Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Rbit: Bitcoin Mining, Mt. Gox, Stat Arb & the Future of Crypto HFT
132 Plays1 month ago
Rbit joins the Insilico Terminal Podcast to talk about his 15-year journey through crypto — from discovering Bitcoin on 4chan, mining BTC in 2011, losing funds on Mt. Gox, and trading the early BitMEX era, to building statistical arbitrage and HFT systems. We cover the 2017 ICO cycle, Bitcoin fork trades, why 2018 pushed him toward quantitative trading, what stat arb actually means, how HFT signals differ from longer-term strategies, why smaller venues can still offer major edge, and how crypto market structure is becoming more competitive as exchanges move closer to TradFi infrastructure.
00:00 Getting into Bitcoin in 2010, mining, Mt. Gox, and coming back in 2017
13:37 ICO mania, BitMEX, and the Bitcoin fork trade 17:14 Why 2018 pushed him toward quant trading
24:52 Building a profitable stat arb business on FTX
40:21 Stat arb vs HFT, and why crypto got more competitive
49:09 What his trading operation looks like today
55:13 The future of crypto market structure and where edge is going
01:05:04 Advice for newer traders and why quant thinking differs from retail
Recommended
Transcript
Introduction and Early Crypto Experiences
00:00:03
Speaker
you
00:00:13
Speaker
Welcome to a new episode of the InSilico Terminal podcast. My guest today, but what do you call yourself actually? Cryptohades or? arbi Arbit.
00:00:26
Speaker
I see, I see. So you're you're my guest today and I like to ask people to introduce themselves. And with you, I also don't really know that much about you, to be honest. i've been You've been suggested as a guest and I've listened to the episode that you did with Worst Contrarian.
00:00:42
Speaker
um So can you introduce yourself for us? Hey everyone, thank you for having me on first of all. ah Yeah, my name is Arbit. I've been in crypto for about almost 15 years now and into Pond trading um from 2019 onwards.
00:01:00
Speaker
onwards So, yeah, um I got into crypto in around 2010. I first started hearing about it on 4chan at the time I was in high school, just know lurking, and it was mentioned a bunch of times. And then about 2011, it was September started I tried to get into it. I started to mine it.
00:01:23
Speaker
yeah um It wasn't very user-friendly at the time, so it was kind of hard to do. um And it was already pretty competitive because people think that before, like 2013 or 2014, it was just free money, but actually it was that was more 2009 and the beginning of 2010.
00:01:41
Speaker
In 2011, it was already pretty competitive. it was really hard to mine an entire block. Yeah. Or even just get any decent payouts from pool mining. and So I started by trying to mine a block for some months. Of of course, it didn't happen. So I started joining some some mining pools, switched to GPU mining. every the time I was like playing video games, I would just, whenever I had idle time on my computer at night, I would leave it on and and mine. So it wasn't even really continuous. But it was an interesting experiment. You know, the whole community was very different than what it is today.
00:02:15
Speaker
And then in October 2011, came out. I tried to mine a little bit of that as well. But at the time, you didn't know what what was going to happen with it, right? So it was, Litecoin was like a like a mysterious shitcoin at the time, right? He didn't know if he was going have any future.
00:02:35
Speaker
And so i just bounced back to BTC. And at that time I didn't even worry about you know backing up my my hard drives and my wallet because you might um mind 0.1 Bitcoin, but it was worth cents. So it wasn't even worth it. What was the price of Bitcoin at that time? I think the first time I actually was able to see a price, because it wasn't so obvious at the time, ad came out on the iPhone. and like the iPhone 3G, you remember that one?
00:03:04
Speaker
who And an app came out. The UI was like Windows 95. It was terrible. And you could see the price of Bitcoin and it was like it's slightly below a dollar. It was like 78 cents or something. oh whoa And that was the first time I actually was
Trading on Kraken and the Mt. Gox Impact
00:03:20
Speaker
able to see it. And so, yeah, and basically it was worth nothing at the time. so I wouldn't even worry about backing up my wallet or just, you know, whatever, just money it again or...
00:03:30
Speaker
I really care. Of course, that was quite a bit of money in hindsight. um And then around 2013, the first ASICs came out. There was these little USB ASIC miners that you could plug into your computer and they will give you a bunch of asterisks that was better than you your CPU because it was ah made for mining. It wasn't very powerful. It wasn't really competitive at all, but it was very cheap. You could buy one for maybe 30 bucks or something on eBay. And so I used a couple of those.
00:04:05
Speaker
and um And then when basically in September 2013, Kraken came out and it was the first place where you collapse and at least of my knowledge where you could actually wire in a bunch of money and buy crypto And, so at the time i was getting to freshman year university, I had, I'd start buying like little 50, a hundred bucks during there and deposited my little money stack. And, and then I moved it to to um empty box because that's where all the action was. It was kind of like the.
00:04:39
Speaker
i don't know, like the bitmets of that cycle or the FTX of that cycle. Yeah, maybe more FTX. yeah So we'll trade on there. And I remember BTC going to 1000 for the first time.
00:04:55
Speaker
and was like late in November, it was winter. um And I remember it was a very interesting moment at the time. I didn't have that much in, but, you know, I was a broke college kid, so it was, it felt meaningful.
00:05:11
Speaker
And then a few months later, Empty Cuts went down, as we all know, and um at the time I basically, you know, rage quit. because i was
Return to Crypto and the ICO Boom
00:05:20
Speaker
you know I didn't have that much money to buy again. and i was pretty It was all pretty negative at the time. a lot of people at the time even took their own lives and were very very Like the the mood was very, very negative about Crypto at the time.
00:05:38
Speaker
It was a very dark moment um because there weren't that many exchanges. Like when FTX went down, there were, you Binance was huge. there were all sorts of other venues. So people bounced back. At the time, there was nothing.
00:05:50
Speaker
Right. and And so... yeah did did did you like lose all your money yeah on Mt. Gox? Yeah. I see. Yeah. And recently got it back. But, you know, it was not that impactful anymore.
00:06:04
Speaker
Yeah. But mostly I just rage-created the space because, you know, it was just such a negative experience. Pretty bad event though. Yeah. I will just monitor like every every quarter or so. I'll just open TradingView and take a look at the other prices and just make sure they didn't go to like 10K without me.
00:06:21
Speaker
Just, you know, when it started climbing back to a thousand in early 2017, I got that can. At that time, after that three years, I... I got a job at that point and started getting a little bit um of money so I could buy back in And at the time, Ethereum was going from like 100 to 300.
00:06:41
Speaker
was its first, well, not its first, because of course it was much cheaper before. But, um you know, basically Bitcoin was pushing about a thousand again and ETH was pushing to 300 for the first time.
00:06:55
Speaker
And so there was a lot of excitement around it again. And so i bought back and I remember I will try to margin spot trade on Kraken at the time.
00:07:07
Speaker
which was not a very pleasant experience. And and then I immediately found out about BitMEX, which was a relatively new platform at the time. Perpetuals were a brand new thing.
00:07:18
Speaker
And it was brilliant. The UI was brilliant. All the the only user experience was incredible for the time. and so it just deposited some Bitcoin there and just spent two years exclusively trading on BitMEX.
00:07:33
Speaker
And then... um
00:07:37
Speaker
ah Give me second. Were you like a true believer from the beginning when you found it that early? Were you like into the whole decentralization? Well at the time I was... It was a very cyberpunk movement.
00:07:54
Speaker
i was never really into finance. i didn't I never dreamed of being a trader or punt or anything like that. I was much more into, you know, arts and kind of the opposite of that. But Crypto did feel like um like something new and exciting and that could actually change how that world kind of works. and kind of went against the rules, you know, it felt like you could independent and kind of withdraw from whatever would see at the time as kind of the word of boomers and and ah you know how your parents do things, right?
00:08:34
Speaker
So it's kind of a rebellious attitude applied to to finance, which was really unique and really interesting. So back then I was just a kid, but that was kind of interesting to me.
00:08:45
Speaker
I think it's kind of crazy. i don't know if I've ever talked to someone else before that has been in crypto for this long, because there's always like this meme of a being early, and then we're we're all early. we're like I mean, I guess now we're no one is really early anymore because like institutions are here and whatever. but
Market Challenges and Shift to Quantitative Trading
00:09:00
Speaker
um for me like you you reached the point in your story where i first like kind of got into crypto in 2017 and um it's it's a bit insane to realize i that so much stuff has already happened before that that i that i yeah well i still felt kind of earlier but like you were like really really early and even you didn't feel early because mining was already not the the like quite competitive so i don't feel early at all
00:09:25
Speaker
you know, the people got in really early could hit a faucet and get 25 bitcoins for free. Yeah. You know, or, you know, crazy stuff like that, or mine tens of thousands of them from their home computer.
00:09:37
Speaker
So i it didn't feel early to me. um But yeah, of course, in hindsight, yeah. And also most of the people who got in early, like me or earlier, did not make necessarily make that much money because holding was much harder when something is worth just cents and there is there is barely any community about it and and there is no way to even easily convert it to fiat and You know, you just get associated to like Silk Road track pillars and stuff like Like it didn't feel like the future necessarily. When people said, you know, this thing is going to go to a million, going to to a hundred thousand. It didn't feel like a forecast. It felt like a very, very, very distant bike dream. Crazy idea. Crazy idea, yeah. So barely nobody, I think, held on to that money. And the people who did hold on, they are not able to sell.
00:10:35
Speaker
yeah If you believe in it that much that you held for 15 years, you're not going to sell. you're not better Why would you ever sell at that point? yeah It's very, very rare for a person to to be both ah to have that vision at the time and then held for so long and then actually be able to sell and to kind of balance it out and and have a...
00:10:56
Speaker
you know, maybe convert half of it to the Tradfly assets or whatever and just live your life. It's usually they're extreme kind of people that get that early and actually held all the way.
00:11:08
Speaker
i think we we kind of look at people uh like like kobe or maybe you to to i guess you're not as known as kobe but like to to people that know you you're also like quite an og so we we look at people that have been in crypto for very long as if like they've had it easy you know because it's like you you could buy at extremely cheap prices you could buy eth ico and all of these things and blah blah blah but stuff like Mt. Gox might have just like taken you out completely if you put all your your savings in there and it's it's just gone and like all of these struggles that like were there throughout the years like we don't even really think about anymore because they're kind of just like a being forgotten or whatever but even just like the even in 2017 it was kind of difficult like I just I was still quite young but it was very difficult to get money on the exchanges and like I remember I bought, ah I don't know if people will know what this is, but I guess yes in Europe kind of people know, but I bought Bitcoin with PaySafe cards.
00:12:06
Speaker
What's that? you know what PaySafe card is? PaySafe? No, PaySafe. It's like something where you could go to to a gas station or whatever, and then they have these cards and it's basically like a gift card, but like like a Steam gift card or something, but you can use it like everywhere. And I sometimes bought them to like get stuff from League of Legends, like buy League of Legends skins. And then I also like got those to like buy Bitcoin with it before I had like a ah proper bank account or like PayPal account or whatever.
00:12:34
Speaker
Mm-hmm. Yeah, and lots of people use tricks like that. Unfortunately, by 2017, I kept using Kraken because it's really good service for if you had and off ramping. But yeah, in general, the user experience, it was terrible back then. Yeah, yeah, yeah. It's so much worse than today. And we'll talk about it more later. And it's it's still not that great, so. Yeah, yeah.
00:12:56
Speaker
But imagine like before Uniswap, DEX trading was like EDX and ETH for Delta. And the ETH transaction fees were like $300. And the wait times were so long. And there was no, like UI was so ugly. and There was nothing to, nobody to reach out to if it didn't work. There was no community around it.
00:13:17
Speaker
twitter comes Twitter was not that big. It was, you know, everything was so much harder. And I think even Kobe, and I don't know him that well, but I think even Kobe made most of his money from 2017 to 2020. Yeah, I think so.
00:13:31
Speaker
I don't think he was a huge whale in 2011 or so. Yeah. No, I don't think so. I think he also made his money later.
00:13:40
Speaker
So we we got to up to 2017 and you like getting back in into crypto. how how How did it go from there? What did you do? Well, first of all, I i lost that era of like the Ethereum ICO in 2016 and all that golden era.
00:13:56
Speaker
But then in 2017, the ICO mania was really, really taking off. And looking back at it, it was kind a huge phenomenon, but it wasn't that... long journeys of time. it was actually, most of it was probably about nine months or a year, because by the end of 2017, by last quarter, it was already extremely difficult to get into ICIOS. They started putting all sorts of limits to it. You could only You know, there was like one chance in a hundred that you were going to be selected. And even if you were, you could only deposit to like a thousand bucks or even 300 bucks.
00:14:28
Speaker
So it was really difficult to make any money with it by the end of it. But at the beginning of it, it was gold mine. It was so easy. You could just throw money pretty much anywhere, um except the ones that I did.
00:14:42
Speaker
No, it's a joke, but like Tesla's for example, which was like the biggest one, I yeah put a lot of money into that one, ah day i in percentage terms, to my network. And that one dragged out for like two years, two and a half years, and then it it never went back above basically its ICO price.
00:14:58
Speaker
Really? Yeah, it barely
Profitable Trading on FTX and Statistical Arbitrage
00:15:00
Speaker
did. Oh, I didn't know that. It basically sold a break even for two years, but it locked up so much of my capital for so long. Yeah, I can of imagine. And every other ICO was like, you wait two weeks and you sell out 50X, 20X.
00:15:13
Speaker
So that was kind of a golden era. But most of my trading in 2017 was actually on BitMEX. Like in DeFi Summer, a lot of the trading was in Uniswap and the trenches were different, but in 2017, the trenches were were basically the XBT contract on BitMEX.
00:15:30
Speaker
but The whole community was there in the trial box just fighting it out in your book. And the spreads were huge, the the fees were huge because BitMEX had the maker-taker model that was basically 7.5-dip staker and and make a rebate for everybody. And the spread was so big and in business points.
00:15:48
Speaker
So it was very expensive to trade, but it was super ah volatile. It was super fun. So my core strategy for 2017 was just trading the Bitcoin forks. Because there were was Bitcoin Cash, Bitcoin Gold, and I think Bitcoin Biomog.
00:16:02
Speaker
And every time there was a fork, it was so inefficient because Everybody will long or because you will get um ah one to one ah for your Bitcoin, basically in the new fork. And so I will long and then it will be a website with a countdown in blocks.
00:16:20
Speaker
And then at the last block, i will just 20x short. And then on the fork, it will go down like 15%. It will trigger the server ah the the API overload and the liquidation cascade. And then I will try to long the bottom, basically. yeah And I will just 10x my money. like three times over. So it was like a free short because everyone immediately sold their new whatever Bitcoin cache or... I see. I mean, the Bitcoin cache not even because it was just a snapshot. So you will actually will see the Bitcoin cache a bunch of times later.
00:16:52
Speaker
But everybody wanted to be in the snapshot. I see. Yeah, that makes sense. And that was the expectation because it was public English. So everybody kind of had that idea. So you were just kind of front-front basically.
00:17:03
Speaker
And yeah, that was a my moneymaker for 2017. And then 2018 was much, much, much more difficult. was just price going back down and up, to basically ranging from 6K to like 12K.
00:17:18
Speaker
And then for three months, it flatlined at 6K. And it was so tough at the time because it would literally just go up a tick every few minutes and go up like 10, 20 ticks over an hour and then go back down.
00:17:34
Speaker
in one tick and then it will just keep doing that for months and i was trying to like they trade that it was was awful and so that's when i kind of decided that yeah i will i would rather try to to take a look at the quantitative side of things or auto trading at the time. I didn't know what the font even was, but I thought at least if I learned how to do that, it's probably more profitable. It's better on my um my mental health and it's more productive because at least like even if I lose all my money, I still have some
00:18:10
Speaker
some skills that I can use in other fields. You know, I can learn how to program. I can learn how to do data analysis or machine learning models, you know, all this stuff that I can just, you know, it's ah it's a hard skill. It's not just doing a chart.
00:18:22
Speaker
Did you have a job back then or like were you still in university or was it all that you did? No, actually, I was quite successful at my job, which was in a completely unrelated field. So I was actually pretty lucky as well. And then I kind of used that my savings from that job to get back into computer science and and kind of re-roll my career into creative and still trading.
00:18:46
Speaker
I see. And you you just, so i um I imagine back then the the resources weren't that plentiful for for learning stuff like that. So, and you're, I guess, entirely self-taught.
00:19:00
Speaker
how How did you manage that? Well, there were a bunch of books that i read back in the day and um A lot of i love the most valuable things I learned, i learned from people on Twitter, actually. some Some people that I was lucky enough to to meet and and they believed in me and they helped me and mentored me. And ah for example, MicroSapplePod is somebody who helped me quite a lot over the years. He's a great, great guy.
00:19:25
Speaker
And so, yeah, it was basically just trial and error and then going up to people with more experience than me and saying, hey, I tried to do this, this, and that, and these were my... conclusions, but I wasn't able to solve this problem.
00:19:38
Speaker
ah Am I looking in the right direction?" And they would be like, yes, but you forgot about this. And what if you look at it this way? And just you know nudging me in the right direction and the and then doing the homework. But there was no AI or elements at the time. So yeah i think he was i think just trying really hard was more valuable in a sense.
00:19:59
Speaker
because you really, really had to do it. that There was no shortcut. You had to do your own work. And so I think people are also more keen to help you out when they saw your ah you know your your hard work.
00:20:15
Speaker
And now I guess it's you know it's really hard to tell if somebody just put prompting to Claude came up with something or if they actually put in the effort
00:20:26
Speaker
How long did it take you until you were like successful with quant trading or profitable, I guess? A couple of years. um So I started actually deploying first systems in 2019. So my very first so my my very first uh, tries where basically i was a retail trader back then. So basically I was trying to do some kind of TA inspired things and i was trying to automate them. And then immediately I discovered kind of alternative sampling schemes. So you you could sample your candles by volume and
00:21:01
Speaker
and and stuff that is not time. And so I started experimenting with that. I thought that, you know, I don't see anybody talking about this, so probably, you know, it's alpha. It wasn't. um But that was kind of my first attempt. And that taught me how to, you know, make a website that's going connect to an exchange, pull data and sample data and try to make a simple model and simple features and execute some trades. So all the basics.
00:21:25
Speaker
And um After that, I tried to do pairs trading, which I'm going to talk about later because it's connected to the startup.
00:21:37
Speaker
um Basically, I would trade each BTC or LTC, BTC, like relative value kind of stuff that was a little bit easier to predict because it was basically a just one very strong downtrend most of the time with some spikes up when when price went up. but It was mostly pretty you know pretty steady.
00:22:01
Speaker
and So that worked out much better. And I started deploying some capital to that, but ah operationally, it wasn't super easy to manage because you know managing servers and AWS, it has a learning curve as well. So I had some downtime and have to learn a bunch of lessons about that.
00:22:20
Speaker
And then in 2021, I met my current partner who came from a commodities startup background. And so he kind of showed me
Market Dynamics Post-FTX Collapse
00:22:30
Speaker
how he used do it TradFi.
00:22:32
Speaker
And so we kind of merged the approach. I i understood, OK, that's actually what I'm trying to do. um i didn't even know it had a name you know and so we started doing that on ftx and when we called ftx um which became kind of popular after the the code crash in 2020 because of the liquidation engine they're basically on bitmax it was it was i don't know if you were there during that no i wasn't just i've just heard heard the stories about it Yeah, it was pretty... awesome For me, it was one of the most funds ever had because I remember we were in our in a group call in Discord with a bunch of my friends, creator friends, and we just stayed up all the whole night.
00:23:13
Speaker
And it just went down to the 3K and the server got shut off. And then it came back on. for like half an hour, it was a 10% spread on Bitcoin and nobody would cross it. And it was just a mess. Yeah, that's crazy.
00:23:26
Speaker
And the liquidation engine on Bitcoin was pretty... pretty unforgiving and on FTX instead there was much less leverage in the system and the recordation engine was a little bit more chilled so they got a huge um boost in user base after that.
00:23:45
Speaker
did they Did they already have tether repairs back then? Was that like already a thing? ah Yeah. Because BitMix only had the the inverse, right? Yes. No, FTX was just ah spot and perps. The nice thing about FTX also was the collateral handling. So basically you could just throw anything you wanted into your wallet and it was all kind of collateral.
00:24:04
Speaker
I see. So there was a and the UI was great, the app was great, so it was a huge step up, even if BitMEX UI was really good. And so you could trade spot and perp on the same subaccount without and changing your your collateral.
00:24:21
Speaker
Yeah. And so that was great. And you could do carry trades and bases on the same subaccounts so easily. And it was the same endpoint for for the API for both spot and perps and teachers. So it was all very, very handy from that point of view. And also, it was very easy to get zero maker fees.
00:24:43
Speaker
You could just stake, I think, 30 FTT or 100 FTT or something. oh whoa And you will get zero maker. And that was kind of hard to get from like Binance. You had to do a lot of volume. in So it was all these little details that really drawn people to the to the platform back then.
00:25:04
Speaker
and And so when we deployed the startup to FTX, we were immediately very profitable. And for 2021 and 2022, were doing nine. like sharp nine after costs, which is insane. We're doing 30% month on leverage.
00:25:21
Speaker
Yeah. Back to back. It was crazy. It was crazy. And we were not very advanced at all yet. and Our execution was just take care. Basically, you just take your orders.
00:25:33
Speaker
We'd significantly sleep with it well. But the the the alpha was just to too big. And after that, when FTX went down, we tried to copy paste the same system to Binance Futures. We were like, you know, it's the same asset. Why shouldn't it work? It didn't work.
Future of Crypto Trading and High-Frequency Strategies
00:25:50
Speaker
at all. And so for like six months basically we shut everything down and we reworked the whole system from from scratch. why Why is that that it didn't work on Binance?
00:26:03
Speaker
Good question. Well, first of all, the market regime completely changed. the Completely. It was a huge impact on the on the market. And ah I guess
00:26:18
Speaker
Well, i'm I'm not actually sure. I looked into it back then a little bit, but um it was probably a mix of just the microstructure being different. And we were probably unknowingly picking you up on some on some signals that were idiosyncratic to FTX as well, and not just at the AC level.
00:26:38
Speaker
ah And ah I don't know. It was probably a mix of many small details. But the most important part being that after that, the market regime changed completely and the the competition changed. Lots of people got wiped out some ah big trans-life firms that were trying, they were starting to trade on there, got kind of spooked and and backed out. Like, CGL, I think, was trading on Binance Futures, and they stopped after through that. And then they they eventually came back in.
00:27:09
Speaker
um So yeah, for for a while after FTX went down, the markets were very inefficient as well for some days or even weeks. And then eventually it kind of settled.
00:27:23
Speaker
so what like what exactly is statistical arbitrage? Can explain that for the plebs of us? Yeah. Let me start with a simple example. So let's say you want to predict the future returns of Bitcoin.
00:27:37
Speaker
There is a lot of external factors to go into it. There is the S&P and geopolitical factors and ah inflation and news. And there is so much going into it that it's really quite hard to to predict because you really you really can't predict more than a small percentage of it.
00:27:57
Speaker
even if you were really, really good. And so the basic idea was, well, let's take another asset like Litecoin or Ethereum, and these two assets have are very correlated and they have the same common factors. So if there is a geopolitical news or an ES move, it's going to move both assets in similar ways.
00:28:17
Speaker
Not in the same magnitude, not exactly at the same time, but mostly in the same direction, right? And so if you trade the relative value between two assets, it's ah kind of a very simple way to basically visualise away the common factors that they share.
00:28:34
Speaker
And so trading the relative value between the two assets lets you focus on the um on on less or much less and lets you express a view that is much simpler. So, for example, I think BTC is ah an asset that has much better fundamentals than Litecoin.
00:28:52
Speaker
And so I think the long term over the next year, BTC is going to either go up more or retain more its value than Litecoin. And so I might long one and short the other.
00:29:04
Speaker
Long Bitcoin and short Litecoin. And that's a very simple way to to explain it. And then if you expand that to a bunch more assets, let's say you want to trade Bitcoin, Ethereum, Solana, Hype and Litecoin and Doge maybe.
00:29:19
Speaker
ah you might want to rank them and you might assign some some values to that according to your understanding and your evaluation of the fundamentals. So you might want to long hype the most because you think you know it's a very strong token. you know We don't need to to say anything about it. Everybody knows it. And then Solana has been performing really well, you might put that one second. And then maybe Bitcoin, then maybe Ethereum and and those on Litecoin.
00:29:47
Speaker
past And so if you then kind of weight them or distribute those in a kind of normal way, you get some weights. um And that's an example of a a factor that TradVy will be called quality minus junk.
00:30:04
Speaker
So you long, high-quality things, and you short, low-quality things. yeah And that's kind of like a tilt. in You call it a tilt. So it's something that doesn't really change over time.
00:30:15
Speaker
It's kind of a static expression of value, of relative value, in a cross-section of assets. Okay. And so this is an example of something that is really quite simple to do. You don't really need any tools to do it.
00:30:30
Speaker
ah You could just do it from your own intuition and knowledge about the space or from, ah you know, from statistics that you can get from CoinGeck or whatever. and And something that it's simple to trade because you might just want to keep the weights kind of kind of within bounds and rebalance that every week or every day. And you could literally trade this off of a spreadsheet. this might, if you do it well enough, can easily get above too sharp.
00:31:00
Speaker
So this is something that I would do if if I was just starting out, for example. And if you expand that a little bit more, you could find factors that are not just tilts, but they are kind of timing factors. So you could say, for example, if hype and soul pump way too much, ah for example, more than two or three standard deviations above,
00:31:22
Speaker
what Bitcoin and Ethereum are doing. And if Litecoin and Doge dump more than about the same compared to Bitcoin and Ethereum or compared to to the market in general, you might want to short Hype and Sol and long Doge and Litecoin for, let's say, a week.
00:31:40
Speaker
Okay. So that's an example of mean reversion effect. And that is more of a timing effect than a tilt. And so when you execute those positions, when you put on those positions, matters more than just the just your valuation of of the quality of these assets compared to one another.
00:32:01
Speaker
And a lot of the alphas you find in Stardard are obviously timing because there is only so much that you can do with ah just tilt. ah And so that's another example of something that you can do. Another example would be carry.
00:32:16
Speaker
So normally when you do a trade carry in a quantitative trading system, it will be and kind of screener that looks at all the assets and their funding rates. yeah And you look for funding rates that are really high.
00:32:30
Speaker
really low and that are persistent and on assets that have enough liquidity to to support your own trading. And so then you will position yourself to get paid the funding and you try to hedge your position either on spot or on perps on other venues that have the same underlying.
00:32:49
Speaker
And you will collect the funding. And in Starterb you can do the opposite. You can position yourself according with in the same direction of the funding if you just um are careful about assets that have kind of extreme funding rates and that are liquid, but mostly the funding rates predict future returns um because also funding rates are capped. And generally, if something has you know a negative funding rate, and
Building a Trading Operation and Career Insights
00:33:20
Speaker
it's generally... Basically, you you can write the direction of the funding rate and even after you pay the the funding fees, you can get out ah profitable.
00:33:31
Speaker
So you're using the funding rate as an indication if it's going to go up or down basically? Yeah. Yeah, you you can do that. And that's another example of another factor, another very simple factor you can do.
00:33:43
Speaker
And if you put all of this together and you do it not just on the six assets that I was talking about, but on, let's say, 200 assets, you got yourself an example of a long, short starter portfolio.
00:33:57
Speaker
And then, of course, you want to be careful about your exposures and you want to make sure that you are mostly either delta neutral, dollar neutral or beta neutral.
00:34:08
Speaker
You want to make sure that not single no single position is too big. You want to make sure that you are sizing your position in the right way. may For example, according to the inverse of the volatility of the assets.
00:34:21
Speaker
That's the really simplest way to do it. and um And then you can just rebalance this position over time. And this is a very simple explanation of what a startup book may look like.
00:34:31
Speaker
And then if you get fancy, it gets to hundreds or even thousands of signals like that and fancy ways to mix those signals and fancy optimization pipelines to make sure that you're getting the most out of your signal mix or predictions.
00:34:50
Speaker
How many factors or signals did you use in your trading? um More than 100, less than 1,000. But I guess if if something like the example at the beginning where you talked about fundamentals, then it's not all like quantitative, if it's also like discretionary, at least in part.
00:35:09
Speaker
Yeah, you can so there is ways to do that. For example, a way to do that in ah in a way that is more quantitative, I guess, is You measure the correlation of each of the as of these assets to to Bitcoin.
00:35:21
Speaker
And generally, the stuff that is more correlated to Bitcoin is higher quality, and the stuff that is less correlated to it is lower quality. Another proxy is size. So you go by market cap.
00:35:33
Speaker
So you long the stuff that has a very big market cap and you short the bottom stuff. And so these three examples are mostly proxies of the same thing. So in any way you do it, you're going to get similar results.
00:35:47
Speaker
And if you should look at the P&L of these three factors, they're going to be very correlated and they're going to look very similar. And you could, for example, mix them into one. um But they're essentially just three ways of expressing the same thing.
00:36:00
Speaker
I see. That makes sense. So how do you go about Even though you will call the market cap factor, you will call it size and not quality minus junk. And it's really, you know, mostly they're very similar concepts.
00:36:15
Speaker
So how did you go about like building systems around that? like how do you How do you find new signals to look at? and i don't know how how much you can reveal about this ah or like how much you still do in trading, but like which which signals? said Maybe maybe that you can give some examples about like what worked in the past that doesn't work now anymore that you like found interesting.
00:36:38
Speaker
um So when I build systems, first of all, I don't really stare at charts. I don't really trade manually anymore. For the last five years, at least.
00:36:50
Speaker
um It mostly comes from intuition. It's mostly thinking about the data that you have, or maybe you buy some new data, for example, and you start thinking, what if...
00:37:05
Speaker
ah this measurement of this data could be informative about the future in some kind of way, or could give me some, some intuition about, you know, um,
00:37:19
Speaker
future returns in any sort of way. So um one example would be the the funding. So you will just think, what if I wonder if funding rates predict returns.
00:37:30
Speaker
I wonder if they predict returns kind of equally across different styles of the universe, for example. they do so equally at the top with majors than at the bottom of the the deciles.
00:37:44
Speaker
And I wonder if they can if they can beat the funding payment costs. And I wonder if they're stable enough and how often do they change? And and so you will just get the data at that point and it kind of start plotting stuff and and diving into it.
00:38:01
Speaker
Or an example could be maybe... let's say there is a sudden increase in volatility, right? How does that impact future returns depending on the asset so a sudden volatility increase on bitcoin might signal future returns that are nearly positive of their next over the next three days let's say or a week or whatever and at the same time the same sudden increasing volatility on a super low market cap shitcoin might might uh
00:38:35
Speaker
might say that somebody's trying to pump the price or it pumping and dumping the coin, and that might signal that that coin is going to perform worse than anything else because it's going to dump back down 60% or more over the next week.
00:38:53
Speaker
And so depending on the on thisci of this or the size of the coin or other qualities or interactions that you might find, something that you can measure, such as volatility, change over time, can have different impacts on the price. So that's just a basic idea and intuition that you will you will just think about well looking at data or you know, looking at whatever trading your data that you have on your research pipeline and you might have this intuition and you might try to measure it and build a signal around it. So you will try to express that idea in the cleanest and simplest way possible and then try to explore all the different interactions that it can have with with other
00:39:37
Speaker
data that you have and then kind of, you know, try to transform it in some way because maybe you you can transform it in a way that makes it more linear and makes you perform better, makes you have less turnover, so less cost.
00:39:49
Speaker
and So there is a bunch of aspects that you will look at and then you will look at how much it loads onto other stuff. So ah what if this idea is actually very similar to something else that I did in the past that didn't use the same data, didn't use maybe volatility as a measure, but it used something else.
00:40:07
Speaker
But actually, when you look at the ah the final result, it's actually kind of similar. So we are effectively measuring the same phenomenon in two different ways. And then maybe I didn't think about it, but then you find out that it's really not that uncorrelated.
00:40:23
Speaker
um
00:40:26
Speaker
Is everything that you do kind of like higher time frame stuff because you mostly mentioned like days and weeks ahead? No, so actually in startup you will mostly work with about a day, intraday to daily to a couple days horizons.
00:40:42
Speaker
I also do a fair amount of HFT and market making and did in the past. We started doing some basis, some market making on FTX. And then we did some small venue HFT stuff. That's the stuff that I posted on Twitter with the Crazysharp. And then we spent about maybe six or nine months market making on Bybit, which was pretty tough in 2024.
00:41:14
Speaker
um And then now we are gearing towards becoming market makers on Binance futures as well. ah We still have, I mean, the the market-making infrastructure that we have, we currently mostly use it as execution layers for startup.
00:41:29
Speaker
So we have really good execution for our startup waste, which is kind of a must if you increase your capacity. i mean, your AUM over a certain amount, because course you're trading like 100 grand, it doesn't really matter. But if you're trading 20 million, um it's it's pretty meaningful because your T-cost can very easily kill your edge.
00:41:49
Speaker
um so yeah that's So yeah, we do both and we're going to do more HFT in the future. i I do think it's going to become more tough in the future to pull that off, especially on on major exchanges because they move gradually more towards TradFi infrastructure. But we can talk about those there as well.
00:42:15
Speaker
yeah I was just going to ask a about that actually, not about the the future, sure but like about if it's already become more difficult, everything that you've been doing like Startup and HFT over the last couple of years as participants have become more sophisticated.
00:42:31
Speaker
Yes, but Startup is getting really tough as well. Quarter one, 2026, I think it was. I'm pretty sure it was the worst quarter of the last two years for a lot of startup shops and us included.
00:42:43
Speaker
It was really tough moment. HFT-wise… Why is that? like Why why is it has this been the toughest? Well, it was very, very influenced by geopolitics and ah news and wars and… I see. ah Yeah. and Most of the volatility, here I would say, was driven by those factors.
00:43:06
Speaker
So that makes it kind of hard and kind of toxic environment. Because it's less predictable. Yes. There's lots of competition. Funding rates are are very...
00:43:17
Speaker
um and very high and a lot of the stuff like altcoins, most of the universe, ah all the altcoins and shitcoins are just going down to the ground over and over and over. And so there's not much that you can do about that, but just short them, but also you pay tons of funding.
00:43:36
Speaker
Yeah. ah So it's kind of a tough position to be in right now. But I think it'll get better. And like HFT in general and market making?
00:43:48
Speaker
HFT also getting definitely more competitive. um Generally, HFT becomes easier when there is lots of volumes and volatility.
00:43:59
Speaker
There's more dislocations. these locations there is more i mean, the the the volume becomes less concentrated on on the majors and it kind of spreads down to altcoins and, you know, whenever you have a bull run it's usually the best time to be an hft especially taker um but also maker so yeah it's it's not the best time for hft either but you can see lots of very competent participants going down to very questionable venues ah to try to squeeze a little bit more alpha.
00:44:34
Speaker
yeah um Yeah, it's definitely a tough moment in that aspect as well. But, you know, it's nice to build during tough moments to reap the benefits when markets turn around.
00:44:46
Speaker
Why do you not not look to ah less popular venues but try to like compete on Binance market making? Isn't that like the most competitive market that exists in crypto, basically? Pretty much. um That's a good question. and i actually advise the opposite if somebody asks because table selection is is a huge huge user advantage in crypto and the market is so fragmented that you should take advantage of that. um And actually, the the crazy sharp thoughts that I posted on those were on kind tail venues, venue somewhere.
00:45:22
Speaker
and It was kind of ah of an odd place to trade where it was kind of hard to price those assets. I i can't... dive into too much detail but basically usually when you go into tail venues most of the time you're just trying to get the binance prices there as fast as possible and having the fastest api access as possible that one was kind of different breed it was a i can't say much but it was when you were pricing the assets that were listed was was kind of you couldn't do it from binance prices okay you had to
00:45:58
Speaker
to get creative. So that was our edge. It wasn't even speed, really. And so that was just a Python infrastructure. It wasn't particularly particularly fast or anything.
00:46:11
Speaker
Um, and that was a unique situation and decayed pretty fast, but before decayed, we were pulling like 40 X on capital in six months. Oh, wow. And it's really, really fun when you can, when you can do that.
00:46:25
Speaker
And also i will say, when you look at kind of scrappy places like that and you estimate how much money you can make, it's usually more than you estimate. Mm-hmm. So you will look at a place and and be like, you know, it's not worth my time. Actually, most cases it will surprise you.
00:46:42
Speaker
That number will be quite a bit higher than you'd think.
00:46:47
Speaker
So like ah what what what makes you turn to Binance? Is it just like size size constraints ah at your point? um Well, it's operational constraints. ah So when you trade small venues, tail venues, most of your time is going to be spending operational issues. So you're going have to maintain lots of different infrastructures for each one of the venues and it kind of make them communicate see and you're kind have to be a little bit custom for each one of them um especially if you're trading like perplexes and stuff that has a sort of unchained mechanics or batch auctions instead of continuous auctions like apple liquid so everything needs to be a little bit custom and
00:47:28
Speaker
most of the time you're just going to spend it on shitty documentation, APIs going down, um errors that are not documented, documentation that is in from in Chinese and you have to translate it with LLMs. So all sorts of issues like that. And that to me wasn't really what I was dreaming of doing, you know.
00:47:51
Speaker
It was either that or just spend most of my time doing research and trading on more reliable vennies. But it's much harder, but if you get it correct correctly, if you get it working, the payoff is is's quite big and you get to spend your time differently. We're also small team. We don't have a lot of manpower on the operational side. My partner and handles the operations, but um yeah There is a lot of stuff to do and is already kind of struggling to keep up just doing one venue basically.
00:48:27
Speaker
um So it's not easy to to pull off. but I see some people on Twitter, some smaller traders focusing on that and I think that's the right thing to do. i think liquidity goblin as well on your podcast on this episode. was mentioning doing lots of small things all over the place and that's a great approach. it certainly makes you money and it's certainly much easier than doing just buy-ins, but just do the way that you spend your time changes drastically.
00:48:51
Speaker
You're just getting woken up all night every time there is... yeah It's a lot more annoying to do. Yes, it's extremely annoying to do. it's just a choice. I think we...
00:49:03
Speaker
yeah we are willingly leaving money on the table to change the problems that we have to deal with with something else is uh your trading operation nowadays like a fund basically or find small like a fund u her um So, yeah, we trade our own prop money and we also do SMAs.
00:49:25
Speaker
So we do handle clients' money on their accounts. So it's a little bit lighter lighter on the bureaucracy side than just having a full-on actual fund. um But yeah, basically my my day-to-day looks like...
00:49:42
Speaker
Basically, we get daily performance emails with all the recaps from all the strategies and breakdowns of P&L by account, instrument, and analysis of transaction costs and funding rates and everything.
00:49:54
Speaker
ah Then I have some Grafana dashboards with the live monitoring of the P&L, the latencies, the fields, orders, exposures, and everything. We have a PagerDuty setup for emergencies and for any bug or error or any value that breaches some set limits.
00:50:12
Speaker
um Then we have InSilico open as well. We do InSilico for that actually because oftentimes, especially with clients, you don't have access to the account. You just have an API key.
00:50:23
Speaker
And so we will use InSillyCode as kind of a third-party monitoring verification layer. So something that doesn't depend on the exchange, but doesn't also depends on our code. So yeah if we break our own code, we just look at InSillyCode and just make sure that we're not hallucinating, you know, that everything is right. And also, InSillyCode, you can um intervene if something goes wrong.
00:50:43
Speaker
right So if Yago for some reason doesn't happen, but for some reason goes crazy and and gets a position wrong you can turn it off and then go and so and and just get rid of the position without having access to the account so i think liquidity goblin mentioned something similar that's also really like uses it in in emergencies yes it's really really handy and the ui is very responsive so that's good and uh appreciate that yeah um not a paid show and um been looking at logs a lot lms have made much easier to monitor that as well so if you have a nice logging infrastructure you can you know talk often just you know read these logs for today and make a report on how the system is behaving etc that's really handy and then most of my the rest of my day is just working on alphas and models
00:51:29
Speaker
I handle basically all the alphas and and models and strategies for for our team. and And so that's all me basically. So I spend a lot of time trying new ideas and trying to improve our existing existing ideas.
00:51:43
Speaker
And um for either HFT or startup, and it's really nice actually to do both things because you but they're kind of different in how you will approach alphas and and features and modeling.
00:51:57
Speaker
um And so it's nice to have a view on both and learn things from from both and be able to apply them to the other thing.
00:52:10
Speaker
How old does the difference show, if you can like expand on that? Sure. um Well, for example, I already talked about examples of startup signals that you might use.
00:52:22
Speaker
While in HFT, those things barely matter because the the timeframes that are so much smaller. They are maybe, you know, can be 10 seconds or a minute or that kind of timeframe.
00:52:36
Speaker
And so what you will look at instead is something like, for example, auto-booking balance. So especially on like Bitcoin and Ethereum and instruments that have most of the action in the top of book ah because of very small tick sizes and high rebates. And so you will look at kind of what's the imbalance, where do the other micro makers want to get filled at, and quote, what's their bias. And that's pretty informative already. It's a very strong feature, so you will look at that and it's predictive power over different time horizons and how it correlates to other things. and
00:53:14
Speaker
So the kind of kind of the scope of things, it's very, very ah different. So momentum, for instance, you might measure momentum in a startup and it might get really fancy and trying to do it cross-sectionally, but in HFT, it will kind of do it differently.
00:53:32
Speaker
and You will care more about the microstructure, you will think more about... how many levels is something moving in one direction, how much time compared to other things. And you will just look at the ah few seconds of history, maybe 30 seconds of history or so, and you will look at kind of with a microscope at how things are behaving.
00:53:51
Speaker
So it's kind of the opposite, where it's like looking at a herd of animals and then looking at looking at the at and the atoms that are making one of the organs of the same animal. And you're kind of looking at the same thing, but from completely different perspective.
00:54:07
Speaker
do Do you have an an example for a signal in HFT? Our broken balance is by far the most common and and very strong and and very intuitive signal example that I can make.
00:54:19
Speaker
and Then there is all sorts of you know momentum and relative value. You can see if something moved compared to something else that is correlated to it or across venues. So you might want to look, for example, at Let's say you're trading on Hyperliquid, you might look at the Bitcoin price a relative on Binance futures relative to Hyperliquid.
00:54:41
Speaker
And if you see it, it's part copy, it means he it moved on Binance but didn't move on Hyperliquid. And so that's a signal. So you kind of look at small things like that. see, yeah.
00:54:54
Speaker
you've you've been like you You don't really have like a a finance background or anything, right? Not really. I'm basically self-taught, yeah. I study computer science, so. I see. Yeah. that gives me like a little bit of uh research but i think it's like very it's it's very impressive or it's like extremely impressive actually that you've just built all of this from the ground up by yourself um like a plane to help from amazing people yeah if it was if i wasn't on twitter if it can twitter didn't exist i don't think i would be able to do it
00:55:27
Speaker
do Do you like ah see yourself doing this for forever or like for a longer amount of time? I guess we can also branch a bit into like the the thoughts on your thoughts on the future of crypto.
00:55:39
Speaker
Well, I'm not sure about myself. I think I'm going to keep doing it for for the foreseeable future for the next five years or so. I'm not sure I'll be doing the same thing in 20 years or so.
00:55:52
Speaker
um you know I enjoy the kind of the the challenge of doing this and the process of doing it, but it was never my dream to do this when I was a kid. you know So I guess it's kind of different. I see kids these days in high school saying, how do I maximize my chances of getting to Jane Street after college? Like, teach, you're 15, just calm down.
00:56:19
Speaker
So, it was that was never really my my approach. um and Regarding the future the my view on the future of crypto, i think I think I'm seeing centralized exchanges moving more and more towards a Troutify kind of setup, because also I think that affects well on their value from the perspective of VCs. and the ah So I see Tracking moved, for example, to Beaks in August 2025.
00:56:47
Speaker
ah So that's the the provider data center provider for lots of Troutify exchanges. And so they're moving away from the cloud. They're moving to more proven um ways to do things, which is generally better yeah for the end user. Latencies go down, it's more reliable, crashes less. you know and It's a net improvement, but actually for people like me, it's very bad news because suddenly financial firms that have been active in Wall Street for a decade can just mostly like copy paste a bunch of their systems into something they know very well because it's the same hardware that you're going to use for the New Stock Exchange or whatever.
00:57:32
Speaker
um so i think i'm not looking forward to it i think in a couple of years by 2027 2028 it's going to be very very difficult to have any alpha on binance and uh I see TradFi exchanges moving kind of in the opposite direction. So listing more and more ETFs and futures crypto and evaluating the possibility of going twenty four seven And it kind of kind of the the perpetual future idea is kind of um you know getting into TradFi slowly.
00:58:04
Speaker
and um them perpetxes and their bo batch auction approach is also interesting and is becoming very popular right now, Confeproliquid. And prediction markets are also huge and and they actually want to get into perpetuals as well. So everything is trying to do everything.
00:58:24
Speaker
And it's kind of merging into one big blob, which I think is probably good for the end user. Not so great for for me. um We'll see.
00:58:38
Speaker
We'll see. I think there's goingnna we're going to have to be flexible and ready to move to on where the opportunity is going to be. Yeah. Have you ventured into any TradFi assets?
00:58:51
Speaker
As they've gotten like a bit more popular the chain and anything recently? Not really. Or do you plan to? We might have to, yeah. We might want to expand for that. For example, one of the things that you can do on platforms that list both crypto and TradFi assets, in the example that I said earlier, where you might want to long high-quality stuff and short, low-quality stuff, you could do in the same book, you could long you know, SMP or Tesla or NVIDIA or whatever, and Bitcoin and hype and short shit coins as the short side or the short lag.
00:59:26
Speaker
And that's very powerful. I think that's a really good performance that you can have for very little sophistication and something that we'll definitely do if I was at retail right now.
00:59:38
Speaker
And um do you have like, I think you already like gave a lot of good good things throughout this this episode, but do you have like in general some advice for people that are just like trying to start it out now and then want to do something similar as you? And maybe also like some of the the mistakes and the lessons that you've been through throughout your 15 year journey in in crypto, apart from like don't put all your money on Mt. Gox.
01:00:09
Speaker
Well, that was actually the best holding ever. so okay My best trade. the um Well, it's one of the biggest mistakes is, i don't know if I can call it a mistake, but Underestimating and then overestimating the competition is something that I think happens to everybody when you're starting out.
01:00:32
Speaker
ah Correctly estimating the competition competition is very, very hard. At the beginning, I was underestimating them yeah because it was just a detail. um Nobody's going to do this use this indicator on this type of candle as well. No, that there was actually bullshit. that The market makers and HFT people were already super sophisticated compared to me at the time. And then um then i made the opposite mistake later on where I would have a system that was pretty good, but I would think, okay, there's no way that this is...
01:01:04
Speaker
ah competitive with other people. And then later, a couple years later, i will meet somebody who will tell me that on that same venue two years ago, they were making a crazy amount of money with a super simple system that was actually simpler than mine.
01:01:16
Speaker
And so actually um kind of trusting what I have and believing in it and um kind of just worrying about doing and kind of 80-20 rule.
01:01:27
Speaker
ah So worrying about the things that matter and just yeah making sure that the important things, you get them right. So latency and pricing and the important bits, and then not worrying too much about the rest and just and just get going, start trading, start putting some capital into it.
01:01:47
Speaker
And um you know don't be afraid and don't wait until you have kind of the perfect system because it barely exists. And um then I wish probably that I had um that I spend more time doing table selection again, especially when starting out, like in 2019, 2020. I think there were many things that I could have done that were very uncompetitive back then and that yielded very, very good returns.
01:02:17
Speaker
and Not going all-in on Hyperliquid was another big regret at the time. i had it Actually, I called with Jeff and he was asking me. Really? Yeah, he was telling the funding rates are crazy. Please come to Hyperliquid and trade on it. and You can't do this much APY just to carry. um that it was very you know it was at the beginning as well the exchange was brand new and i was like yeah that's awesome but you know i have so much to do i'll put it in my to-do list and i put it at the bottom and it never actually got to it um but yeah that's a mistake that a lot of people made um i would say for people up and coming
01:02:56
Speaker
um I will certainly look into prediction markets. I think they're going to become even bigger in the future and they're going to try to expand to, to to you know, perps and stuff probably as they announced last week. um So there's going to be a bunch of opportunity there.
01:03:11
Speaker
And I will try to get into scrappy stuff, especially right now that with LLMs, it's much easier to manage breadth over depth. So it'ss it became much simpler to make custom one-up systems to harvest specific things and maintain them yeah in a way that is doable for a small team of a single person compared to how it was five years ago.
01:03:34
Speaker
And it's actually with Rust, for example, becoming popular, it's easier to maintain even a high performance system or it's also easier to write trip pretty well-performing Python with Numba, Cyton and Bidim, you know,
01:03:49
Speaker
GIL version with LMs, so you don't need to to go you know to go crazy. um So I'll definitely into that.
01:04:00
Speaker
And the hard part at the end of the day is learning to think about trading in this way that is very distant to what you will learn as a retail trader and then system design.
01:04:12
Speaker
So getting things right in a way that is that allows you to have good performance, um good modularity, and the the right trade-off, and enough freedom and flexibility to implement any strategy that you might want to implement.
01:04:29
Speaker
um So kind of that's, I think, the hardest part. And it took me a bunch of rewrites to get it right. but What do you mean by like distance so to I reach a trade of things?
01:04:40
Speaker
um That you will necessarily... So for example, as a retail trader, you might think in terms of 200-day moving average or you know stuff like that. ah You will think in a single instrument terms. And for example, the things that I said earlier about startup and examples of relative value trades and...
01:04:59
Speaker
um quality man is junk all these concepts they are not they were not intuitive to me when i started they over time they they it became apparent that they were interesting things i dive into them and i learned about them but at the beginning it was not so obvious let's see and then also for hft t it's It's very intuitive. For example, a phrase that Rico says a lot is, um if you go too low with the timeframe, it gets too noisy.
01:05:33
Speaker
So you have to zoom out and it's actually the opposite. It's much easier to predict very, very small timeframe. It's super easy to predict one second ahead or five seconds ahead. The problem is monetizing it.
01:05:44
Speaker
And actually the more you zoom out, the harder it is because everything comes into place. Yeah. Into play, sorry. Geopolitics and and news and and just, you know, multi-year cycles and inflation presidencies and whatever, it becomes contaminated with all sorts of stuff.
01:06:03
Speaker
So all these shifts in mentality are, are I'll say, all important and kind of you have to go through them
01:06:16
Speaker
Yeah, makes sense. I think this is a good point to wrap it up. We've been through quite a lot of topics and um thank you very much for coming on. Thank you for having me.
01:06:28
Speaker
And goodbye.