John Gu – Crypto Market Making & The Cold Start Problem (S7E30)

Hey everyone, Corey here. Thanks for tuning into another episode of Flirting with Models. If you're enjoying the show, I'd greatly appreciate it if you'd take a moment to rate, review, and most importantly, share with a friend. A word of mouth is how this podcast grows. And if you'd like to learn more about newfound's platform of Returnstack Mutual Funds, ETFs, and model portfolios, head over to Returnstacks.com. Now on with the show. All right, John, are you ready to go? Yes, sir. All right. Three, two, one. Let's jam. Hello and welcome, everyone. I'm Corey Hofstein and this is Flirting with Models, the podcast that pulls back the curtain to discover the human factor behind the quantitative strategy. Corey Hofstein is the co-founder and chief investment officer of newfound research. Due to industry regulations, he will not discuss any of newfound research funds on this podcast. All opinions expressed by podcast participants are solely their own opinion and do not reflect the opinion of newfound research. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of newfound research may maintain positions and securities discussed in this podcast. For more information, visit thinknewfound.com. My guest today is John Goop, founder and CEO of Caledon, one of the most active market makers in crypto, and a firm that has provided liquidity to more than 200 token launches. John's path runs through MIT, Alpha Simplex, Citadel's principal strategies group, and tower research before landing in Singapore at the dawn of the ICO era, where what started as a trade on the kimchi premium became the foundation for one of the most active liquidity providers in digital assets. The thesis of our conversation is what I'll call the cold start problem. In traditional markets, every newly listed stock arrives with scaffolding already in place, a designated market maker, a reference price, a universe of comparables, and decades of regulatory infrastructure. Crypto inverts that. A new token can launch with no order book, no comparables, and no clear demand curve. Someone has to quote a two-sided market into that void and how they do it shapes whether the asset becomes a real tradeable thing or a graveyard of widespreads and stranded liquidity. John and I dig into how you bootstrap liquidity from zero, how the quoting playbook evolves as a market matures, the economics of token market making contracts, and how that same infrastructure now bridges into structured products and treasury solutions for token foundations. Please enjoy my conversation with John Goop. John, welcome to the show. Thank you for joining me. It's always a special treat for me when I get to record with a guest from halfway around the world. I'm not mistaken, recording from Singapore at the moment, which lining up those time zones for me is never so easy. So I really appreciate your flexibility and you joining me today. Yeah, thanks for having me really appreciate it. I listened to quite a few number of your podcasts, really enjoy them, so it's really honor to be here. Thank you for the kind words. I'm excited to dive in here because this is a bit of a novel conversation for what we've had on the podcast so far, which I love to have seven seasons in and I don't know something like 80 episodes when I can find a guest who can really bring some novel perspective and viewpoints. It's a special opportunity. So let's dive in. Now, I know John from our pre-call, you had an interesting path to get to where you are today. You did your studies at MIT. You spent some time at Alpha Simplex, right in Boston with Andrew Lowe. You were in Citadel's principal strategies group. You were at Tower Research and then ultimately ended up moving to Singapore. Walk me through this arc and maybe talk to me a little bit about how each step along the way has informed the way you think about markets today. So I went to MIT. I actually had a tough decision of whether or not to go because it was quite expensive and I had a false scholarship to the local University of Maryland. But my dad was being the Asian parent that he is. You definitely are going. When I got there, I actually was quite struck by how ahead a lot of people were. So I got there. I decided to study computer science and met these people who had already built their own 3D ray tracing engines and published games and like, wow, this is a whole different world. Among these people, I also met folks who really knew that they wanted to do finance. So they had lined up all the summer internships at the Goldman's and JP Morgan's and I had honestly no idea what I wanted to do. Study computer science really loved learning about all the stuff. So at MIT you can take all the classes that you want. They don't have a limit. And the way I got connected with Andrew is because he did a lot of cross disciplinary work. So I ended up doing my master's thesis with him, which is how I went to Alpha Simplex. But just to give you a sense of where my head was at, I lived on the Boston side and MIT's and Cambridge getting around took quite a bit if you were just walking. So a lot of people would roll or blade. And I remember once going to a Goldman Sachs internship hiring session where I just showed up in my roller blades. I just got in. I was sweaty in my t-shirt and shorts. And I get in and everyone there, all the other students were in suits and is sitting down. So I did an intense question and answer session and I was a psycho man, kind of in the wrong place. That was MIT and I really loved the experience of learning a lot. Ended up doing my master's thesis with Andrew. And afterwards he happened to be working on Alpha Simplex at the time and asked me to interview at the firm. And I did and was made an offer to be a research scientist. That was the title. And the work that I did there was quite cool because it was a small company. I did a lot of research into statistical arbitrage strategies using technical data as was fundamental data like balance sheet information earnings, eventually financial data, as well as CTA type of strategies. So trend following in various futures markets. But I think what really struck me at the time was Andrew was working on what he called the adaptive markets hypothesis. He was actually formulating it around the time that I was working with him. He was describing it to me about how markets are really a living thing. Attention focuses from one thing to another. People and opportunities shift around. And that's really what defines a relative degrees of efficiency in markets. And it's not a static problem. You can't go into like a equilibrium state all at once. So I really like that. And that left an impression on me and sort of won the mental models that I used to think about markets. And then my next stop was that Citadel and Citadel was actually quite a different experience. I worked for what's called principle strategies at the time, which did two things. One was risk arbitrage, the merger, acquisitions. And the other one was a more speculative event space trading. So here events encompass the wide category of things like C-suite turnover, strategic alternatives where a company exploring sales or something else. There I learned about a much wider set of data that's available. And that was super interesting because I built some strategies where, for example, if you look at CEO turnover and look at the performance of the company post CEO turnover and whether it outperforms, it's a very strong indicator of continued outperformance. Because these things, information gets incorporated into markets very slowly. A classic example that I think a lot of people will identify is when Steve Balmer left Microsoft in the CEO role. Nadella really, really did a great job. But the stock price was just amazing. And so I was really happy to see that. And I was really happy to see that. Because I was working and I focused on work. I left power in 2016. Didn't really have a job lined up at the time. But was curious to see what else was out there. And then came across crypto. Because I was talking to a friend. And who became my co-founder, Michael. And he pointed out how there were a lot of different arbitrage opportunities. prices looked very different across different exchanges. There were two trades that I took advantage of initially, just experimenting on my own. One was the Kimi-Pichi premium, where I think crypto prices in Korea traded at somewhere between 20% to 50% premium to exchanges outside of Korea. That was actually very hard to take advantage of, because there were a lot of operational gating factors. There were capital controls. You couldn't really exchange Korean one for dollars, but very easily. Because I was trading a lot of Korean markets, actually, prior to Alpha Lab now called Caledan, I had those contacts, so I reached out to them and say, "Hey, did you know there was this R? Let's collaborate." But the other interesting opportunity was just these very short-term arbitrage opportunities between exchanges. And I built a very simple automated algorithm to take advantage of that. And it did a lot better than I thought it would. I thought it would be a small thing, just a curiosity, and it ended up being quite profitable. I think your insight that my background's kind of organic is very good, because I realized I used to think I was a very structured person, and I think in a structured way, but I'm actually quite intuitive in how I approach things. So I didn't have a plan, and I remember walking with my wife, we were on vacation, we were talking about these trades, and she was like, "Well, look, you can do it as a lifestyle business. It's good, we don't really have to work, we can just relax, or you can build a business." And I was thinking about that, there's no right answer, right? It's really a question of values, and it wasn't that I was far-sighted in seeing the opportunity, 'cause I was almost certain the trades themselves would disappear. But what I thought was nice was I got the opportunity to leverage that to build a business, which is something that I had always wanted to do, and that was what I thought about while I left Tower that I wanted to explore opportunities to start business. Just some background, my parents immigrated to the US in the 1980s from China, and when we got to the US, they barely spoke in English. My dad was an exchange student, my mom worked as a babysitter, even though she was a doctor in China. And over the years, they eventually started their own healthcare IT company, and they grew organically, and at some point, they employed 500, 600 people, and I sort of saw that growing up, and I was always really impressed, and I kind of wanted to emulate that a bit. And then this was an opportunity and launch pad for me to do that. When we talked, I decided, let's go for it, let's try to build something real. That's how Caledang got started. - I love that story about your parents. That could be a whole podcast episode in and of itself. - I'm very proud of them to see where they've gotten to from where they started. It's just unbelievable. It's really amazing. - So I can't go any further without addressing the name Caledang. Obviously, I presume a Dune fan. - Yes, exactly. - Any particular meeting to choosing Caledang over anything else? - So we were originally called Alpha Lab. That's how he started. Back then, it was before a lot of token treasuries and crypto companies started using the labs suffix. We thought, you know, a lab that produces alpha, that was quite nice, but then we started getting confused with a bunch of people, 'cause everyone's like, are you Alpha finance labs? We decided that if we wanted to build a brand of any sort, we needed to come up with a more distinctive name. Caledang is a home planet of the house of the treaties. They're known to be very tough, but fair. And it's also water liquidity, a reference to a water planet. So that's why we thought it would really fit it. Well, - I love it. Well, let's dive into that being fair and liquidity and all those elements sort of leaving nicely into my next question, which is a core part of your business today is being a, for lack of a better phrase, almost designated market maker for token launches. And this is gonna be a big part of the conversation over the next couple of questions. This is really sort of a cold start problem. When you have traditional markets, you do have a designated market maker. You often have some sort of reference or IPO price. You can talk towards, you've got comparables, regulatory scaffolding, in crypto, you know, a new token might launch with no order book depth, no comparables, uncertain demand, no idea who's gonna be trading it, who holds it, whether you have all the information about who's actually gonna be allocated or not. This is something you and your team have now done 200 plus times if I'm not mistaken. And so I was hoping you could walk me through what the sort of the first day in the first week of providing liquidity to a brand new token actually looks like from the market maker seat. - This problem typically you only need to face if you're doing this as a service, as opposed to making a choice to trade it because of all the problems that you mentioned. The second thing I would say is this isn't so different from CIFI. I think it was back during the ICO boom when there was nothing really out there, but the space has matured and it's really a matter of degree. So basically nowadays you do actually have some comparables, you have other tokens that look like it, other tokens that same VCs might have invested in. You do have prior VC rounds to reference. So there's a lot of information that's out there. There's no fundamental value to lean on, like you do when you have a business that's listed. I will remind people that during the dot com bubble there was also no fundamental value to rely on, even though a lot of companies went public. It's based on the story and the future expectation people's excitement and still a lot of the same factors are at play here. In terms of the initial things that you do, I think a lot of it is that informational discovery. So you speak to the project, where do they see their tokens, how do they think about it, how are they communicating with the exchanges? Typically one token lists, they do speak to the exchange and come to an understanding of what reasonable ranges are, and then the exchanges will communicate to the market maker and say, hey, we're opening, we're communicating to folks that we believe it'll be around here. This is where we would like you to place our orders. In the absence of a lot of information, people do anchor to that price. So it's not path independent. Once that anchoring comes, it's a little bit sticky and at just in time. With that in mind, then us and other market makers, we sort of discote in smaller sizes and generally wide. Once trading happens, then you have a lot of trading and it's quite volatile. The thing that we try to do then is try to balance supply and demand. Whenever we trade, we take our inventory risk and then we try to make sure that it's not so imbalanced that it becomes on a canonical for us. So maybe we can just sort of hone in on that exact cold start moment. You don't have a sense of fair value, though maybe in other markets you don't as well, you don't have the historical data. There is to a certain extent a little bit of sort of pricing into the void. And there is that moment where the token goes live that you do need to provide a quote. There's a mid. So maybe you can talk a little bit about the art of coming up with what that initial mid is. - I think the mids are not so important as it is a bit an ask. You can arbitrarily choose a mid and discote really wide. And I think that's what people do. We adjust our width to be wide and then we slowly narrow in. And then we use the feedback from the trading itself to actually tune our mid. There's sort of a lot of different models. What makes this work ultimately though is the fact that we're compensated to do this because the activity itself, the market making activity itself, there are some market makers that actively paint a story with their trading. They try to kind of represent something because they know for example that they might control very large amount of the inventory. We don't do that. It's a pure business decision, not a moral one. People have different risk tolerances. And for us, we just thought it's not worth it. And so for us, we're very passive or price takers. And it's almost always a money losing proposition, which is why the market makers often compensated by the project to be there and to provide the liquidity and the optionality to trade. - One of the evolutions we've seen recently has been the introduction of pre-market perps. And just pre-market trading in general and OTC markets surround points, for example, before tokens start trading live. You have that for some markets, typically the larger markets that are coming, but not for small tokens. How does the calculus change when you talk about having, "Okay, this has a nice liquid pre-market perp "that's trading versus something that doesn't "both your ability to hedge your business risk "as well as your ability to hone in on an initial fair bit ask." You also had that even before pre-market perps and points, which is on-chain, people can actually create tokens and deposit them into on-chain automated market makers. So there have been price discovery before listing on centralized exchanges. What you said makes a lot of sense, but not necessarily because these mechanisms existed before, but because they are associated with bigger projects that have more mind share. So what then happens when these ones list is that it's almost interesting, they still use market makers, but because of the interest, I think the market makers probably play a lesser role because there's more distributed ownership, people are more excited about it, there's more natural organic trading. And so in a lot of senses, those are ones that a lot of market makers might trade for profit. They wouldn't do it necessarily to the KPIs that the projects want, but it does change the calculus in the sense that it's much easier to rely on what going on with the token as a market participant, not just a market maker. You can incorporate that information into your pricing. My naive presumption here would be that the quoting strategy that you apply day one is going to be different than the quoting strategy you apply day 30. That the playbook is going to change as you get more organic flow that has developed and maybe the order in balance is changing particularly for new tokens where there might be a lot of people who are air dropped that are looking to exit and you might end up with a large imbalance of inventory depending on again how the market plays out. How does the approach evolve? A sort of price discovery, mature from that initial point of inception, the tokens lied everyone can trade it to 30, 40 days. You might have very different participants that are trading that token. I think there's a bifurcated outcome. In one case, what you just mentioned absolutely happened. There's an initial price discovery where a lot of people have supply, on locks, or people cell and then it bounces back. You have this dynamic that things settle in and it settles in in such a way that there's enough organic trading that it's self-sustaining. Basically, it becomes a healthy market thereafter either because there's enough interest, there's enough holders, there's just enough mind share that comes with the token. I think there's another path where none of that really develops and what you have then is just the market makers there and no real organic trading. In those cases, you have a few in its spectrum, but I think it eventually settles into one of the two where you have maybe very tightly controlled inventory. The price action is not necessarily reliable as indicator of long-term value. You have a lot of potentially insider manipulation and that becomes possible because of tight inventory. It just makes it very hard for market makers to be on the other side of that because there's no organic trading to really offset that at all. That's something that we have to really contend with in the crypto markets and sometimes if that happens a lot, you sort of go to the project and say this is what it is. We really can't do it in this way. It does depend a bit on the model in which they're paying you because some of those models do make it easier to provide liquidity than others do. How does listing venue change the cold start problem? So for example, or project hires you to be the lead market maker on Binance or they hire you to be the lead market maker on OKX or Coinbase. How does that venue decision ultimately change how you think about managing the problem? Oftentimes the projects can only get listed on certain exchanges if they have a certain degree of engagement and traction. So that already tells you about interest for the project, interest in the token and the people who might want to trade it. So oftentimes a listing on say a Binance or a Coinbase, I imagine the projects need to pass more rigorous sets of I guess filtering by the exchange. So once it does list, there's generally more tensions viewed as a positive signal. This happened more back in the day, but you still observe it. When the projects do actually announce a launch on Binance or Binance announces it, there's a price increase. I think Binance is viewed as almost a distribution channel for the tokens because it accesses this large customer base and it brings attention to the project in addition to what has existed before. Oftentimes, what ends up happening is the price discovery process and organic volumes. You can count on them being around more for larger tier one exchanges, whereas for the lower tier exchanges, it's tougher. And we also know that wash trading is kind of an endemic problem in the industry. So oftentimes with smaller exchanges, I think people really question whether the volumes are there. I don't know that it helps us calibrate necessarily a price other than directionally favorable, but you can sort of trust what you see more in the data more. And you generally have an expectation that things normalize to a steady state that's a bit more reliable on some of the bigger exchanges. I want to talk about the two large business models here and you referenced it in an earlier answer that there's a distinction between market making for profit where you've got this open competition. Edge comes from a variety of places including speed and understanding order book and balance and better pricing and forecasting flow and then market making as a service where you're getting paid by the project and I think to quote you here, the goal is to not lose your edge in being paid by the service, right? And I was hoping you could unpack this distinction and maybe how it changes how you actually market make operationally, how you think about implementing your market making structures. Market making for profit ultimately you kind of have a choice of whether to participate and how to participate. I wouldn't even think about it as market making for profit more like trading for profit. So you could trade either by being a market maker, by being opportunistic market maker, by being a market taker or by arbitrage in between different markets. There's a richness of different strategies that you can apply almost all that rely on their being organic trading involved. If you're a taker, you need liquidity there, but the only way the liquidity is there is if there's organic trading. Market making for as a service, I imagine what firms are trying to do is to bootstrap the marketplace. So this is not unlike Uber where initially they gave very good incentives for people to participate and to ride so people come for one thing, but then they stay for another. That's what these token projects are trying to do. They're trying to create enough organic interest so that people stay and continue holding onto the token, trading the token. All the problems that we talked about exist. We have KPIs so we have certain uptime KPIs where we have to be in these markets. We have size APIs where we have to quote some minimum amount of size at various spreads. All that is meant to give people active takers the ability to trade when they want to trade at significant size. And we have to manage this across a lot of different markets, potentially depending on how many exchanges these tokens launch on and the bigger tokens will launch on multiple exchanges. And then we have to manage inventory risk across these various exchanges. Those are some of the challenges that we face. And all the meanwhile, oftentimes, we're compensated with options. We actually have embedded Delta to these tokens because of the options that we're trying to manage in a responsible way. These are some of the challenges that we wouldn't have just trading for profit. Let's talk about the business model a little bit. As far as I understand that there's sort of two big ways in which this business works. There's a structure where the listing project actually takes P&L risk within some bounds. And then there's another where the market maker receives a loan plus an option to buy the token, which you just sort of reference. And you can I guess maybe think about that as a covered call to a certain way. How do the models differ in terms of how you think about quoting hedging and ultimately managing your long term relationship with the project? And are there certain conditions under which you'll accept one of those models and not another? In the early days, I think those are the two structures that normalized. So the loan option model is more popular and prevalent with bigger projects. And the what we call the retainer model is for smaller projects. But before that, I think during the wild days, there were more beast spoke deals that actually targeted certain price targets, which often sort of set quite weird incentive because market makers shouldn't be able to influence price targets per se that they were certainly incentivized based on hitting certain price targets. Those sorts of structures are perhaps what make certain tradfied folks quite uncomfortable with crypto. But to your point, the way we do it at least, and I think people do approach this differently, there's a few ways in which you can look at the option. One way in which you can look at the option is that it's basically a VC bet and you're paying for the option through labor rather than buying it outright from the token project. The way we look at it is that the options enable us to provide the quidity through delta hedging. Delta hedging, you're basically selling into the market as it goes up and then you're buying as it goes down so that fundamentally dampens the volatility of the project, which is good because I think the projects don't realize that what can go up quite quickly can come down quite quickly too. And it gives us a way in which we can deal with the large inventory fluctuations because otherwise you could really take on quite a bit of inventory without meaning to and that presents a lot of risk. In this model, because the delta hedging nature and the ability for you to manage the inventory using the option, the risk of market making sits with a market maker. In the retainer model, what instead happens is that the risk sits with a project. You would do your market making, you would tell the project what the PNDAL is and they would ask you, "Hey, why did you lose so much money? What happened here?" But fundamentally, the market maker take on no risk and is paid a service fee directly. So it might be something like, I don't know, $2,000-$5,000 a month depending on the number of exchanges. changes and the number of pairs. A way of approaching that that we sometimes do is, well, just tell the project, hey, what we'll do is synthetically hedge. Basically we'll act as if we had an option. Well, layer liquidity according to that, basically that gives you a lot more certainty around what the cost will be depending on the price path, right? But you can give a more reasonable bound in that structure because the project takes on the PNL risk, there's sort of a lot more degrees of freedom that people apply the problem with. It does become a client management question and a relationship management question where you have to give the project some sense of what might happen, why things are happening the way they are. Can you talk to me a little bit more about the inventory risk issues? One of the things that sort of naively comes to mind for me here is, again, going back to sort of the two different types of projects that might come to market. That's very hyped in which there's a large airdrop that you might expect initially a lot of sellers that you're going to have to absorb. Then the other projects that maybe are less hyped that aren't going to have as many initial sellers that maybe it's a more balanced initial market and you might need more inventory. At the end of the day, I presume you don't want to just be left holding a lot of inventory. So calibrating that initial amount of inventory to supply and demand seems like a really difficult problem, but an important one for the sustainability of the business. Can you talk a little bit about that? One of the things you can do, you are not all pretty unvaccinated. You sort of understand a bit of the tokenomics of the project. You understand how much supply might be coming online. You sort of understand the thinking of the people, right? Because basically a lot of the people who participate do it on a semi-professional basis in terms of farming the airdrops and whatnot. So you know what they're likely going to do. And you're actually right. Oftentimes, what happens is that there's initial selling pressure and then things normalize. So for us, there's sort of really limited things that we can do in this dynamic. And I think what it is is about thinking about how you deploy your inventory to be most effective. Do you step in or do you let it fall and then step in? What we've noticed is that it's not path independent. It's not like the same amount of trading has the same price impact. It's not the case that if you deployed the same amount of inventory, regardless of how quickly or at what price it results in the same buffering impact. You can deploy the inventory a smarter way. In the way that we delta hedge our option though, initially what we need to do is sell the delta, the embedded delta of the option. We would be sellers along with the other sellers and we make this quite transparent, but we just sort of space it out in time and we make sure to minimize the market impact to which we can. A lot of this is thinking about building up a buffer reserves so that later on when the price do drop, we can buy it back. And then similarly if the price is rise, we can continue selling into that price rise. I want to talk a little bit about maybe the aggregate portfolio construction angle here because right now we're talking about market making for a single project in isolation, whether it's trading for profit or as a service. You at Caledon simultaneously market make on over, I believe 200 plus assets, different crypto tokens that are going to have various correlation structures and tail risks and these wildly fat tail distributions. How do you think about aggregate inventory exposure? How do you decompose what's idiosyncratic to a given token versus systematic and again with all of the way these correlation regimes shift and liquidity cascades in the way that Dexas and centralized exchanges are organized. You got a unique problem on your hand. So how does your risk management adapt all of this? That's a pretty complex problem. It's one that we think about a lot and continue to work on. One thing that you mentioned is that correlation shifts in time and I think we see that a lot. Oftentimes correlation is liquidity phenomenon. So if you look at BTC, it's become more correlated with general traditional asset markets. And often that's because the same holders of risk make the same decisions of risk on risk off. And so BTC becomes more correlated with equity markets as BTC becomes more intertwined with the traditional financial markets. What we've seen though is really that for small tokens or new tokens, almost all the risk is idiosyncratic because it's sort of dominated by all of these dynamics that we've just been discussing. Day one or maybe day 10 dynamics that's still playing out. There's not so much impact of what the overall market is doing, but also we hold usually smaller inventories of these. We just kind of took a more pragmatic approach, which is that we essentially have two risks that we look at. And then this will evolve in time. We don't think about factor risks. Crypto factors are still quite new and we can't really decompose them so well. But we think about an overall market risk that we control and then we think about an idiosyncratic risk. What we do when we sort of manage our overall portfolio is we skew our pricing differently for different coins depending on how correlated to the market. It is. And this correlation is not a numerically estimated measure. It's more of a kind of a sign measure. So I don't know. Some of your listeners might be familiar with bar factors and bar factors are derived in such a way where the loadings on certain factors are actually assigned and they back out the factor themselves. So it's a little bit like that. For bigger tokens like BTC and ETH, there is going to be a lot more, I would say, market based skew or overall skewing of the price from portfolio level exposures, whereas for the smaller tokens, it's almost all of it from their own inventory. And we just have tighter bounds on those. So it's a very pragmatic approach and kind of works because our trading goes in and out so quickly. If we're holding positions for much longer, I think you would need to have more sophisticated management of these risk factors, but for us, we're able to turn things around quickly. So generally, the risks pretty tight. Historically, there has been a perception and I think it's probably fairly accurate that spoofing is endemic in crypto markets. I don't know how true that is today. But for listeners who are coming from traditional markets, this is explicitly illegal. In crypto, it's probably reasonably illegal still, but people have been able to get away with it because it's largely unregulated and difficult to pin down. How does the prevalence of spoofing change the way that you read the order book and think about adverse selection when you're making markets? So I actually started doing a bit of high frequency trading in traditional financial markets before spoofing became explicitly illegal. This was the global financial crisis with Dodd-Frank that became explicitly illegal. And people actually prosecuted it, I think the SEC and the FBI prosecuted it under wire fraud statues. So I think those same statues might apply to crypto as well, but as you said, people do it anyways. What's interesting was the spoofing was really targeted toward high frequency trading strategies because they tend to react very quickly. It's quite sophisticated as well. I actually, personal note, worked in close proximity to someone who was one of the biggest spoofers. It traded 10, 20% of the ES market, one of the most liquid markets in the world, and was able to successfully manipulate that for profit. And what's really interesting is there's sort of degrees of sophistication where you do this on a primary market where you don't lose much money or you break even, but you derive the profits from another market that's influenced by it. So what example might be that you trade in ES, the eminions, but you make your money on SPYs and all the other stocks that move with it because there's effectively a large set of derivative markets. Let's call them derivative or influence markets that takes a signal from a primary market. And you see a lot of this in crypto as well, right, where a lot of these economic hacks are because someone references a smaller market and then so much liquidity builds there and people can manipulate the small market and then make profits on the big market. You see that in crypto, you see both the very basic manipulation, which is somewhat easy to filter out, but also the more advanced manipulation, but you also have all of these other problems like insider trading and very tightly controlled supply and all that. All which is to say that we're sort of operating a very adversarial environment. And this is where the table selection becomes very important. You want to choose the tokens and the venues that have good edge and there's a lot of different filters and market analysis that we can do, which identify these markets. So we pick our outside of market making for as a service, which we're being paid to do for our actual prop trading. We sort of pick our markets very carefully. I want to talk a little bit about venues. I think what we saw a couple of years back with the rise and subsequent collapse of FTX was a pretty seismic shift in the way people view venues within crypto. All of a sudden this idea of counterparty risk went from maybe theoretical to existential overnight. Curious, how is market making infrastructure, view of venues, credit relationships and the ideas of maintaining capital of. efficiency across venues, been rebuilt and rethought since then. And I'm particularly curious because one of the things that evolved out of the FTX demise was the rise of perp dexas and hyperliquid, especially, which emulates so much of what a centralized exchange is, but is now doing everything fairly transparently on chain. And there's some meaningful trade-offs there in the way that those exchanges are managed with their centralized vaults and ADL being such an important component. So maybe you can talk a little bit about how the landscape has shifted and maybe impacts its hat on your business. So FTX did impact us, but not nearly to the degree that I think it impacted some of the other firms in the space at the time. We had maybe about 5% of our equity on FTX, but I do remember frantically pulling money out of other exchanges just to make sure that there weren't any sort of systemic issues lurking. And I think exactly right, risk became top of mind for a lot of people. The way I conceptualize it is this is exactly like adaptive market hypothesis, collectively attention shifted from one thing to another. It became much more fashionable and defensible to talk about risk, whereas before maybe you were seen as being the mace air. What then happened was a push towards, I think, certain infrastructure being built that can manage that risk. One of the big firms that came out of that was Hidden Road, so they underwrote insurance and counter-party risk. So that became an interesting product for a lot of people. You have Copper and Bicco now offering essentially credit intermediation services. You can post your collateral or trading funds with them and they'll reflect it onto certain exchanges with whom they have partnerships. And then the third thing I think we saw is probably more bilateral trading relationships develop. The way I think about crypto is quite similar to FX or commodities in market structure. You have a lot of bilateral trading relationships that's established where people can vet each other on bilateral basis. There's a lot more transparency than with centralized exchange like finance. But also you did have more transparency with even those, for example, their reserves being published publicly. Those are some developments. And as you also mentioned with Perp Dex's essentially this idea of using smart contracts to hold customer funds and customer assets as opposed to it being a centralized entity into which you have more limited visibility. The FTX Fiasco created a lot of these opportunities for new structures and new companies to be formed. One of the interesting things with, and this is probably true with the centralized exchanges, where you certainly see it with the decentralized exchanges and the decentralized Perp Dex's that have come out over the last couple of years is that because you have, maybe I'll call it venue sprawl where there's so many venues that you can trade across, you don't have what you have in traditional finance where you have one centralized arguably protected venue where there's a waterfall of ways in which customer assets are protected. If you're trading something like E-minis, the likelihood of your position being cut against you during a significant market sell-off is very limited. The client comes first with all these venues needing to protect their own solvency. They have a waterfall of ways in which they can try to manage this leverage cascade risk. But when you mix lack of centralized clearing plus a huge amount of leverage in crypto, it seems like at least once a year there's a massive liquidation wipeout and you have these auto-deleveraging effects that then have knock on wipeout effects for many traders who are maybe trading long short and trying to hedge their positions and losing one leg of their trade. Terri, so you think about that risk? It is very different on a venue by venue basis. They all operate differently. It is very different than traditional finance. I think with any Delta neutral trade, something like a funding radar or anything where you're carrying a long and short position across different exchanges, this becomes quite risky. There's no easy answers here. I think it's one of the reasons why you're paid a premium, right? Is that you're carrying this risk. There are ways in which you can kind of manage on individual exchange bases. It depends on your strategy. So one thing is that these ADL risks are proportional to your return. So the more money you make, the more likely you're to get liquidated ADL. I think they do this to give you some sense that you're okay, at least you made money. What you can do is when you've detected that you've made a lot of money on a particular position, you can just trade out of it back in. That's obviously expensive, but it is a way to manage the risk. Other things are more, I think, at the strategy level where you just have to look at, while how big is your notional position, this particular thing relative to open interest? When you spread that open interest around a bit more, all that is more trying to limit the impact should it happen. There's unfortunately no straightforward answers here. With the rise of perp decks, I feel like traditional defy has been maybe a little bit forgotten in the narrative, but it is something that Caledan actively participates in. Not only do you participate in market making and centralize exchanges, but traditional defy pools as well, where you're providing liquidity against these automated market making curves, and you have to manage things like impermanent loss and you're deploying capital into smart contracts with different trust assumptions and audit needs. It's really a fundamentally different exercise than looking at a centralized limit order book. As I mentioned, you operate across both simultaneously. When you talk about the structural differences between the centralized approach versus the true defy liquidity pools, not just an execution, but how you think about risk and capital allocation and the nature of maybe edge itself. The way I would think about decentralized venues is one, there is a counterparty risk, but the counterparty risk is really against the smart contract. Then there's a question about how secure is a smart contract, is there an exploit? That's a very tough question. You have audifers that specialize in this and they still miss it. Our thinking is it's hard for us to effectively evaluate that better than other people can, people who are specialized, so we don't even try. We just rely on existing audits if it's new, but really you rely on the battle hardening of some of these protocols where they have been hardened over many years, they have been exploits, they've been fixed and so forth. The other feature of D5NU is that it's very expensive to do transactions. That's actually why the Uniswap, the constant product automated market maker was such an innovation, it was because you could do it actually cost effectively. It was cost effective from a gas perspective, but it was very inefficient from a capital perspective. You started with Uniswap V2 where you had to put in a lot more capital to allow a certain amount of trading with a certain amount of market impact, but it was very capital-infishing. Then Uniswap V3 came out which introduced this range-based liquidity. If you think about that, it's a limit, that's just like a limit order, but then you need a more active management. You see this transition path back in effect into an order book of some sort. For us, though, all these inefficiencies essentially create some opportunity. One that's very well exploited is this C5D5RB. Essentially, oftentimes centralized exchanges or limit order book exchanges are where price formation happens and then the D5 side lands. That creates an arbitrage opportunity in which you can push those prices back into line. That's when you are on the taking side. You can also selectively be liquidity providers into these polls and then also make money, but you're almost always hedging your position into the C5 exchanges as well. It's very much a trade-off between how much fees you earn versus how much adverse selection you experience from the arbiters. For example, if you have a poll where the only people who trade are the ARBS, then there's no way you're making money because they're only trading when they can overcome the fees. But if there's enough noise traders where they're going back and forth and they're paying the fees, then that can wash out where you're actually still making money. Those are some of the challenges. Risk management is entirely separate because the smart contract holds your custody. That's just are you in or out and then if you're in how much. The last thing I would say is from an infrastructure investment perspective, DeFi is quite different. Because again, the high costs are such that everything's batched together and because now there's real economic consequence, the structure of DeFi markets are such that depends on the protocol, but for Ethereum, there's an auction-like process whereby you essentially bribe the block producers to put your transactions first. You're often paying a lot of your profits there. For us, Phelana, what's interesting is it's so cheap that people just spam it. The way in which you gain an edge is to just kind of spam the chain with a bunch of transactions and you hope one of them gets through. It creates a lot of very interesting incentives for these profit-motivated players, but it is quite different from the traditional game of high-frequency, same ideas, but different implementation. In your experience with the breadth and depth of experience you have working with a number a project. to bring new tokens to market, what would your advice be to those who are doing their initial listing and looking to maximize the long-term health of the token? If you take a step back and think about what the token is, it depends a bit on what the project founders are going for. Projects I think the token is a product. Everything else is kind of a marketing support for that token. In that case, maybe this discussion is irrelevant. But another way to view it is a mechanism by which you're rewarding participants in the ecosystem with economic value that's being generated. The key is that there needs to be economic value. The token also has other benefits like coordinating behavior that otherwise might be harder to do, etc. The underlying fundamental is that there's real economic value and that the token reflects that. Oftentimes, I think people focus a bit too much on the token itself versus building something that has sustainable value. The main thing I would say is make sure your project actually has real value and real revenue and real use. That ultimately will drive the token. I think nowadays people are also a bit more sensitive to the actual mechanism and the rights given to the token before it was like, trust me, that's also starting to mature a bit. Nevertheless, I don't think there's a substitute in the long term for the project itself actually being valuable and the token participating in that value. We push on that a little because I think in crypto, especially the token is the product and the product is the token. There should be ideally a business behind it. We see that in some cases where they're using revenue to do programmatic buybacks. There is a fundamental cash flow behind the token. You see this reflexive relationship that when a token isn't doing well, it seems to almost have a negative marketing impact on the project which drives revenue lower. If the token price doesn't do well and you don't necessarily see that in traditional markets, it's like, well, Nike stock going down doesn't really impact me buying Nike shoes. For whatever reason in crypto, the price of the token to going down seems to impact people's use of certain protocols. I would argue potentially the price health of a token actually is for whatever reason in the state of crypto right now. Important to the sustainability of the project. Curious your thoughts on that. I think so. I think the token does serve a marketing function. There is a reflexivity element. If you look at, for example, hyper-liquid, there's real cash flow. Their actual launch started. It was interesting. It did go quite well, but they started quite low, for example, like $3 or $2. Now it's at 40. It's not about going out with a bang. I hear everything that you're saying. I think reflexivity is not just a thing about crypto markets. If we look, for example, at the Wall Street bet stocks like AMC or GameStop, they were able to raise capital off the back of those pumps. And that really gave them kind of a lifeline to continue doing other wacky stuff like buy mines and such. Or if you think about how stock price may or may not influence employees who have stock option grants, it does have maybe not as dramatic, but it does have real economic impact on the company and people's motivations and things like that. So that reflexivity is there. And one of the critiques of I think crypto is oftentimes the token seems to prefer for us. It's just to have one. And for the health of the project, it actually detracts people's attention. The team's attention to manage now this effectively thing that trades, which is a very different problem from working on the project and what the project is meant to do. That speculative option embedded into the token is also what allows a lot of these projects to kind of raise capital, for example. When we evaluate equity based investment versus token based investment, the token based investment have much more near term liquidity. And there's sort of a sense that it's used to be that even if the project fails, there's a kind of a speculative value that remains in the token once it has captured critical mass. You look at certain projects like ETC, who knows where it is. Yes, it's still trading. It's still very non-trivial sort of market caps, but there's not really a team where anyone working on them anymore. It's just become its own little game that persists, I guess. You've recently extended your services to include OTC structured products here, building out the spoke derivatives, forwards option structured notes for token foundations, funds, and family offices. Curious how you came into this line of service and how you're thinking about sort of the transition or differences between your primary line of business, which is right, market making as a service to this new line of business. The first one would be, they're quite different. I see a few areas overlap. One area overlap is perhaps the hedging capability. Having access to linear capability where you have access to a very wide breadth of places to hedge is a comparative advantage because you can write the products, you can design the products. A lot of this follows sort of trad-fi design. It's nothing new, but to actually be able to manage the risk afterwards is quite tough because of the liquidity within crypto is just much less than a lot of traditional financial markets. One aspect of it is leveraging the existing market making infrastructure, prop trading infrastructure to do the hedging. The other one is, as we pointed out, there's no central clearing in crypto, so a lot of it's based on an evaluation counter-party risk as well as trust. We've kind of survived through all the ups and downs and blow-ups, and we've done it with just our own capital, our own balance sheet, never with investor capital. There is some credibility on our part that we can leverage to be counter-parties to a lot of these clients who are looking to benefit from structure products. The third one is we're in touch with a lot of these foundations who can be customers for them, and because we already have that working professional relationship, it's a way into essentially offering them more services from which they can benefit. Just pause on the client side here for a bit. I suspect, again, for a lot of my traditional finance listeners, this is all going to be a little bit novel, and I want to create the context. If you're a token foundation, you might be sitting on hundreds of millions of dollars worth of your own token, but it's going to be illiquid, maybe locked up, obviously highly concentrated. Your ability to monetize it and use that capital is limited because people are watching what you're doing with it, and/or you might just blow through whatever liquidity pool or order book there is, trying to convert it into some other asset you can spend. What does Treasury management actually look like for these foundations? Maybe you can talk about the core problems they're trying to solve and use this capital for and how, given those problems, the types of structures they're coming to you to request that you're building out for them. The first thing I'll say is that Treasury's are quite diverse group. There are some Treasury's that have been around for a long time, and most of their assets are actually maybe in BTC or ETH, and they just haven't done anything with it. There are some other Treasury's that have hired very, very professional people from traditional finance that are quite knowledgeable about option structures and the limits are in. It's a very diverse group with different needs, and the way that I think about structure products is that you have these Lego pieces. The Lego pieces themselves are quite technical, but you don't really talk to the Treasury's about that. Talk to them about their needs and whether constraints are essentially used that to synthesize certain payoff structures or ideas to satisfy their needs. What are some of these needs? I think EO generations one. We hold a lot of assets. They tend to be in crypto, not just their native tokens, but other crypto, and these typically don't have a way to earn assets. That's one. Another need is, hey, we have a runway. We have these operations, little costs. How do we have a program where we can finance these operational costs or sell our tokens to cover these operational costs without spoofing the market? The needs are quite diverse, and that's where customer education comes in. That's why we have a team that goes and speaks with these Treasuries and understand their needs and then come up with something custom. I will say that the concern, the signal and concern that you mention is very real. I think it would be like if Elon Musk was like, oh, I'm selling on my Tesla, everyone would freak out. You have the same concerns and tradfights as much more concentrated in crypto. I think there is room like in tradfighter solutions where you pre-commit. You're not trying to time the market. You pre-commit to a certain program. People understand that. In crypto, we think there's probably a way to do that as well, where you can satisfy your needs without necessarily spoofing the market into in terms of thinking that you have insider information where you lost faith or stuff like that. Some of the structures that we create does help to shield the Treasury from having that negative signal to the market. What we're actually doing is monetizing volatility, monetizing convexity, and trading counter flow to where the market naturally trades. That also helps a bit. One of the things I noticed about many of the retail crypto participants is that there's generally this negative view about market makers for whatever reason that the market makers are just destroying price or playing some nefarious role or that there's backdoor deals. What do you think the biggest misperceptions are? How would you clear the record as to the role what market makers are really doing in these markets? That's hard to do because I think there's maybe some truth to some of those perceptions because market makers as a group is a diverse group with different actors that have different choices that they make. As I mentioned, there are even now these active market makers that very much try to paint a picture. Have you read the book "Reminism of a Stock Market Operator"? Absolutely. It paints a story about how you can manipulate markets. It's like pump and dumps and penny stocks. That happens still in crypto. Crypto projects do engage with these active market makers, the attraction of a high token price is very appealing. All that does happen. Some fortunate that it does happen, but it's there. When we think about market makers, perhaps the key is not to use such an overloaded term to describe what it is that people do. That's why we've always chosen to stay away from, I think, these more manipulative or deceptive practices. It's just not really in our DNA given what we came from. I think I've seen some pictures floating around on Twitter where participants have labeled different key market makers and what the historical price action has looked like for the, you know, and you say, "Oh, this new token is being run by this market maker. That means the price action is going to look like this or XYZ." It's a little tongue-in-cheek to your point. Maybe there is some historical precedent for it does look like that. There's a difference between cause and effect there, too, right? It's sort of like these tutoring services that claim 99% of their people they've tutored made it into Ivy Leagues. Very likely they just chose the people who would have made it anyways. I think a lot of the initial price action isn't really within the control of the market makers. They can create certain things in the short term, but it's hard to sort of maintain that without a set of conditions being true, one of which is a very tightly controlled spot sort of inventory. What's interesting is we actually see that when we trade because one of the things we do is help people hedge and one of the very different things about crypto is this idea of a funding rate and perpetuals that don't really exist in TrapFi. The funding rates can just behave really, really strangely and super hard difficult to hedge. It's a place where we've spent quite a bit of time doing research, but it sort of has that flavor of a tail wagging the dog. So imagine you have a very small spot inventory, maybe even just a few million dollars, but you have an open interest that's like an order of magnitude higher than that. If you manipulate spot inventory, you can affect a lot of economic outcomes for the holders of the perpetuals. It creates a really interesting kind of market dynamic there that can be manipulative someone chose to do so. These are the things that make it a bit interesting, but also difficult, I think, for TrapFi buyers to navigate the crypto markets. From the outside, it does appear that crypto in many ways is speedrunning a lot of the lessons and product development that TrapFi has had over the decades and centuries. I'm certainly not the first to have stated that. Here's where you see the market today and where you see it heading, particularly as TrapFi institutional participation deepens in these markets. I think a lot of this is credit to the roadmap that TrapFi already laid out. If you look at the structures and the things that have been built, it's pretty faithfully replicating a lot of what already exists in TrapFi. I think it has a couple of advantages. One is that it's like a nice regulatory arb. You sort of have a hyperliquid where they can sort of say, "Well, we're not custody and customer assets." Then the other thing, of course, is you have programmatic money and these relationships that are codified so you can see them play out without any circuit breakers, a much more raw way so you can get a lot more learnings from them. The way I think about crypto, I have a couple of theories. One is that it'll just replicate Forex and commodities markets. You have a lot of the same challenges. You have essentially these little regional markets that are there for regulatory and other structural reasons that are isolated. You have essentially a lack of central clearing. You have the need to actually settle something. It's not like oil where you have to deliver foam place to another, but it's not quite just risk that you're settling. It's actual physical things that you're settling. I see it really developing in that direction where it's a highly, highly distributed, highly fragmented market where you have a lot of different bilateral relationships. There's a very central role for a market maker not in the market making as a service sense, but in the trading principle, risk sense to balance flows across these different markets. Then you also have DeFi, which is yet another form of fragmentation and takes its own form of management. The interesting thing about DeFi that I'm not sure about is people have already started talking about permissioning in DeFi. In which case, then it's more maybe just a ledger in settlement. It's a technological innovation, but fundamentally will be integrated into the traditional financial system. There's also the non-promission side, which maybe then it becomes this thing that will always exist, but it's a little bit more nichey and will have ultimately less capital. It'll still be there in its little sandbox, but not to the degree that it's there today. That's how I think about the system evolving, but we'll see. That's the closest analog I can get. John, I always like to end these conversations with a bit of a personal question. The question I've been asking all my guests lately is, what is something in this has to be outside of work that you were currently obsessing over could be an idea, a book, an activity, a hobby, music, something that just currently has you completely captivated? My kids are advocating for a puppy. We've been talking about a puppy. How you have a puppy in Singapore where the rooms are quite small. Also, I've been obsessing a bit over AI because a lot of people use it for search and it's great for that. I think you can fundamentally create different workflows for yourself that is ultimately a new way of approaching things. I'll give an example. I was talking to a friend and he said what he did was to actually feed in a lot of personal information and preferences into AI, to almost build a virtual version of himself, to talk to himself about how he would approach certain decisions. I can also think about feeding in. I'm a collector of mental models. I really like this idea of feeding in a few guiding mental models and saying, hey, what's going on, I think it will really change the nature of work and what we define as work very, very quickly. Thinking about, are we even teaching our kids right things? Is it going to be valuable? Like a lot of those questions now sort of sit with me. Have you ever watched an interview with a vampire in the movie? I have not. So, in there, there's this scene where basically there's this leader of a coven of vampires that have been around for hundreds or thousands of years. What he says is that one thing that people really struggle with is keeping up with change and the changing of times. I feel like change is so fast these days. It's very hard to anticipate what things might be like. It's quite disruptive because you plan for one thing, you spend years planning for one thing and it's changed. I don't envy my kids basically. They're going into a world where they have to struggle with a lot. You know, it's something I think about a lot too. I actually feel like AI has been something that a lot of guests have answered with lately, but everyone with their own unique particular obsession about AI. It's certainly those with kids. For all of us with kids, it's a question of kids that are currently in high school or middle school are going to have a very different relationship with this problem than kids who are toddlers right now who might just grow up immersed in it and the world will have already changed before they're aware of it. So, life is moving quickly. It is. Yeah. Well, John, thank you so much for joining me. I really enjoyed the conversation and thank you again for taking the time. Thank you for your time as well. [Music]