Hummingbot Hummingbot@Liquidity 2020, 23 Oct 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Hummingbot@Liquidity2020 - Quant Panel ft. Michael Feng of Hummingbot & Kevin Zhou of Galois Capital

Description

Learn more about the Liquidity 2020 summit: https://liquidity2020.dystopialabs.com/

**Important Disclaimer**

This demo is for instructional and educational purposes only. Any parameters used are purely for demo purposes only. We are not giving any legal, tax, financial or investment advice.

Every user is responsible for their use and configuration of Hummingbot as well as their participation in liquidity mining.

Please see https://hummingbot.io/liquidity-mining-policy/ for the full disclaimer.

A

Hey everyone uh we are here for our next uh panel. We have kevin and mike uh they'll, be doing a talk around uh quantz um I'll. Let you guys take it from your. Do you mind introducing yourself while the audience you guys can confirm if you can hear them.

B

Of course, uh hey everyone, uh I'm mike, I'm the co-founder and ceo of hummingbots we're an open source project that helps anyone in the world, use it to basically democratize.

A

And become a high frequency.

B

C

uh Guys, I'm here with, uh do you guys.

B

C

A

So I hear it on your end. uh There's two reasons that might happen: uh it's because you have crowdcast open on another window.

C

A

uh Another way is for you just to refresh your browser.

C

That's probably what it was: uh okay, great, so um hi guys. uh My name is kevin, um I'm a managing partner at galwa capital and uh we are a trading firm which focuses primarily on uh liquidity provision, um and uh you know our uh approach to. It is definitely very quantitative.

B

Okay, um well thanks everyone so yeah. I I think that you know it's been a long day. I think this than day three of the conference, um so I'm sure everyone's pretty tired. Now and uh but I think this session, uh it's really just more of a conversation uh between me and kevin. uh So as background I've known kevin for three years now, uh we actually also used to run a quan fund um when we first started, but but uh you know, we only ran it for a little while.

B

So so we're not really the expert we quickly pivoted to hummingbot, uh because we realized we're more engineers at heart, uh rather than traders and- and I think that you know, building software and making that available to traders. um I think, at least for me, was just a lot more um less stressful than having to worry about managing people's money. You know every day, and so I I think that um I I give a lot of credit to kevin because he's been running galway capital for the last.

B

uh What three years now, four years, maybe uh three years now yeah.

C

B

And I think it's obviously he's you know he's much more it's better than I am at managing money for other people, uh but I think he's also much more experienced in terms of how a professional trading firm views the crypto market. So I'm really glad to have him here today and hopefully we can have a very like-hearted conversation about.

B

uh You know uh what it's like to run a fun fund, um how it's different from you know, maybe like what people today are doing in terms of trading and also what like, how a quant fund views all the developments that happened this summer in d5 and with the um what it means for the future of of trading.

C

Yeah, so I I think uh uh you know, I appreciate the kind words there, but I actually think mike that you actually have a harder job than me, uh because when you're building product um you have customers who don't have customers- and you know just dealing with customers- is definitely not easy. So uh you know I definitely prefer what we do. It is more. I think it is probably more stressful, but at the same time I think there's less headaches overall, uh because there's just fewer constituents to deal with.

C

So you know, I think it's just the trade-off. um You know to each whatever they prefer.

B

Yeah, that's a fair point. um So, first off, can you tell us about how um like how what is galvas capital's origin story, and how did uh how did you end up where you are today.

C

Sure so uh we got started in january of 2018, and uh you know I I met my co-founder. We both went to school, uh basically to be quants. We were in the masters of financial engineering program at berkeley, uh which is where I met him. uh You know we we've both done some stuff. Since then um you know he was working at a machine learning startup. uh He was head of data science over there. uh For me, um I was working at two different bitcoin exchanges. uh You know after I graduated from the master's program.

C

One was a very early one called buttercream uh lasted for about two years and it went under during the bear market of uh 2013 and then afterwards I ran the trading desk out kraken for uh another two years. So after four years at these exchanges, I thought it was about time to you, know: go more to the buy side, uh take a little more action in the market and that's why we set up um yahwah capital, and uh you know it's been a long journey for us. uh You know for almost three years.

C

I think, which is you know, crypto time? Quite uh you know, you see a lot of things happen and definitely the landscape has changed a lot, um but uh you know I don't think, there's anything else. I'd rather do so very happy uh to be doing the work.

B

Well, I'm sure a lot of people in the audience are crypto traders, and I think um you know I'm sure some of them have had aspirations of starting their own hedge fund uh in the future. uh So so I think that you know obviously the most I think most hedge fund managers start by just trading for themselves. You know, I'm sure you did that, and I did that too. um What are some of the differences between managing capital for other people versus trading for yourself.

C

Yeah, I think, there's a quite a few differences there and I think um you know I definitely agree with you. I think a great way to get started is just to trade. Your own account uh your pa and once you've built up a very nice track record. Maybe at that point um other people might be willing to give you some some capital. um I think uh you know there's really um different different parts to it.

C

There's sort of the trading side, which is, I think, uh very similar to how you trade for yourself, but then there's sort of the business side, which I think is quite a bit different, um and you know that involves you know. uh You know hiring people, um you know creating the legal entities figuring out the right structure. Consider considerations about um you know: tax uh law. um You know dealing with administrators who deal with the accounting.

C

um You know all sorts of things like that, as well as structuring the fees and what you're going to charge. um You know investors, um you know, and then, on top of that, finding the capital, convincing people to give you capital, um and then uh you know dealing with the mechanics of that capital flow. I mean: are you taking subscriptions redemptions every month, every quarter every year you know. How does that work? You know stuff like that.

C

um So I think there's a lot that goes into it outside the trading component, um though I think um you know I think, to get started. I think that that component at least has to have um at least I think that you know good enough of a track record um that it's worthwhile to try and expand and build out a firm around it. um I would say that you know another uh particular difference when it comes to the trading side.

C

uh You know, which is maybe a minor difference, which is um how do you think about risk? You know if it's your own money, you know at the end of the day, if you lose it, you only have yourself to blame. Maybe you feel pretty bad, but you don't really have any particular obligations or fiduciary responsibilities to other people. um While you know if you're managing other people's money, you know there could be certain constraints on the kinds of risks that you can take the kinds of instruments um that you can trade.

C

So you know with that. In mind, I think generally you're you're going to be a little bit more conservative right um and I think you know- and I think it really depends um on manager. um Certainly there's always this.

C

um You know very pernicious, uh you know very bad kind of incentive uh and you know, I think, colloquial it's called uh opm or other people's money where you know, if you think about how the fees are structured and whatnot, um some fund managers might think about that as a free option on someone else's money and maybe wants to take exorbitant risk because they keep some of the upside. But the investor takes all the downside. um You know, obviously that's very bad for long-term business.

C

It's unethical, um you know it's not a good way to go, but I think certainly some people do think about it. That way, so I think you know in convincing investors uh to give you capital um there's not only sort of a a a show and a proof of competency that you're going to manage. You know that you can manage their capital wealth, but that you will that you know there's enough transparency, and uh you know you have a good enough character as a person uh that they trust that you're not going to do anything.

C

uh You know along these. These you know very pernicious incentive um uh lines, so I think you know- and I think that's sort of what I also look for uh when I invest into other managers, because you know I obviously I also invest uh you know I'm both managing capital, but also investing uh into external managers. uh You know I always look for two things.

C

um First, um are they competent and second, are they trustworthy and I think both are extremely important uh if any one of those um is not there, uh you know you can have a you know, kind of trustworthy guy, that's completely incompetent he's randomly gonna lose your money away. You know innocently enough, but you know that could happen or you could have some guys very competent, but also uh you know very distrustworthy and then they're going to find ways of making money off of you rather than making money for you right.

C

So I think both of those are you know very key attributes, and I think um you know that that goes. I think that plays into um how to raise capital- and it's about pitching investors, that both of those three two things um are the kids yeah.

B

Yeah yeah, I really like what you said about managing capital, other people and like how the pressure they put on you, because I don't remember this like viscerally this one day when we're so managing a hedge fund, it was december 22 2017 the market dropped 22 in that day um and after 10 you know our fund was actually not in the market. um You know it. The bot just bought after the market would that went down 10. and then the market then dropped a subsequent another 10 12, and so uh I'm thinking I'm like wow.

B

Like you know, it's like. I knew that the bot would just keep on holding, regardless of how you know how far the market drops. I think we actually ended up pulling the plug and selling, um and basically just you know, just you know it's basically just selling manually in order to protect our investors capital, but I think that it was our own capital. We probably would have made a different decision on that day um and we probably actually would have ended up by making more money, uh because market actually did recover the next day.

B

But uh I think that it's kind of like I feel like when you manage other people's money. It's like you almost have to like be more like use the risk parameters you're managing too for them, as opposed to perhaps like taking more risk or less risk for yourself.

C

Yeah, I definitely agree, I think um I think. Overall, uh you know, I definitely take less risk when I'm managing other people's money, and it's like it's a lot due to what you're saying is that their preferences are just different than mine. uh You know generally us folks who are in crypto who trade crypto we're used to this kind of volatility. You know for them.

C

This is the most volatile asset class ever that they've ever been exposed to at least for some of them who are not native to crypto, and I would say another thing too, which is you know when you lose your own money? um You know you just feel like an idiot when you lose when you lose other people's money. You both feel like an idiot, and you know now you have to have a lot of very uncomfortable calls and conversations you know. So it's just like doubly bad.

C

You still feel like an idiot, but now you're gonna have to have a lot of awkward conversations. So you know that's just it's just real real, tough, sometimes.

B

Yeah yeah, I agree. I agree, um okay, so actually by the way I want to say, if anyone has questions, please drop it into the chat. um Okay, can I have known this for a long time, so we can just keep talking like this for like hours. So um if you guys have questions, please put in the chat and we'll try to address them as soon as you have them, um so so switching gears a bit. Let's talk about some of the more like relevant developments that happen this summer.

B

um Obviously I think the one of the biggest things has been the rise of these automatic market makers like curve, unisoft and balancer and how they have accumulated a lot of capital uh and kind of, like you know, and they're, almost like a like a competitor potentially to some of the larger exchanges like finance and so forth. uh That have appeared, um and so I guess my first question is: um are you guys actually actually participating in those amms either as a liquidity provider or as, like a you know, a potential arbitrager.

C

Yeah, I think we've played around a little bit with both, but it's not uh our core focus. um I think uh you know it definitely has gotten bigger uh and quicker uh to that. To that um to that size uh much quicker than I expected, um I think if it continues to go that way, then we'll have to look into it a little bit more seriously, but yeah we've dabbled here and there um and it's uh it's interesting, the innovations that people come out with in the space. You know.

C

I always think that yeah and especially, I think the relevant right now. There's this huge debate, that's going on in crypto twitter uh on one side of people who are saying amms, don't really work in the long run and there's other folks that say. Oh you know this is the best thing since slight spread, um you know, maybe maybe the truth is somewhere in the middle. It's really hard to say.

C

I have my own opinions and I always find it funny, because whenever something like this happens, it's always like there's always one side that sees it as okay, we've completely reinvented. um You know this wheel and it's just a way better wheel than has ever been done before, and then other people are looking at and saying. Well, this isn't a wheel at all. This is just a square wheel. You know strictly work what we had before. Why are you reinventing things in such a clunky and uh bad way?

C

So you know, I think, there's a there's a lot to be debated. I think on both sides. uh Each side makes some good points.

C

I personally am on the side that I don't think that the a m replaces the function of a club, and I think it's particularly difficult for market makers to be liquidity providers on the emm's without in the long run, getting picked off and and what I mean by that is, if you think about the structure of it, um everybody pulls liquidity to the exact same price levels and the behavior is completely deterministic right.

C

So it's as if you were facing in the market, um somebody who you knew on the other side, you knew exactly what their behavior would be for every action that you took. um I don't think that gives enough flexibility for the market maker um in order to make profits in the long run. I think eventually it you know you're, basically absorbing you're, going to be absorbing a lot of toxic flow um as that flow gets smarter. Now, obviously, there's some things that compensate for right. Now there are fees associated with these a m's.

C

um You know the flow itself is fairly soft. It's mostly retail guys. It's mostly um ethereum folks, it's mostly people who are not. uh You know what we we would consider like. You know sharks compared to like, let's say the prop shops in chicago uh stuff like that, uh but I do think that- and I do expect that over time the flow is gonna get more toxic. Liquidity is going to start evaporating because lp start losing money.

C

uh I think impermanent losses turn into permanent losses, uh and then uh you know generally a little bit pessimistic about it, but you know it could go either way. I mean, I understand the argument on the other side too yeah. Well, I don't want to just call it. You know we're happy to see people just try out new experiments, yeah.

B

So yeah I mean for us. Actually I feel like as a company that makes software for market makers.

A

I would say a year ago I was definitely in the.

B

Square, the square uh you know the square peg camp where I was like this is like how how could it? How could this work?

B

How could this even work yeah and I would say, I've been proven wrong over the last like you know, year or so, because you know especially I'll say last few months as as you know, and I think to me, I think the the the reason is because um I would say two things, two reasons I think the first yeah I think without in the in the um in the abstract, if you, if you just have like just you, shouldn't, earn fees, then you're probably going to lose money if you're a participant in a normal amount.

B

um I think that it is easier to to be a liquid provider for these, like stable coin mms, like curve, for example, where there is kind of like.

C

You're not really.

B

Taking as much in permanent loss risk um because I think that's easier uh and I think the other one, the other way is, I think, because of liquidity. Mining quote unquote where token projects are air, dropping tokens on on to people who are liquidity providers. I think it's added like another source of revenue and maybe in certain cases like that kind of, does compensate for the impermanent law source of certain changes.

C

Yeah yeah. Definitely um I think um I think overall, I want to talk about both of those points that you brought up. I generally agree. I think um the first is more about, like the correlation of the two assets that you're providing liquidity to right right like, for example, um if you had like w uh left right, you had wet and then you had another type of like synthetic. um You know maybe from like. I don't know synthetics or something like that. um You know those two things are going to be highly correlated.

C

I mean pretty much it's going to capture the spread, there's not going to be much loss. um You know, as opposed to uh you know, bitcoin, you know rap bitcoin versus uh uh you know. You know they might not be as correlated and then you can think about things that are just literally anti-correlated right and then that would be where employment loss is the greatest, um and so I think you know you know when we look at like the stable coin, uh amms uh like, for example, the ones on curve um for the most part.

C

It's like that, though, with I think one caveat, which is that you subsume a lot of tail risk um now, maybe very low probability tail risk, but you do subsume a lot of tail risk uh in that you know all of these stable coins are correlated with each other until there's a peg break right and egg break may not come for another like three years or five years or whatnot, uh but when it happens, one thing can just completely get divorced and correlation from the other uh and you're going to be subsuming.

C

All of that now, maybe what you make along the way right and all the sort of bit-ass bounce that you're capturing along the way over those five years more than compensates you for taking like, on that day, a fifty percent loss right, because maybe you've already made like seven hundred percent eight hundred 800 over those five years.

C

So you know uh that's definitely possible, but I think um you know it is important to know that you know it's not purely a safe thing, just to trade, two things that right now look super correlated and theory should be correlated. Maybe there is some tail risk to consider that tether intrinsically is a little bit different than die or is intrinsically a little bit different than usdc, um and then um the other thing that I want to talk about is that um you know if we look at um you know, I agree with you.

C

If there was no fee, then all you all you take on is in permanent loss. You'll always be worse right because you're always passing the market. If the market moves up, you're, always holding the asset, that's worth less.

C

If the market moves down, you're also always holding the asset, that's worth less right, but the fees compensate for it and really, I think what it is is that in the literature, the way they they talk about, this kind of phenomenon uh is very similar to the idea of why you wouldn't even post up a limit order in the first place, even on a clock or a central uh limit order book right.

C

The reason, if you think about what that limit order is effectively, it is a sold option to the rest of the market to buy or sell from you at a certain price right now. If the market was perfectly informed, they would only do that if you're on the wrong side right, if the intrinsic value of that thing is higher or lower than where you're pricing it they would take your quote, you would take losses basically right um now.

C

On the other hand, if the market was completely uninformed, then even with a one penny spread, your expectations go positive right, because the market is entirely noise. You quote one penny spread you're collecting half a penny every time, someone trades on in expectation, um you don't have to be very small marketing right. So a lot of it just comes down to how informed is the market?

C

If they're 100 informed, you can't even market make, because you always just get picked off no matter what, um if it's 100 uninformed, then even with a minuscule, uh infinitesimal spread, you can always make money in the long run, so the market is somewhere in between um fully informed or fully uninformed, and therefore it even makes sense to post that limit orders. uh You know in the book in the central limit order book now. I think a lot of that same phenomenon is happening with the amms.

C

In that there's, a lot of uninformed flow, which is then subsidizing the liquidity provider for all the times they have to take losses when an informflow hits them right. So as of right now it does seem very profitable. It does seem like there is more more a greater proportion of an informed flow in terms of um the fees that it pays, the uh liquidity provider on unit swamp. Then the the liquid provider loses, in you know, permanent loss due to infection.

C

um But what I think what my point is is that I think over time, that will change and as the size of these, the creepers gets greater and greater. At some point, it makes more sense for me to go hit that at the same time that I hit you know if I wanted liquidity up or down right by yourself to hit that and the centralized exchange at the same time build it into my small rider to hit everybody at the same.

C

At the exact same time, nobody can react right and then, at that point I get the best feel possible and the whole thing is pretty toxic, because I still not order across multiple venues right so, like I think um you know when, when when there's more sophistication that develops on the taking side or liquidity, then we'll, I think, we'll see the character of the flow turn more toxic and more informed, and then we'll see liquidity providers pull back because they're they start taking more losses than what he gives them.

C

So that's sort of what my narrative is and what my story is for how I see it playing out, but you know, obviously I could just be wrong. I mean it could just be the whole thing. You know, there's always enough unemployment flow, always enough retail flow. I mean at least today people think of it as a fair system.

C

So as long as that, that's the case whether it is or not doesn't matter if they think it is, um then there's always going to be a lot of people on both sides providing equity and trading against the equity providers. Right, um I think over time as a narrative changes. If it does change, maybe then we'll start to see a kind of race to the bottom. Everybody starts pulling out on both sides. So sorry, I just started rambling, but that's sort of what was yeah.

B

I don't know, but I I.

C

B

You made a really good point about it's like there's, actually, the way that am, I designed it's actually quite similar to how order books and marketing design from like a motivation standpoint. So, as a market maker, you.

A

Always want to.

B

Chase the markets where they have a maximum of uninformed flow right, because that's the best market for you and I think, liquidity providers and am should also think the same way and it's like it's like, and I think it's what's what's. um I, like you brought the temporal aspect because I do think that it's like there are certain times when certain markets and certain amount markets are really great for liquid riders. Maybe it's like the first time that you know uni, like um some new asset has listed on emm and they're, also liquidity mining.

B

So there's a lot of like uninformed people out there. Some people are trying to buy it. Some people are trying to sell it. It's just like you know like a lot like a lot of volatility in the market, a lot of um a lot of tracks. A lot of you know activity, and it's that point. That's when you want to be at llp, uh but after the thing settles, the death settles and now it's like everyone's bought. The token has bought it.

B

uh Everyone sold it sold it and now you're, basically facing the market you're facing mostly arbitrage or coming in to the price.

B

I think at that point it may not make as much sense to be an lp as it was earlier on in that markets history, um so I think, there's a tempo element to it and I think, there's kind of like overall, my I haven't tried this, but uh I I feel like, as if I really running pawn fund, I would try to like almost like um toggle between being lp versus being arbitrage and just like.

B

You know basically, however, like if I think the market's better for the lp I'll be nlp, but I think the market's mature and it's more, you know just floating on arbitrage of flipped arbitrage.

C

Yeah, I think that makes a lot of sense and I think, um even though, uh as an lp, you can't actually control where, where you put up your prices and you can't differentiate your prices from other lp's in the same pool as you, um you do have something which is a unique action that you can take to be different than other lps, which is that you can pull liquidity and provide liquidity at any time that you would like right so, like, overall uh being an lp in, like let's say, unit swap pool you're, basically short ball and short gamma right.

C

uh Just to explain of all means volatility. Gamma is uh well, it's determined options, uh but you know it's basically, uh your short large moves happening up or down right. The jumpy large moves that go down um so uh generally, when you want to be an lp when it's best to be an lp is when you think um volatility is gonna uh continue to be very low right. But if, but, if you think all of a sudden uh tomorrow, volatility is going to spike, then maybe that's when you start pulling.

C

um You know all of your. uh You know. Lp shares out of the pool, and then maybe you switch on the other side and start doing the arms right, because then you know that uh the short ball people are gonna start getting picked off, as the price starts oscillating like wild right, there's another consideration too, which is that um as of right now, it's been happening a little bit, uh but I think not not to its full extent, which is you know.

C

This is all this debate about minor extractable value and I think, there's a lot of merit here, which is this idea that at the end of the day, the miner is the one who sequences all the transactions on chain right. So um if there's a huge charge opportunity, why should they? Let you take it right? Why not replace your name with their name and they substitute for you in that transaction and take the art themselves right. um So I think um you know overall there's two ways that miners can extract value.

C

One is like this direct way, which is like replace name with their name right. The other is through uh competition within between arbitrages right now, if you're an auto charger and I'm not a charger, we see that there's ten dollars on the table for the taking now we both need to get into the next block. Maybe I'm willing to pay five dollars as a gas fee uh to get that included in the next block? Maybe you're willing to pay six? Maybe then I'm gonna pay seven. You know pretty soon, I'm willing to pay 99.99.

C

You know uh 9.99 right and then at that point you know I only capture a penny, all that value all that excess value of the arm actually goes to the minor right. So I think there's a lot to be said about you know how how viable is in the end game, when you know everybody becomes sophisticated, the miners, the you know the arbitragers, the market makers, the takers, you know everybody becomes sophisticated.

C

Is it even possible to even do art on these amms that the miners, you know or you know, can't just take themselves, yeah.

B

I I actually would argue it is and the reason I think it is because I think I think the miners can really capture uh it's like. Yes, I think if both legs are on the ethereum blockchain, then absolutely miners are gonna like take all those arps, but realistically, most of the time only one arb is on ethereum.

B

The other arb is an essential exchange, and so, if miners only see that most people only see that one leg of the trade it's hard for them to arm because they don't know what you're, how you're hedging that trade.

C

Right yeah, I I agree that in that case they wouldn't be able to take it, but I think in the end game the miners themselves. They all have trading desks at this point right now: they're, not the most sophisticated trading desk because they're not specialized in it, but I think over over time they can at least specialize in this one particular strategy, um which I think you know it's like see. A low price here see a high price. There you know, buy the cheap one sell the expensive one.

C

I mean eventually they'll be able like within the next five years. Probably they'll be able to do that right. So.

B

Yeah, that's true, that's true, and also if they were running hummingbots, they would already be doing this. Oh.

C

There you go well, they should just do that. You know kind of they could just cut in line and just have it done. You know that's right.

C

They probably have some very good deals in terms of fees with decentralized exchanges because they probably pushed so much uh mining flow through there too right. So you know they're, probably on a very good feat here uh for that too. So uh so I think you know a lot of interesting questions. You know hard to say, but uh you know those are my thoughts. Yeah, okay,.

B

All right: well, it's been an incredibly. We can go on with like for hours literally hours, um but uh we're only five minutes left. I want to make sure we as well answer some some um some. You know questions to the audience. So let me let me go through a few questions that people have asked. um So one of them is um say someone has: can you guys elaborate on risk parameters used when managing other people's money? um Sure yeah? So I would say we only ran a small fine, at least for us.

B

It was like it's very simple. It was like you know we were. We were trying to basically ensure um uh I think we were trying to ensure something like less than um like 20 annualized volatility. So that was like a target, uh but uh it was, I would say it was it was hard to manage. So I would say for us it was a little more like finger in the air type of management in terms of, but I think for us. We we knew that our our investors were um more risk averse than we were.

B

uh That's why we're they're investing in us. So for us we wanted to make sure that um our users had no more than like a 10 drawdown uh every month. That was like our biggest fear.

C

Yeah, you know, I think um I think, there's a lot of ad hoc stuff and there's also you know some quantitative stuff. I I would say that a lot of it is uh more of an art than a science like you know, just as an example. For example, you know you have all these different metrics. Like you look at this, you know. Back-Tested strategy has like a sharp ratio and a sorting ratio, a trainer ratio, information ratio. You got all these ratios well, which one matters more right, like you could say.

C

Oh the sortino ratio matters more because downside this is well. You know worse, you'd, say sharp. You know you could talk about leverage scalability stuff, like that, um you know, but overall the judgment still comes down to the to the manager on what they value and also inferring what the investor cares about right.

C

You can have all these numbers, but at the end of the day you know how you even weigh them between each other uh matters a lot too and then, on top of that, you know, there's more ad hoc risks right, for example, if you're, if you're putting up at most 10 of your capital on any individual exchange, then that's different than putting up at most 30 percent of your capital, any one exchange right. So let's say let's say you have ten exchanges.

C

Now, arguably you could say: okay, I'm not gonna put more than ten percent of my capital on any one exchange. Now, if any exchange gets hacked, um then you know we can only lose numbers like ten percent or maybe even less than that, because there's some recoverability uh value right.

C

But now maybe now you have a lot of frictions right because now, like all of a sudden, there's a huge r between hobby and uh binance- and you know you're out of one asset on one but you're, and but you have more assets of that on another exchange, uh but then you're the for asset number, two, you know you have it all on one, the other exchange, but not on the first exchange right. So now you have to do there's a lot of operational rebalancing. You have to pay the gas costs.

C

um You know, there's also the hassle there. So you know it creates frictions which may lower the return. So now we've got work that maybe keeping 30 percent. It's fine. uh You know what is the ethical chance of an exchange getting half per year like? Is it like for major exchange? Is it five percent annualized? Well then, you know, maybe that's what you can take into consideration. You know um you know, do you, uh you know some other considerations like. Do you take that smart contract risk?

C

You know your you, let's say: there's a huge um yield form that's uh just released yesterday and the yields are insane they're like the four-digit five-digit yields. You know, and you know then, in two days that yield is going to be like double digit yields right. It's going to it's going to lose two digits or three digits in two days right, but you don't have time. You don't have time to go. Read over all over the code. You know everything about it.

C

You understand its full mechanics um at what point can you say I've done enough due diligence. This is a. This is fine to put five percent of our capital at risk. At this thing, that's earning absurd uh yield. uh If we lose it, we lose it. But you know, but how much like how much research do you do right? Like do you for one hour? Do you research, five hours right all this time? The yield is decaying right right now and a lot.

C

A lot of it is just judgment right and a lot of it is about what your uh lp is, what your investors will accept right and ultimately, it's their money, and ultimately, you have to reflect what their preferences are right. So.

B

Yeah yeah this.

C

Is too, they all have different preferences too.

B

Sorry, I want to make sure I get to the last question because we only have like one minute left, but I think the last question is a good one. um What's the role of a money manager or hedge fund manager when d5 automates trades, uh so the example is token sets, but I think we can also use wi-fi for this. uh So so like it's in a world where there are on-chain hedge funds, um you know: what's the role for like normal hedge funds,.

C

Well, I think um the difference between most of these on-chain hedge funds and uh what you know most I think, quant shops do- um is that the on-chain stuff is fairly simple. You know mostly it's like, um if you think about wi-fi, for example, it's a robo advisor. um You know it's constructed pretty well, uh we were actually very long wi-fi uh for for quite a long time um and did well there, but I think overall, you know, that's not really an actively managed strategy right, we're not really looking at.

C

Like historic data about you know, back testing models using neural networks, we're not using like very sophisticated things, because at the end of the day, um you know all that logic has to go on chain right, like that. Robo advisor has to have logic coded in solidity for how it's going to operate right. So you know for the most part, I think you know it. It has to be kept very simple or also you know it's gonna be very hard to automate. How do you retrain a model? How do you recalibrate type parameters?

C

You know there's all these these sorts of ideas. um I certainly don't think that you could have like brentec right renaissance technologies models codified into solidarity and made into like an on-chain kind of platform, because so much of the value is derived from the secrecy right. The secret smallest is that so that nobody else can copy you. If you see one particular strategy working very well uh on ethereum. uh Well, you should just clone over that code right and you should make sure your own modifications.

C

You should learn something from that um that output disappear very very quickly as more people pile into it, uh and also on top of that, a lot of these polls are not apps right. So it's just like okay. Maybe this thing can make like 50 uh a year on 10 million, but once you once everybody starts piling into it goes to 100 million.

C

uh Well now you start to see like what happens right, and you start to see that already a little bit with the wi-fi uh with all their vaults right here, all the y vaults. You see that with live core, you see it. You know all over the place. You know you see the yields coming down because everybody's chasing it right everybody's, putting in their capital saturating out the market. Now, if you could have something, that's more protected, a private strategy, a secret strategy. um Well then, you know at that point you know.

C

Hopefully you can keep it going for as long as possible. Alpha will still decay, but you know hopefully longer than you know, in the public domain.

B

Yeah, what's coming with, uh you know aztec and project protocols, so uh it could be possible in the not too distant future.

C

That'd be interesting, you know if one of these protocols comes up with this kind of like obfuscated code or some kind of privacy, you know zero knowledge, proof or whatnot. We might give it. Oh, we might, we might code up our own. uh You know automated strategy and put it on. You know that might be pretty cool. It might be fun to do.

A

Yeah definitely all right cool, so we are out of time uh we need to transition to the next session. uh Thank you guys so much for this. This is really great.

C

Thank you thanks for having me.

A

Yeah, of course,.