [nfts](/list/nfts) X% **Social topic influence** [hyperliquid](/topic/hyperliquid) #87, [exchanges](/topic/exchanges) #621, [binance](/topic/binance) #278, [decentralized](/topic/decentralized) #391, [$650m](/topic/$650m) #5, [bitmex](/topic/bitmex) #17, [debt](/topic/debt) #1682, [balance sheet](/topic/balance-sheet) #1659, [cascade](/topic/cascade) 5%, [crypto](/topic/crypto) X% **Top accounts mentioned or mentioned by** [@danrobinson](/creator/undefined) [@conejocapital](/creator/undefined) [@jlldavies](/creator/undefined) [@hyperliquidx](/creator/undefined) [@d2_finance](/creator/undefined) [@0xkilgore](/creator/undefined) [@cryptohayes](/creator/undefined) [@0xnagu](/creator/undefined) [@binance](/creator/undefined) [@zacxbt](/creator/undefined) [@8bitbytebit](/creator/undefined) [@freecitywarlock](/creator/undefined) [@nad_roid](/creator/undefined) [@velmeshev](/creator/undefined) [@hishboy](/creator/undefined) [@shuri2060_defi](/creator/undefined) [@paradex](/creator/undefined) [@bitmex](/creator/undefined) [@sychype](/creator/undefined) [@0x_vikt0r](/creator/undefined) **Top assets mentioned** [Numeraire (NMR)](/topic/numeraire) [Art Gobblers (artgobblers)](/topic/art-gobblers) ### Top Social Posts Top posts by engagements in the last XX hours "Why is this algorithm bad especially on decentralized exchanges X. If ADL is used repeatedly overshoot (much like a liquidation cascade) X. Not Sybil resistant can split positions to avoid being ADL'd On 10/10 the overshoot was ($650M)" [X Link](https://x.com/tarunchitra/status/1998451873196724407) 2025-12-09T17:57Z 78.3K followers, 4342 engagements "Why do they use this algorithm As @CryptoHayes details in his latest post Huobi first created this algorithm as a heuristic for dated crypto futures BitMEX later repurposed it for perpetuals futures and since then most exchanges copy-pasta'd it" [X Link](https://x.com/tarunchitra/status/1998451856163614733) 2025-12-09T17:57Z 78.3K followers, 4349 engagements "As another large HYPE holder accused me yesterday: "show me the tweet/paper and I'll show you the bags" There are a few things here that are worth pointing out despite everyone FUDing not reading the paper and making conclusions based on the post (skill issue illiteracy)" [X Link](https://x.com/tarunchitra/status/1998985133303701673) 2025-12-11T05:16Z 78.3K followers, 170.4K engagements "In this figure I keep the square shaded to handle the case that the water-filling is correct (e.g. it only liquidates the correct amount of PNL) I was unable to find that in the data if there is proof they were partially closed that'd be awesome" [X Link](https://x.com/tarunchitra/status/1998985136323592277) 2025-12-11T05:16Z 78.3K followers, 21.1K engagements "@0xnagu @danrobinson Great analysis will work on updating it in the coming weeks with some of your suggestions" [X Link](https://x.com/tarunchitra/status/1999346031021232135) 2025-12-12T05:10Z 78.3K followers, 3876 engagements "👀 I wonder if this means the X or so proximity gap papers will have Lean proofs by the end of the year as well" [X Link](https://x.com/tarunchitra/status/1996946715182600381) 2025-12-05T14:16Z 78.2K followers, 5677 engagements "Also by the way pro-rata (which @paradex calls socialized loss) *is actually* an ADL algorithm despite what you might read This becomes clear once you see it via the lens of water-filling algorithms its just one that makes different trade-offs than Queue" [X Link](https://x.com/tarunchitra/status/1998451989207011343) 2025-12-09T17:57Z 78.2K followers, 4519 engagements "What about central clearing This is where the exchange holds the bad debt on their balance sheet. It has the positive of not liquidating your best customers so revenue is preserved. But it is fundamentally not : as the exchange grows its ability to successfully preserves solvency decays @cryptohayes documents that exchanges doing this faced an enormous amount of adverse selection and and blew up" [X Link](https://x.com/tarunchitra/status/1998452010325258521) 2025-12-09T17:57Z 78.2K followers, 3177 engagements "We show that one can construct an infinite set of easy to compute risk-aware strategies that non-uniformly haircut winners (see below) and both theoretically and practically outperform Queue and Pro-Rata in terms of solvency fairness and revenue" [X Link](https://x.com/tarunchitra/status/1998452074888179955) 2025-12-09T17:57Z 78.2K followers, 3003 engagements "Moreover we backtest a number of these strategies against the real data from @HyperliquidX on 10/10 and show that virtually all of these algorithms would have $" [X Link](https://x.com/tarunchitra/status/1998452091308958205) 2025-12-09T17:58Z 78.2K followers, 2976 engagements "We also show that these algorithms outperform on fairness and revenue The trilemma implies that they have to trade-off solvency for this(e.g. the exchange has to hold some bad debt on its balance sheet) But the magnitude of loss is 10x smaller with optimized ADL algorithms" [X Link](https://x.com/tarunchitra/status/1998452110950810016) 2025-12-09T17:58Z 78.2K followers, 2946 engagements "Did @HyperliquidX autodeleverage (ADL) $650m of PNL that it didnt have to Was this 28x more than the minimal necessary Did almost every exchange (incl. @binance) copy-pasta a Huobi heuristic from 2015 Can we do better in 2026 (+ a new paper)" [X Link](https://x.com/tarunchitra/status/1998451762232177041) 2025-12-09T17:56Z 78.3K followers, 297.9K engagements "What water-filling algorithm do @HyperliquidX and @binance use Greedy Queue algorithm: X. Rank traders by PNL % x Leverage X. Fully close positions until the deficit is covered This algorithm (e.g. liquidates more PNL than the deficit requires)" [X Link](https://x.com/tarunchitra/status/1998451843295580522) 2025-12-09T17:57Z 78.3K followers, 4698 engagements "If the perpetuals exchange were decentralized I should be able to see the code and verify this It is not public for Hyperliquid Decentralized exchange should have public verifiable ADL policies" [X Link](https://x.com/tarunchitra/status/1998985138789838914) 2025-12-11T05:16Z 78.3K followers, 42K engagements "Drift's implementation of pro-rata is public should you want to go into minutiae without making blanket statements about why equity weighting is wrong (w/o proposing *why* position size is better especially in fairness terms)" [X Link](https://x.com/tarunchitra/status/1998985141176381793) 2025-12-11T05:16Z 78.3K followers, 23.4K engagements "1) Youve written more to describe this case than what the Hyperliquid documentation which is case in point about how having to interpret what ADL is *actually* executed (in numeraire terms or not) is a flaw of the system 2) I agree that if there are two queues (one for seizing positions / PNL and one for executing/crossing as you describe) that interact with each other then you have to include that in the accounting. But there is no documentation explicitly saying that and no code showing you how that would happen; so the paper only focuses on the closures as thats what you can verify from" [X Link](https://x.com/tarunchitra/status/1999013977150411255) 2025-12-11T07:10Z 78.3K followers, 8198 engagements "*Three* rules: Sybil resistant is the most important one (and Queue is not Sybil resistant; thats fine for binance since they have identity not for hyperliquid) The other part you seem to miss is that Queue is fine in isolation (if you look at most single events it is fine). But there is compounding of loss when ADL is repeatedly used versus with pro-rata (or any more distributive method) The Queue haircuts are of the form (1 X c X 0) whereas pro-rata is c * (1 1). Iterating the loss measurement (e.g. the value of overshoot conditional on a haircut vector) on the first vector is inherently" [X Link](https://x.com/tarunchitra/status/1999026473294835987) 2025-12-11T08:00Z 78.3K followers, 8608 engagements "Yeah thats fair; it still has a gap to pro rata though and a rounding error since the contract vector c is fixed and integral. The two algorithms are approximately equal if c_i is flat / approximately constant. This entire conversation is proving the point that the lack of real documentation and open source code makes this system impossible to trust fully if I cant even reason about the gap to pro-rata. The gap controls whether there is any compounding of the deficit on multiple ADL events" [X Link](https://x.com/tarunchitra/status/1999095210282828021) 2025-12-11T12:33Z 78.3K followers, 8330 engagements "OH: It must be annoying to have the guy who coinvented art gobblers constantly dunking on you" [X Link](https://x.com/tarunchitra/status/1999465759937020099) 2025-12-12T13:06Z 78.3K followers, 51.1K engagements
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
@tarunchitra Tarun Chitra
Tarun Chitra posts on X about hyperliquid, exchanges, binance, decentralized the most. They currently have XXXXXX followers and XX posts still getting attention that total XXXXX engagements in the last XX hours.
Engagements: XXXXX #
- X Week XXXXXXX +8,750%
- X Month XXXXXXX +305%
- X Months XXXXXXX -XX%
- X Year XXXXXXXXX -XXXX%
Mentions: XX #
- X Week XX +400%
- X Month XX +7.70%
- X Months XXX -XX%
- X Year XXX -XX%
Followers: XXXXXX #
- X Week XXXXXX +0.63%
- X Month XXXXXX +1.20%
- X Months XXXXXX +6.90%
- X Year XXXXXX +18%
CreatorRank: XXXXXXX #
Social Influence
Social category influence exchanges #919 finance XX% cryptocurrencies XX% technology brands X% nfts X%
Social topic influence hyperliquid #87, exchanges #621, binance #278, decentralized #391, $650m #5, bitmex #17, debt #1682, balance sheet #1659, cascade 5%, crypto X%
Top accounts mentioned or mentioned by @danrobinson @conejocapital @jlldavies @hyperliquidx @d2_finance @0xkilgore @cryptohayes @0xnagu @binance @zacxbt @8bitbytebit @freecitywarlock @nad_roid @velmeshev @hishboy @shuri2060_defi @paradex @bitmex @sychype @0x_vikt0r
Top assets mentioned Numeraire (NMR) Art Gobblers (artgobblers)
Top Social Posts
Top posts by engagements in the last XX hours
"Why is this algorithm bad especially on decentralized exchanges X. If ADL is used repeatedly overshoot (much like a liquidation cascade) X. Not Sybil resistant can split positions to avoid being ADL'd On 10/10 the overshoot was ($650M)"
X Link 2025-12-09T17:57Z 78.3K followers, 4342 engagements
"Why do they use this algorithm As @CryptoHayes details in his latest post Huobi first created this algorithm as a heuristic for dated crypto futures BitMEX later repurposed it for perpetuals futures and since then most exchanges copy-pasta'd it"
X Link 2025-12-09T17:57Z 78.3K followers, 4349 engagements
"As another large HYPE holder accused me yesterday: "show me the tweet/paper and I'll show you the bags" There are a few things here that are worth pointing out despite everyone FUDing not reading the paper and making conclusions based on the post (skill issue illiteracy)"
X Link 2025-12-11T05:16Z 78.3K followers, 170.4K engagements
"In this figure I keep the square shaded to handle the case that the water-filling is correct (e.g. it only liquidates the correct amount of PNL) I was unable to find that in the data if there is proof they were partially closed that'd be awesome"
X Link 2025-12-11T05:16Z 78.3K followers, 21.1K engagements
"@0xnagu @danrobinson Great analysis will work on updating it in the coming weeks with some of your suggestions"
X Link 2025-12-12T05:10Z 78.3K followers, 3876 engagements
"👀 I wonder if this means the X or so proximity gap papers will have Lean proofs by the end of the year as well"
X Link 2025-12-05T14:16Z 78.2K followers, 5677 engagements
"Also by the way pro-rata (which @paradex calls socialized loss) is actually an ADL algorithm despite what you might read This becomes clear once you see it via the lens of water-filling algorithms its just one that makes different trade-offs than Queue"
X Link 2025-12-09T17:57Z 78.2K followers, 4519 engagements
"What about central clearing This is where the exchange holds the bad debt on their balance sheet. It has the positive of not liquidating your best customers so revenue is preserved. But it is fundamentally not : as the exchange grows its ability to successfully preserves solvency decays @cryptohayes documents that exchanges doing this faced an enormous amount of adverse selection and and blew up"
X Link 2025-12-09T17:57Z 78.2K followers, 3177 engagements
"We show that one can construct an infinite set of easy to compute risk-aware strategies that non-uniformly haircut winners (see below) and both theoretically and practically outperform Queue and Pro-Rata in terms of solvency fairness and revenue"
X Link 2025-12-09T17:57Z 78.2K followers, 3003 engagements
"Moreover we backtest a number of these strategies against the real data from @HyperliquidX on 10/10 and show that virtually all of these algorithms would have $"
X Link 2025-12-09T17:58Z 78.2K followers, 2976 engagements
"We also show that these algorithms outperform on fairness and revenue The trilemma implies that they have to trade-off solvency for this(e.g. the exchange has to hold some bad debt on its balance sheet) But the magnitude of loss is 10x smaller with optimized ADL algorithms"
X Link 2025-12-09T17:58Z 78.2K followers, 2946 engagements
"Did @HyperliquidX autodeleverage (ADL) $650m of PNL that it didnt have to Was this 28x more than the minimal necessary Did almost every exchange (incl. @binance) copy-pasta a Huobi heuristic from 2015 Can we do better in 2026 (+ a new paper)"
X Link 2025-12-09T17:56Z 78.3K followers, 297.9K engagements
"What water-filling algorithm do @HyperliquidX and @binance use Greedy Queue algorithm: X. Rank traders by PNL % x Leverage X. Fully close positions until the deficit is covered This algorithm (e.g. liquidates more PNL than the deficit requires)"
X Link 2025-12-09T17:57Z 78.3K followers, 4698 engagements
"If the perpetuals exchange were decentralized I should be able to see the code and verify this It is not public for Hyperliquid Decentralized exchange should have public verifiable ADL policies"
X Link 2025-12-11T05:16Z 78.3K followers, 42K engagements
"Drift's implementation of pro-rata is public should you want to go into minutiae without making blanket statements about why equity weighting is wrong (w/o proposing why position size is better especially in fairness terms)"
X Link 2025-12-11T05:16Z 78.3K followers, 23.4K engagements
"1) Youve written more to describe this case than what the Hyperliquid documentation which is case in point about how having to interpret what ADL is actually executed (in numeraire terms or not) is a flaw of the system 2) I agree that if there are two queues (one for seizing positions / PNL and one for executing/crossing as you describe) that interact with each other then you have to include that in the accounting. But there is no documentation explicitly saying that and no code showing you how that would happen; so the paper only focuses on the closures as thats what you can verify from"
X Link 2025-12-11T07:10Z 78.3K followers, 8198 engagements
"Three rules: Sybil resistant is the most important one (and Queue is not Sybil resistant; thats fine for binance since they have identity not for hyperliquid) The other part you seem to miss is that Queue is fine in isolation (if you look at most single events it is fine). But there is compounding of loss when ADL is repeatedly used versus with pro-rata (or any more distributive method) The Queue haircuts are of the form (1 X c X 0) whereas pro-rata is c * (1 1). Iterating the loss measurement (e.g. the value of overshoot conditional on a haircut vector) on the first vector is inherently"
X Link 2025-12-11T08:00Z 78.3K followers, 8608 engagements
"Yeah thats fair; it still has a gap to pro rata though and a rounding error since the contract vector c is fixed and integral. The two algorithms are approximately equal if c_i is flat / approximately constant. This entire conversation is proving the point that the lack of real documentation and open source code makes this system impossible to trust fully if I cant even reason about the gap to pro-rata. The gap controls whether there is any compounding of the deficit on multiple ADL events"
X Link 2025-12-11T12:33Z 78.3K followers, 8330 engagements
"OH: It must be annoying to have the guy who coinvented art gobblers constantly dunking on you"
X Link 2025-12-12T13:06Z 78.3K followers, 51.1K engagements