Wintermute founder's podcast: The market needs to introduce a circuit breaker mechanism, and there will be no copycat market in the short term

This article is from: The Block

Original compilation｜Odaily Planet Daily Azuma

Editor's note: On October 11, the cryptocurrency market experienced an epic crash (see " The Terrifying Night of 1011: The Crypto Market Plummets, $20 Billion Vanishes"). Although nearly a week has passed, discussions about the reasons for the crash and its subsequent impacts continue unabated.

On October 15, Evgeny Gaevoy, founder and CEO of Wintermute, the industry's leading market maker (which was rumored to have faced a crisis on the day of the crash but has since refuted those claims), participated in The Block's podcast to share his views on the "1011" incident.

The following are the main points from the podcast, compiled by Odaily Planet Daily, with some content edited for readability.

A Completely Chaotic Hour

• Host: Let's get straight to the point. What happened on October 11 was shocking for the entire market. Can you take us back and review what exactly happened that day? What triggered the crash? How did Wintermute respond in such a situation?

Evgeny Gaevoy: To be honest, we still need more time to fully understand the ins and outs of this crash, but one thing is clear— the trigger seems to have been a series of news related to Trump, which gradually led to the largest liquidation event in the history of cryptocurrency.

That day was extremely unusual for everyone—not just for ordinary traders, but also for market makers. The market was completely chaotic within an hour.

We will later discuss the ADL mechanism (refer to "Detailed Explanation of the ADL Mechanism in Perpetual Contracts: Why Does Your Profitable Position Get Liquidated Automatically?"), and we will also talk about how this crash differed from past market fluctuations. One thing is certain: this day was very difficult and unprecedented for many.

It's Still Unclear Who Suffered the Most Losses, Perhaps the Hedge Funds

• Host: Public statistics show that about $19 billion was liquidated that day, but since Binance did not fully disclose the data (the system can only display one liquidation event per second), the actual number could be much larger, at least between $25 billion and $30 billion. This means that the recent liquidation scale was more than five times larger than the previous second-largest liquidation event. Why is this? Is it due to excessive leverage in the system? Or did some key infrastructure fail, preventing market makers like you from intervening in time to stop the chain reaction of liquidations?*

Evgeny Gaevoy: I believe this is the result of multiple factors. On one hand, there was indeed more leverage in the system; on the other hand, the market had more types of tokens, more perpetual contract products, and more large platforms trading these perpetual contracts. Looking back three or four years ago, we didn't have so many perpetual contract products with huge open interest and hidden risks of massive liquidation. In terms of market maturity, while it is indeed more refined and sophisticated than in the past, this development has also given rise to many problems.

Currently, we still do not know who exactly "blew up" and who suffered the most losses, but I suspect that many of the severely affected institutions were actually running long-short hedging strategies, for example, they might have shorted Bitcoin while going long on certain altcoins, thinking this would hedge their risks, only to be "slapped in the face" by the ADL mechanism.

Moreover, when the market experiences extreme declines, various trading paths often get stuck. This is particularly troublesome for market makers. For instance, if you buy on Binance and sell on Coinbase, you might find that your stablecoins on Coinbase are increasing while you have taken on tokens on Binance—only to discover that withdrawals on both sides are blocked, making it impossible to transfer assets.

So, when people say "market makers are exiting the market and unwilling to provide liquidity," it is often not "unwilling," but rather "impossible"—they cannot quote here or place orders there because the assets cannot move at all. This situation occurs not only on centralized exchanges (CEX) but also in DeFi. This is the most troublesome issue—you simply cannot adjust positions across platforms.

The Lack of Transparency in ADL Causes Chaos

• Host: You mentioned ADL (Auto-Deleveraging), and I guess about 90% of cryptocurrency users are hearing this term for the first time. Can you explain how ADL works and why it caused so much chaos in this incident? Also, what impact does it have on market efficiency when market makers cannot be active on multiple exchanges simultaneously?

Evgeny Gaevoy: ADL (Auto-Deleveraging) is essentially the "last line of defense" mechanism for exchanges. Generally, when the margin for your perpetual contract position is insufficient, the exchange will directly liquidate you in the market; if liquidation fails, the insurance fund should theoretically bear the loss.

The ADL mechanism is usually not triggered at all; many exchanges haven't used it for years. It was designed as a "last resort" measure. In extreme cases, such as a massive decline and chain liquidations like on October 11, if the exchange continues to forcibly liquidate through the order book, prices could drop to "zero," and the entire exchange would become insolvent, while the short sellers would profit immensely. Therefore, exchanges try to use ADL to forcibly offset some short positions, which is akin to artificially matching shorts with liquidated longs to create a "virtual offset" and prevent prices from collapsing completely.

In theory, this is an "elegant" solution, but the premise is that execution must be orderly, which was clearly very chaotic this time. The biggest question is—how is the execution price of ADL determined? This will become the focus of many trading institutions questioning exchanges in the coming days or even weeks.

Many institutions had their positions liquidated at extremely unreasonable prices. For example, in our case, some ADL prices were completely illogical; the market price was $1, while our short position was forcibly liquidated by the system at $5. This cannot hedge at all and results in an instant loss.

Is There an "ADL Exemption" Privilege?

• Host: As far as I know, Ethena has signed ADL exemption clauses with certain exchanges. Can large market makers like you obtain similar protections? Why does Ethena have such special treatment?

Evgeny Gaevoy: First of all, I am not entirely sure whether Ethena enjoys this privilege. Additionally, it should be noted that Ethena mainly trades BTC and ETH, which are mainstream coins that rarely trigger ADL. ADL is more relevant to various altcoins and meme coins. If such protective mechanisms exist, we would certainly welcome them.

But should exchanges widely offer such clauses? Not necessarily. I believe that if implemented, it must be transparent; investors need to know which open positions enjoy ADL exemption privileges, otherwise, it will create a detrimental market structure. We certainly welcome such protections, but the premise is that a highly transparent disclosure mechanism must be established; otherwise, the so-called privileges are merely conspiracy theories.

Another key point that is rarely discussed is that some exchanges (like Coinbase and Kraken) have implemented market maker liquidity protection plans. The former FTX also had similar designs, allowing market makers to take over positions that were about to be liquidated, bypassing the insurance fund and ADL. This allows the most risk-tolerant market makers to absorb these risks. However, on mainstream platforms that experienced large-scale liquidations this time, such plans were collectively absent. I believe that restarting such plans would greatly improve market resilience.

Should the Market Introduce "Circuit Breakers"?

• Host: There is also a saying circulating on X that part of the reason this liquidation wave was more severe than before is that Hyperliquid is now one of the top three exchanges by open interest, and many of its data points are very transparent—including liquidation price levels, which are completely invisible on centralized exchanges like Binance, OKX, and Bybit. Some people believe this may have made it easier for certain individuals to estimate, "If we just push the price down to here, we can trigger these liquidations," thus artificially triggering a chain reaction of liquidations. In your opinion, could this transparency have actually fueled this liquidation wave, causing some assets to plummet by 90% or more?

Evgeny Gaevoy: I think if Hyperliquid were the only exchange in the world, then this "targeted attack" conspiracy theory might hold more water—meaning that there were indeed people specifically watching Hyperliquid to trigger this chain liquidation. But in reality, a large amount of open interest still exists on exchanges where liquidation points are not visible, so I think the likelihood of this statement being true is low.

I find it more interesting to consider whether Hyperliquid represents the future direction of the industry. In other words, could its mechanism of "making all liquidation points publicly visible" become an industry standard?

Personally, I believe that Hyperliquid should ultimately find a balance between transparency and privacy—currently, the information disclosure is indeed a bit "excessive." One solution is to enhance privacy; another potential solution is to introduce circuit breakers.

This feature is absent in all centralized exchanges. While I can roughly understand the reasons, it should indeed exist. Especially for stable assets or mainstream tokens, when you see it depeg to $0.6, trading should be paused or switched to auction mode, rather than allowing it to drop endlessly.

In traditional financial markets, almost every exchange—whether for stocks, futures, or commodities—has a circuit breaker mechanism. It prevents the underlying asset from plummeting too much in a short time; the system will automatically pause trading or enter bidding mode, or a combination of both. But in the crypto market, no exchange has such a mechanism, which has always puzzled me. If there were a circuit breaker, it could actually protect many retail investors from being liquidated in a chain reaction.

Of course, some might ask—if only one exchange (like Coinbase) adopts a circuit breaker while Binance does not, would that be useful? After all, much of the price discovery actually happens on Binance. In that case, even if Coinbase stops trading, prices will continue to fluctuate on other platforms (including on-chain markets). So if only a single exchange adopts a circuit breaker, its effectiveness may be limited. To be truly effective, most exchanges need to adopt it in coordination.

This is actually a trade-off issue. You have to choose between two risks: allow prices to plummet and liquidate all long positions? Or choose to pause trading to ensure the exchange maintains solvency?

For example, if Bitcoin drops 20% on a certain exchange, as an exchange, you can clearly determine that this is an abnormal fluctuation rather than a fundamental collapse, and it would be reasonable to activate a circuit breaker; but if an altcoin drops 50%, that might be considered normal fluctuation, and the market can clear itself. So circuit breakers should at least be introduced for specific trading pairs or asset types.

Will Exchanges Actively "Pull the Plug"?

• Host: There have been rumors in the past that some exchanges would "pretend to go offline," actually triggering a circuit breaker. For example, during the pandemic crash in 2020, BitMEX went offline during the crash, and the market speculated that they did this to avoid a 99% price drop. Do you think that was a genuine circuit breaker action? Or was it just because their technical infrastructure couldn't handle it? Furthermore, why do exchanges like Binance still go offline now? Knowing there will be a huge wave of trading demand, why haven't they improved?

Evgeny Gaevoy: I tend to believe that the simplest explanation is often the correct one. In my view, the reason is simple—most centralized exchanges have terrible infrastructure that falls far short of the technical standards of traditional financial markets (like the Chicago Exchange, New York Stock Exchange, or NASDAQ). While there are historical reasons for this, no one is really going to migrate to NASDAQ-level technology in the short term.

It is precisely because of such backward technical levels that these platforms often crash under high loads. I think this is a more reasonable explanation than any conspiracy theory. I do not believe that exchanges would deliberately "go offline to liquidate retail investors" to profit from the insurance fund; that would be too risky.

From a business perspective, for exchanges and market makers, keeping retail investors trading continuously, engaging in repeated games, and maintaining long-term retention is far more profitable than "clearing out retail investors every year." Because once everyone is wiped out, many will leave for good and not return.

Will Institutions Blow Up?

• *Host: I still remember the Luna crash; although it wasn't as severe as this one, the impact was profound. It took us about two to three weeks to realize that Three Arrows Capital (3AC) had actually gone bankrupt. This time, the scale of liquidation is 5 to 10 times larger than that. While there are speculations that some market makers, trading firms, and lending institutions have been severely impacted, so far, I haven't heard of any that have completely blown up or gone bankrupt; at most, I've heard of a few trading firms "losing some money." Do you think any institution will be found to have blown up next? After all, the open interest and liquidation scale this time are record-breaking. *

Evgeny Gaevoy: I believe that compared to 2022, the degree of interconnectedness in the market has decreased significantly. Back then, when Three Arrows collapsed, the entire market was directly dragged down by its long positions.

Now, if a market maker really goes bankrupt, you should ask who it will affect? How long is the chain of influence? What everyone is most worried about is actually the "contagion effect." Do you remember how Alameda acted back then? They started dumping assets on DeFi during the rebound, and everyone could see it very clearly.

If a market maker really goes bankrupt, say Wintermute—this is just a hypothetical—what would happen? We have some loans, and those could all become worthless; we also have some market-making contracts with protocols that might still be valid; after bankruptcy, we could theoretically sell off some assets to recover funds, or just run away (joking); additionally, we have settlement counterparties who might have margin with us, such as BTC or ETH.

So, the real impact range mainly includes the protocols served by the market maker and the counterparties that have margin dealings with the market maker. The worst-case scenario is that they sell off their BTC or ETH to cash out, but the impact of this situation is actually quite limited.

If some smaller market makers are truly "drowned," they might sell off specific tokens they are responsible for making markets in, but honestly, doing so usually doesn't help because the liquidity of these tokens is limited, and the sell-off would be too obvious, and the market would immediately notice.

Overall, the contagion this time is much more limited compared to 2022. Back then, for example, Three Arrows lent to Genesis, and Genesis borrowed from Gemini, creating a web of interconnections that led to a series of bankruptcies, while the current system is much cleaner and has better risk isolation.

Lessons Learned from Extreme Market Conditions

• Host: After this incident, do you have any reflections or lessons learned? For example, are there areas for improvement in your response strategies, risk control, or hedging mechanisms?

Evgeny Gaevoy: The challenge with such events is that they may only happen once a year or every two years. You can learn a lot from them, but if you invest a lot of resources specifically to optimize for such "black swan" events, it may not be cost-effective.

Many market makers actually just exited the market this time because such extreme conditions are not suitable for their systems. We are still in the game, but only participating with very limited positions—the inventory issues mentioned earlier restrict our operational space. Although this is not the first time we have encountered such a situation, it was indeed more difficult than before. Adding ADL (Auto-Deleveraging) to the mix made it even trickier.

We have indeed learned something from this experience, which is to handle ADL events better. Although our system reacts quickly and can immediately detect changes in open interest and automatically adjust positions, when you receive 500 ADL emails from Binance, you still have to manage it manually. Of course, you could design a perfect system that automatically trades perfectly in such extreme situations, but that would not be meaningful for the other 364 days of the year, and it is not worth the investment.

We had a meeting this morning to discuss the next steps for improvement. For example, in our quoting system, we have a lot of "circuit breaker" protections internally, but this time they were triggered too frequently, almost every minute. We may allow it to trigger less aggressively in future extreme conditions.

Overall, we are satisfied with our response performance. Although we incurred some losses on ADL, we also made a lot due to high volatility, which balanced out to a decent overall result. Of course, we could do better, but overall, it’s fine.

What’s frustrating is that there is a lot of FUD now, and we spend a lot of time communicating with counterparties and partner protocols, explaining our inventory situation, etc. While this is troublesome, it is also understandable—after all, everyone is quite nervous.

What Are the Future Expectations?

• *Host: So, looking ahead to the next few months, what do you think? This liquidation scale has set a historical record, far exceeding events like FTX and Luna, but this time it seems no one feels it's the "end of the world," just that many people have suffered significant losses. *

Evgeny Gaevoy: I believe the main impact in the coming months will be that sectors outside of major coins will be affected because this wave of liquidations mainly concentrated on altcoins. The number of altcoins and meme coins in the market now is much greater than it was four years ago, and investors' money is scarcer and more cautious, so I think the market heat for altcoins will noticeably decline. Of course, new retail investors are entering the market every day, so the market will eventually recover, but in the short term, there won't be a major "altcoin season."

Interestingly, Bitcoin, Ethereum, and even Solana performed quite steadily this time. For example, Cosmos (ATOM) once dropped by 99.9%, while BTC and ETH only saw a maximum drop of 15%, which is very mild.

Liquidity Will Further Concentrate on BTC, ETH, and SOL

• Host: Does this mean that the market and the assets themselves have matured?

Evgeny Gaevoy: I believe so. Bitcoin is now an institutional-grade asset. There are ETFs, support from MicroStrategy, and infrastructure like CME futures. Ethereum is basically approaching this status, and Solana is also getting closer.

So I am not worried about Bitcoin experiencing a massive flash crash unless something very strange happens, like a quantum computing attack.

This is actually a positive signal, indicating that some mainstream assets are now "safe enough to hold long-term." The more ETFs there are and the broader the access channels, the more limited their volatility will be, which also means you can hold BTC, ETH, and even SOL with more confidence and higher leverage. In the future, we will also see more leverage and liquidity concentrating on these assets.

Wintermute's Emergency Response Mechanism

• Host: Your team is mainly in London, right? Although you also have overseas offices, I guess most of your trading team and core members are in London. The timing of this market event was quite late for London, almost at night or even early morning. So when your traders are asleep, how do you respond to such emergencies? Do you have teams in other countries that can immediately take over? How automated is your trading now? How do you handle these emergencies if they occur at the worst times in your time zone?

Evgeny Gaevoy: Yes, we actually have to thank the "Trump rallies" for this; we have long been accustomed to such situations—extreme market conditions often occur on weekends or late Friday nights in London.

So while this time was still quite shocking, we were actually prepared. Of course, to be honest, this is terrible for the work-life balance of traders, but this is the case in any trading company.

Generally, we divide the work this way—our London office and Singapore office take turns. Around 10 to 11 PM London time, the Singapore team takes over trading, while the London team starts to "relax" slowly, but this time the market event broke out before we had a chance to relax, so everyone got pulled in. That night was indeed particularly busy.

• Host: I spoke with another market maker, and they said they have devices that wake traders up at night if there are severe market fluctuations. I guess that might be because they are not as large as you and don't have a global team to take turns. Do you do that as well? For example, if an asset suddenly drops 15%, would someone be woken up by the system? Or are you now distributed enough that you can sleep soundly?

Evgeny Gaevoy: Basically, I would only be woken up if the situation were really bad and we were losing a lot of money. So actually, I slept quite well that night.

I woke up the next morning at eight to see a bunch of people on Twitter asking, "Are you dead?" and similar messages, and then I started dealing with that FUD. The team hardly slept that night, but I slept fine.

This is also the "luxury" of running a large proprietary trading company—you have enough traders to take turns, and you can sleep soundly, dealing with issues the next day.

So if I were woken up at three in the morning, it would mean something serious had happened. So far, that hasn't happened.

Views on DeFi Performance

• Host: From this incident, did you notice any particular phenomena on the chain (DeFi)? Although your activities in DeFi are not extensive, you do have some operations. Did anything go wrong or surprise you this time? For example, I noticed that Aave performed quite well during this crash, with not many liquidations and a very robust system. Compared to the past, DeFi actually held up quite well this time.

Evgeny Gaevoy: We saw some pretty terrible situations on our side in DeFi. But indeed, we encountered the same issues in DeFi as in CeFi: inventory issues.

Our positions were on Binance, but we couldn't transfer them out, so we sold everything we could in DeFi, and bought everything we could on Binance, but we couldn't transfer assets over and just had to wait for inventory to flow back. Of course, we could also borrow assets to make markets, but that carries a lot of risks and could lead to liquidation. Another way would be to quote different prices for USDC in different markets (like DeFi and Binance) to engage in cross-market arbitrage, but that is also difficult to operate.

Such extreme events happen only once a year; you can't really build a system specifically for them. We saw that most competitors simply stopped DeFi trading during this event, possibly because their risk control circuit breakers were triggered.

I am quite satisfied with our performance. Although we could have made more money, we indeed just ran out of inventory.

About FUD

• Host: Last question—Wintermute (WM) has almost become the "scapegoat" in the crypto community. Any market fluctuations are blamed on you. For example, when people noticed you deposited hundreds of millions into Binance before the crash, they immediately spread rumors that you were dumping, even though I know that was just a delta neutral trade, yet rumors still flew around. Do you still care about this kind of public opinion? Although you don't rely on retail investors or Twitter sentiment, focusing more on LPs (liquidity providers) and partner protocols, how do you personally cope? Do you get angry, or have you become indifferent?

Evgeny Gaevoy: To be honest, I have completely let it go. I just feel sad that some people can be so foolish. They piece together unrelated data fragments to draw absurd conclusions and are quite confident about it.

For example, some people saw that we deposited $700 million into Binance that day and shouted, "Wintermute is going to dump," but they didn't even notice that we also withdrew about the same amount on the same day. These people are essentially betting on retail investors in altcoins—while we happen to be the ones profiting from their foolishness.

So this is a kind of "ecological relationship"—they are shouting nonsense on crypto Twitter, and we profit from their stupidity. It's a bit sad, but if they all became smart, our trading volume might actually decrease.

We Have Always Been Net Long

• Host: Will you consider expanding into businesses beyond market making in the future? Such as proprietary trading, investments, or others?

Evgeny Gaevoy: Actually, we have always been doing some other businesses, but there is a misunderstanding in the outside world that we are shorting all the time. In fact, we have been almost entirely net long.

We have been bullish overall since 2022 or even earlier. We have a venture capital department that has invested in many projects, and as a result, we have acquired a lot of locked tokens. We also hold a large amount of core assets like BTC, ETH, HYPE, and SOL. We cannot dump because that would directly harm our own holdings.

In terms of risk management, we have clear rules: our long positions will not exceed 25% of net assets, so even if the market crashes tomorrow, we will lose at most 25% and will not go bankrupt. We also do not put more than 35% of net assets on a single platform, so even if Binance were to collapse like FTX tomorrow, we would still survive.

This is why we were able to withstand the FTX collapse and the hacker attacks. Unless the top five exchanges disappear simultaneously, we can survive.

Original link