Original author: AZEx community
Reprint: Daisy, Mars Finance
1. Historical review: What exactly happened?
On August 26th, XPL experienced a few minutes of 'roller coaster' on Hyperliquid:
05:36 A huge buy order sweeps the order book, with individual transaction sizes ranging from tens of thousands to hundreds of thousands of dollars, quickly pushing up the price of XPL.
05:36–05:55 The mark price rises sharply due to internal matching dominating, far exceeding CEX external market references, causing a large number of short positions to fall below the maintenance margin. The system initiates liquidation: liquidation orders are directly sent to the order book, forming a positive feedback cycle of 'sweep book → liquidation → re-sweep book', continuously pushing up the price of XPL.
05:55 The price skyrockets to a peak, increasing nearly +200% within a few minutes, while whale accounts take profits, with single-minute profits exceeding 16 million dollars. Some short accounts were liquidated for millions of dollars within minutes.
05:56 The market depth recovers, and prices quickly fall back, the XPL contract market returns to 'normal', but many short accounts have already lost everything. Almost simultaneously, the ETH perpetual price on the Lighter platform also experienced a pin, briefly hitting 5,100 dollars.
This indicates: This is not just a problem of a single platform, but the concentrated exposure of structural risks in the entire DeFi perpetual contract.
2. What have these situations caused?
Whales make huge profits, while shorts face huge losses. Low-leverage hedgers were also caught.
Many people think that 1x leverage hedging equals 'no risk'. But in this incident, even 1x leveraged short positions with substantial collateral were liquidated during the pin, resulting in millions of dollars in losses. This led many users to conclude: 'I will not touch such isolated markets in the future.' But the truth is much more complex.
3. Core issue: Structural flaws in the order book model
After the XPL incident, many discussions focused on 'single oracle dependency' or 'lack of position limits'. But these did not grasp the core of the problem.
The Perp protocol itself has multiple implementation paths:
Orderbook (order book driven)
Peer-to-Pool (pool against counterparty)
And the mixed form of AMM/Hybrid
The issue today lies in the order book type implementation. Its structural flaw is:
Effective depth and chip distribution
1. The order book appears deep, but the actual effective depth it can bear depends on chip distribution.
2. When chips are concentrated in the hands of a few major players, even pushing a few points can trigger a chain reaction.
Price anchoring relies on internal market transactions
1. In weak markets, order book transactions will directly dominate the mark price.
2. Even with oracles, as long as external spot anchors are not strong enough, this dependency is a weakness.
Liquidations and order books form a positive feedback loop.
1. Liquidation orders themselves need to enter the order book → further push prices → trigger more liquidations.
2. In thinly traded markets, this is an 'inevitable stampede', not a random accident.
As for measures like 'setting position limits for individual users', they are actually meaningless. Because positions can be split into multiple sub-accounts or wallets, the risk at the market level still exists. Therefore, the pin is not due to manipulation by bad actors, but the fate of the order book mechanism under low liquidity conditions.
4. Return to the essence: What exactly is perpetual contract solving?
When you say, 'I want to go long on ETH', what is actually happening behind the scenes?
- If it is a spot transaction, you put up 1000 U to buy ETH; when it goes up, you profit, and when it goes down, you lose.
- If it is a perpetual contract, you put up 1000 U margin and can open a 10x long position, leveraging a position of 10,000 U, amplifying returns while also amplifying risks.
Here we need to ask two key questions:
Where does the money come from?
Your profits inevitably come from the opposing positions (short sellers) or the liquidity pool provided by LPs.
Who determines the price?
Traditional markets: Order book transactions directly reflect prices; the more you buy, the higher the price goes, which is the feedback mechanism of market trends.
On-chain perpetual: Most protocols (like GMX) do not have their own matching books but rely on CEX oracle prices.
5. The problem of oracle models
Oracle prices usually come from CEX spot transactions, which means that on-chain transaction volumes cannot feed back into prices.
Although oracles have delays, the more fundamental problem is:
You opened a position of 100 million U on-chain, but there was no corresponding trading volume in external spots.
In other words, on-chain trading demand cannot influence prices in return, and risks are 'accumulated' in the system.
This is exactly the opposite of the order book model: Order book price feedback is too fast, making it easy to manipulate; oracle price feedback is delayed, making risks easy to release late.
6. Basis and funding rates
This brings another key question: How to correct the price difference (basis) between spot and contract?
In traditional markets, if there are far more bullish than bearish participants, contract prices will be higher than spot prices.
Perpetual contracts introduce funding rate mechanisms to adjust:
Too many longs → funding rates turn positive, longs have to pay shorts;
Too many shorts → funding rates turn negative, shorts have to pay longs.
In theory, funding rates can anchor contract prices back to spot.
But in on-chain perp, the situation is more complicated: If the depth of the spot market is insufficient, even a high funding rate may not correct the basis. Especially for obscure varieties, on-chain contracts may long deviate from the spot, becoming an almost independent 'shadow market'.
7. The illusion of on-chain depth
Many people think that only obscure varieties are easy to manipulate, and leading assets won't have problems. But the reality is: The true depth of on-chain spots is far less than imagined.
Take the top three tokens of each ecosystem as an example:
- In Arbitrum, the depth of mainstream tokens other than ETH often only amounts to a few million dollars within a 0.5% price difference range.
- On leading DEXs like Uniswap, even ecological tokens like UNI do not have sufficient on-chain spot depth to support sudden impacts in the scale of tens of millions of dollars.
What does this mean?
Effective depth is often far lower than nominal depth, especially when chips are concentrated, the actual bearing capacity is even weaker.
In such an environment, the threshold for price manipulation is not high. Even the top three tokens in the ecosystem can be easily pushed up or down in extreme market conditions.
In other words: The structural risk of on-chain perp is not a 'special case' of niche markets, but a 'norm' of the entire ecosystem.
8. The direction of new generation protocol design
From this XPL pin event, we see more clearly: The problem is not a vulnerability of a specific platform, but the structural contradiction between the existing order book and on-chain liquidity.
Therefore, to discuss 'next generation Perp protocols', at least three directions are worth delving into:
1. Risk control in advance: Before every opening, swap, adding or reducing liquidity, and switching positions, the market health after execution should be simulated first. If the risk exceeds the threshold, limit or adjust in advance, rather than waiting for positions to fall below the maintenance margin before being passively liquidated.
2. Spot pool linkage: The main current models on-chain either feedback too quickly (order book) or have delayed feedback (oracles). A better direction is to link contract positions with spot pools, buffering or diluting during risk accumulation through changes in spot market depth. This avoids delays and reduces instantaneous stampedes.
3. LP priority protection: Whether in order book or Peer-to-Pool, LPs are the most vulnerable link. Next generation protocols need to write LP's risk control mechanisms into the protocol layer, making LP risk transparent and controllable, rather than passively absorbing losses in the end.
9. Exploration and opportunities in practice
It's easy to talk about direction, but it's really hard to land it.
But some new attempts are already happening:
Risk control in advance: Before executing a trade, simulate market health to filter out risks in advance.
Linkage between contracts and spot pools: Allow positions to feedback with spot liquidity, avoiding risk accumulation or instantaneous stampedes.
LP priority protection: Write LP risk control into the protocol layer, rather than making LP passively absorb losses.
At the same time, we cannot ignore a larger market fact:
The perpetual contract market generates over 30 billion dollars in fees and dividends each year. In the past, this cake was mostly divided among a few centralized exchanges and professional market makers. If the next generation of protocols can combine AMM technology to decompose 'market-making' into pooled liquidity provision, then more ordinary participants can share in this market dividend. This is not just an innovation in risk control but also a reconstruction of incentive mechanisms.
In these explorations, some new projects are also beginning to try different paths. For example, AZEx is attempting to combine 'pre-execution risk control + dynamic funding rates + market freeze in extreme cases' with 'LP pooled profit sharing' based on the Uniswap v4 Hook mechanism.
Next week, AZEx will open the test network, interested readers can get the latest progress through [https://x.com/azex_io].
10. Conclusion
The XPL pin event reminds us: Risks are not in the charts but in the protocols.
Today's DeFi perpetual contracts are mostly order book driven. As long as liquidity is insufficient and chips are concentrated, similar stories are bound to repeat.
The real competition of the next generation Perp protocol is not UI, points, or rebates, but whether we can design a new Perp protocol that forms a closed loop of 'price discovery, risk control, LP protection', instead of repeating stampedes in extreme market conditions. Can we redistribute the 30 billion dollars market profit back from a few people to more participants?
Next generation protocols not only need to solve risk issues but also need to redistribute dividends. Those who can achieve both will have the opportunity to define the next generation of the DeFi perpetual contract market.
