Dorm Room Start-Up Defies Odds, Rises as Large-Scale Quant Fund · Christina Qi

published on August 2, 2020

Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield boys and girls what is up welcome to episode 198 on this episode is a special guest and someone who in 2018 landed a spot on Forbes 30 under 30 of finance Christina she in her early

20s after graduating from MIT Christina and two classmates Lou : and Jonathan Wang who had collectively been trading futures from a campus dorm room founded Dome yard it was 2012 and dome yard was to be one of the few

High-frequency trading hedge funds this is a little unique because the majority of hfts operate as proprietary trading companies as Christina states during the interview by all means we should have failed given the three of them had

Little trade experience and minimal insight to the hedge fund industry not to mention the three had almost no money Saddam Yard is also one of the few hedge funds that went out and raised money from VCS such as Softbank

And Aaron tech co-founder as a necessary means to launch now eight years on the boston-based fund is will established and will past the startup phase demure trades thousands of times daily sometimes exceeding 10,000 trades across

Various futures products and during its biggest single-day this year the funded turnover in excess of seven billion dollars this episode includes talk of raising capital building infrastructure early hurdles

Trading strategies generating alpha and plenty more it was a real pleasure to speak with Christina and even inspiring to hear what her and her team have built from scratch it's just incredible so please enjoy here is my interview

With Christina Chi for episode 100 and ninety-eight understand it you had been trading European futures in the night with two other classmates while you were in college or university how did this turn into the borrowed idea of

Starting a hedge fund that's a really great question we let's see yeah we were trading from a dorm room and you know it was interesting because we were surprisingly doing pretty well for a dorm room strategy I guess I mean we

Would make we were only I think one of our one of my co-founders he was at Harvard at the time and his roommate there you know Harvard us all these rich kids apparently and apparently one of his roommates his family runs a hedge

Fund actually and they were able to give us a little bit of capital initially to trade and it wasn't a lot in you know as like a hundred thousand bucks or something but we were able to make about forty thousand bucks take home forty

Thousand basically at the end of the day which was a lot for a bunch of little kids in a dorm room and and then we realized oh that's good you know for him he was like oh wow this could be you know a really cool thing to do and would

Be really exciting you know something that keeps you going and keeps you you know waking up in the morning and for me the reasoning was more because I had interned in finance and I had a couple of particularly well one particularly

Poor experience in the industry where there is a lot of politics you know despite being a trading floor and most of the results were based on you know you're judged based on your performance but there was still a ton of politics

Surprisingly it's amazing how much politics is in every industry so I was really bothered by the culture and the atmosphere there the way they treated their younger people on the team as well as some of the minorities on the teams

And the women and minorities and and I just didn't feel comfortable in that environment and ended up you know for me that was a big part of the reason why I wanted to start off on my own I don't think I've ever said that in quite that

Detail before but there was a lot of hazing and a lot of behaviors that I just you know didn't want in a company and didn't want in a daily environment so decided to start off on our own you know

I wish I could tell you Aaron I wish I could tell you like oh this was the best industry for me you know I had researched and bred so many books and did all this stuff like no I didn't I didn't really I was not an expert in

Machine learning or high-frequency trading or finance at all at the time it was just purely driven by a desire to to have a company with a better culture on my end and a better healthier environment and people always think it's

Ironic you know you're a high frequency trading yeah how can you be ethical and moral and it's like well you know we've had a really great environment for the people on the team at least and and we can definitely talk about that more as

Well so yeah so you interned at a couple different places you weren't a fan of this did you you have two other co-founders at Dome yard did they have somewhat similar experiences to you oh we all came from different backgrounds

We actually weren't friends before starting the company together but we did have one thing in common which was the same career goals in mind we all wanted to do something you know related to finance or FinTech and all cared about

The startup industry as well so we did have different backgrounds so one of them had a more math and physics background and the other one was a computer scientist and but we were all young we were all inexperienced so it

Definitely you know is helpful in certain ways and that we had different skill sets and backgrounds but also unhelpful in that we are founding team we didn't have someone like a thought leader in the space you know or someone

With any amount of experience so it was really an unknown path that we were kind of going down that we had never been down before mmm so this capital which you had initially that you were trading around

With how are we trading that we're using algorithms right from the get-go or was it kind of clicked trading how were you making trades and European futures I was a great question I mean we were literally using like Interactive Brokers

You know we had set up some you know very wonky quantitative based strategies they're still systematic they still had certain triggers that we would use you know we might do a basic mean reversion strategy in our basic momentum

Strategist things that you can implement relatively quickly and easily the holding we were were we entered a yeah we were actually intraday by that time already but they weren't high-frequency in terms of the amount of times they

Would trigger you know is not that much during the day so you know it's impossible to create an hft business overnight or in a dorm room it's impossible but we realized that the faster we were the more of an advantage

We would have and surprisingly the more fair of a playing field it was for us which I know sounds ironic but you know the fact that the only thing hindering us was just being a little bit faster you know and that we didn't need any you

Know we didn't need any insider information we didn't need any you know kind of broader sources of data rather than just market data that's all we needed and so that was really great to kind of see and we decided you know why

Not continue making this seeing what we can do on the technology side to improve the latencies we had and back then you know we didn't even know and we knew it was high frequency trading but um this is before flash boys came out by the way

So this is before the reputation of the industry tanked so it was a different time period you know altogether and we weren't driven by you know really anything that was written in the book because that the book hadn't been

Published yet as you've pointed out you had minimal experience trading and in finance like yeah sure you had a little bit you also weren't will capitalized but you had this idea to launch a hedge fund how did you actually set out to get

This off the ground you're right we had no money I also had no credibility and no connections at the time which was not very helpful whatsoever by all statistic by all means we should have failed I think we were just driven by passion

Really was the biggest thing is it just we woke up every day very early and said about doing this and you know it worked every waking minute and that that's really what what drove to get somewhere I think there are a

Couple things we did do differently to make sure that you know we would either have an edge or we could at least maintain some operating runway one was raising venture capital which for a hedge fund is almost still almost

Unheard of but it's becoming a little more common you know there's funds like you know quanto peon for instance back when they had a hedge fund you know they also raised venture capital as well to fuel their unique fund we've also done

The same thing we raised venture capital as well from some great folks in the industry who who believed in us and wanted to maybe pass the baton or see what we were made of you know whatever it was but um that that support was

Fantastic it also gave us the credibility that we didn't have at the time that you know desperately needed so so that was one way that we you know went about things the other was building all of our technologies from scratch as

Well which is both good and bad I don't know if I could really recommend that to anybody today because the the world is a different place than it was even five years ago or eight years ago it's a different scene today in terms of what

The technology landscape looks like but there was not a lot of high frequency technology out there today there is not a lot of open-source either so all we could really do is scrap it together which meant actually the side effect of

That is we have our own IP we have a lot of value valuable stuff in-house that we can always capitalize on one day or you know maybe we could decide to sell it one day if we wanted to you know said that that was really great also if

Something goes wrong we can go in and fix it ourselves I don't have to call you know a hotline and wait on the line and stuff so so that was great but the bad side the downside of it is it takes forever I mean it was an

Unknown timeline an unknown path and an unknown timeline which is scary it's like the last thing that you want to do is not know when you're going to launch you know as a hedge fund and it's it's extremely frustrating happy to elaborate

On that if you'd like to yeah I'd love to ask you some questions around raising capital and also building out your infrastructure just before we do though I have a question about why you were fixated on high-frequency trading from

The outset I wanted to combine finance and technology and didn't even know what fitnah FinTech probably wasn't a thing back then but I have a more technical background same with my co-founders and

You know during some of my internships we would basically me I'll make guesses on to whether let's say we'd guess whether Google stock would go up or down tomorrow and I knew that wasn't the type of strategy I wanted to do I want it to

Be a little bit more mathematical and you know that's when you figure out oh there's you know you could do some market making and stat are based strategies and have a more a more systematic approach to those strategies

And so so that was the biggest impetus of doing I guess quant and then when we were doing quote were like oh wait you know the faster the the more money we made and I know that sounds like a really I

Mean to some people it's a really selfish reason but that was really the biggest thing was wow we could really do this and and being faster we could really start to compete against the big guys in our space and you know make a

Name for ourselves hopefully one day so that was a naive goal back then okay okay now on the subject of raising capital I guess you kind of explained it a little bit but I'll just ask you again why did you need to raise capital and

Was that capital like the investors funds were they intended for building out your infrastructure and building out the actual fund or where they intended to be those funds intended to be used as trading capital so I'll answer your

Second question first which is this is meant purely for operations so management company operations like paying for employees buying software getting data licenses paying for data feeds you know all those expenses that

And legal fees and you know whatever service providers any kind of expenses that lead up to the launch of the fund itself I think the other thing is like if we could have raised money if we could have raised enough money in the

Hedge fund on day one maybe we wouldn't have to have you know raise venture capital to begin with and sell you know why sell a chunk of your company right that's a pretty risky thing to do but um we

You know knew that on day one given our backgrounds we weren't even qualified or capable of raising enough money in the hedge fund and so the biggest thing we needed to do because we're high-frequency trading as well we

Couldn't just launch the fund overnight we had to build we decided you know to build the order management system feed handlers execution gateways you know market engine simulators like all those different tools pretty much from scratch

And so that was gonna take a lot of time and that's why we needed venture capital we told them we were gonna use it to build those technologies and then eventually you know get to a stage where we can launch the fund and raise a

Separate amount of money for the fund the investors are actually largely separate there is a little bit of overlap there's some investors from you know like maybe firms like Renaissance or somewhere where they are both angel

Investors and also you know hedge fund investors too so they might serve two roles but for the most part our hedge fund our locators are a very different group of you know they're fun to funds or family offices very different group

Than the you know traditional VCS and you know other high net-worth individuals that we raised from on the operating side so so we did do that and then I forgot your first question pretty much explained it it was just why did

You need to raise capital oh yeah I'm desperate right in times like this when everyone's in a crisis – right it's like if you if you need the money would I rather own a hundred percent of a tiny pie or what I'd want to expand that pie

And known a small slice the big pie and I was like well for my career path right now you know I'm not that greedy and I'm just starting so I was willing to own a small slice of a big pie if that meant stability for the next couple of years

That's really interesting how there's kind of two classes of investors which invested in your business so you've got the the venture capitalists which were involved in helping to launch the business and then you've got your more

Traditional type investors who are invested in your trading performance is that correct right that's correct yeah so we had two different types of investors we had the long-term venture capitalists who were in it for the you

Know it's like feels like a marriage right because they're in it for like that ten year you know six to ten year haul basically and then there's the kind of shorter term investors on the fun side where you know they could redeem

Any time they wanted to essentially okay can I ask you how you secured your first investor and investor here I'm talking about the VC side of it given you heard very little track record and you know to be honest not much to show how are you

Able to get someone on board on the venture side is actually easier to answer because of entry side literally you just go to the startup events and pitch you know those events usually at the angel on seed stage people invest in

People right they shouldn't be investing in numbers like revenue figures because you shouldn't really have any and if and if they do judge you by that it's a very inaccurate indicator of future success because most companies you look at like

Twitter or I don't know uber you know certain companies like that they didn't have any revenue until much later on and some of them haven't even you know a lot of firms haven't broken even yet even though they've gotten millions billions

Of dollars in funding so so it's based on people at that stage which was great but then when it came down to raising money for the fund it was a different story it's almost like separate strategy in terms of you know having it was a

Little bit more about showing some numbers and we couldn't really because we didn't have any numbers at the time to really show for high-frequency strategies we had to you really just do what we can so like we would show we

Would show like some back tests for instance and then have to have a ton of disclaimers about like you know here's we only did one back test you know we don't over fit we're not trying to you know show you the best one and although

Most investors probably didn't believe it anyway they're like you know they've never seen a bad back to us before right so it's pretty much useless but we still you know tried to explain our methodology we explained what we were

Doing we did offer day one investors more lower fees and higher controls so more liquidity and that helped a lot to you and also you know higher communication channels in terms of maybe I would send

Weekly emails to certain investors so they would you know feel more I don't know like feel better about you know what was happening or at least understand that we know what we were doing that week and then you know

Obviously offering them a discount on the fees and lower you know better liquidity terms so they could maybe redeem at any point in time maybe it get rid of the lock-up period you know things like that that helped a

Lot and one thing we realize is even if we got rid of lockup here you know no not a single investor redeemed you know within let's say 1 or 2 months or even within a year usually when an investor makes decision like that if they're good

Allocator they're in it for the long haul but they just like having that free option available in case there's a pandemic you know in case there's some event where they actually absolutely need the liquidity so but yeah that was

Really helpful from day one to kind of get that and then there also what's interesting Aaron is there also are quite a few investors out there that particularly allocate to day one funds you know and they do expect the kind of

Lower fees and they expect you know maybe side letter or whatever but um you know the the purpose is that they are interested in taking that risk with newer managers and they might be interested especially in the quant space

I'd say a lot of they're interested in that new wave of quant managers who you know they view as kind of geniuses from a dorm room basically so hopefully people like us I don't wouldn't call us geniuses but you know just folks from

That kind of background and give them a shot and see you know if they have that next big thing it might be a little bit hard for you to compare but did you find there was perhaps some additional interest in your offering given that

There are very few funds out there which actually offer investors exposure to hft strategies it's a good question absolutely yeah especially surprisingly I would say after flash boys came out that was when

We got the most amount of interest ironically you know there were a lot of allocators who had read the book and they were like wow you know let's google this what are the high frequency trading funds out there and it turns out every

Other the you know the big high frequency trading funds they all changed their language on their webs now they're you know they'll say electronic market making or low-latency trading you know they'll try to get rid

Of that bad reputation and image to associate with high frequency we were the only firm that wasn't afraid at the time to just proudly say look we're high-frequency trading you know I've never bought order flow or frontrunner

Background it'll run a single person out there and but we do high frequency trading we're exactly the definition that you would expect of what high frequency trading is because we're trading literally in the you know

Nanosecond resolution here getting data at that you know at the high speeds and then trading thousands of thousands of times per day you know recently like this year the the highest we reached in terms of volume per day was 71 billion

So you know that that is pretty much the definition of what high-frequency trading is and so we were doing that we were like well we're hft so people Google us we show up you know SEO we're we would show up at the very top Without

Really Trying and people just reach out to something like hey you know we heard about you can we schedule phone calls so so that happened as well after a certain point in time and was relatively you know nice and helpful I mean obviously

People would always have concerns like hey you know what do you do front run and like how do you handle order flow and things like that and we would just simply say like we don't you know we don't buy order flow at all and it's

Just not the type of trading that we do and we don't you know do what they described in flash boys were there blasting hole you know through mountains under the river or whatever to try to wire between different exchanges back

Then you know we literally were just trading within a single exchange actually so we didn't need the speeds between you know transatlantic cables or whatever we didn't need any of that kind of stuff so it made it a lot easier back

Then and then you know over time as well the I think a lot of the myths in the book have been debunked or explained by folks over time so most of the investors have gotten a lot more you know educated about our field which has been helpful

To I'd like to spend a little time speaking about how you build out the infrastructure so you mentioned earlier that you built everything from the ground up from scratch internally is this unique for HFT

Fun to do so I wouldn't say it's unique it's something that you know especially if you're starting out in an industry that's relatively still relatively new at the time or just hasn't quite been commodified yet you know it's something

That you kind of just have to do to get up and running I will say today because after flash boys came around actually there are a lot more startups and I started doing hft technology you know maybe like feed handlers for hft or you

Know thinking about open sourced kind of tools like that as well so that has made it a lot easier to get to launch for a lot of folks and usually today when you know startups have a small budget I just recommend look out there see what's see

What you can use that's either you know cheap or free or open source and and surprisingly there's a lot of stuff out there so you know that I guess centuries-old question of bio versus build to be honest I felt like we didn't

Really have a choice but to we can even if we wanted to buy we couldn't afford it so I build it was on our end right right I mean oh yeah it was probably a lot of cost upfront and a lot of time spent but does it feel like that's kind

Of optimal for you now like that's the you know the ideal scenario is that you own all your own software now it's great today it was really awful before we launched I mean you know there were many frustrating days because we just had

Such an unknown timeline of like we had to we had roadmaps he tried to map it out you know investors demanded to see roadmaps and budgets and things like that and we could try our best to do the roadmap and then it felt like I don't

Know if you know like Zeno's paradox maybe I don't know it was like one of those paradoxes we fell into where one task turned into to to turn into four and it just kind of grew and grew and you know you would think a trading

System sounds so simple and then you look at all the nuances and details and all the data sets and every single nuance and every data set and and just all this stuff and it was quite a lot to deal with for you know a small team at

The time so yeah I was definitely unexpected and quite frustrating to be honest before getting to launch I think every every company underestimates how much time it takes to get to launch unless you're like you know if you

Non-quantitative or if you ready know what technologies you're using then sure maybe you can hook it up overnight and you know start trading but um even then it's it still takes time you got to set up the legal structure of the fund you

Got to figure out you know who your auditors administrators etc they all have to be ready to turn it on with you at the same time you know so all those things combined made it quite a lot of work on our end and you know pretty

Frustrating two and a half years before we got to launch yeah now this might sound like a bit of a naive question but just for anyone who might not understand what's the right word here might not understand what goes into building and

Hft business which is probably a lot of folks what did you have to build like what took two and a half years what all the different components that had to be developed there's so many you know it's like pretty much a type of we had to

Build the entire trading system and ecosystem from scratch so everything from because we need direct market access right so we're literally placing trades ourselves at the exchange level and so because of that we had to build a

System you know order management system basically system to be able to place orders to see orders to track orders you know to figure out what's going on to cancel things and that also included risk management tools that we had to get

Basically verified by all the different you know brokers exchanges that everyone had to make sure that our risk checks were enough and that they were stable and that if there was some you know a crazy event that happened that um you

Know our systems wouldn't go rogue and so we had different ways to check those types of strategies as well to make sure that we didn't accidentally you know tank the markets ourselves single-handedly or make any big errors

Or mistakes on that and so so that requires quite a lot of work it's very different from a tech startup by the way and that's one thing I've noticed was you know with tech startups you can launch an MVP and it can be bare-bones

It can be broken it can be pretty bad and that's okay you know your clients early-stage day one clients should forgive you for that because they know you're a startup when you're trading it has to be perfect you can't have any

Error any room for error at that level because you're literally if you mess you know you're done for you don't have a second chance in this space so that kind of process of making sure everything was perfect definitely took a

Lot of patience and you know energy and effort and a lot of you know checks making sure all the checks balances whatever I'll see how it was in place as well yeah so that was a big thing and then you know how do you process data

Right it's another big question of what are how do you get the data that we you know we wanted we had data directly from the exchanges but then how do you process that data and turn that data into an actionable you know some kind of

Value that we could extract from it and so how do you do that as well you know that took a lot of time to figure out and to process and to and then you know data pipelines we they would always break we would always have some kind of

Issue that happened and so figuring out how do we timestamp that data how do we process that data was also a big question that we would have in mind as well yeah in general it's a it's quite a nuanced process and it just had to all

Be relatively perfect on day one or else you know hedge funds rarely get second chances I would say unless you have unless you're famous and your name is already known and order you try to make yourself known but I'm

Back then we weren't so you know we literally just had one shot and knew that we couldn't mess it up so a little kind of scary but I'm where we do feel fortunate and lucky to have made it this far today hmm and sign that what gave

You the confidence you know after two and a half years of building us out that when it came time to switch everything on what gave you the confidence that the strategies were actually going to be profitable oh man

To be honest we we did it it's hard to say you can do as many back tests as you want we had a simulator as well you could we could simulate you know strategies and orders and fills basically in the live environment –

Which was great but even then you know it's hard to replicate a hundred percent what's going to happen in reality so we did as much as we could test as much as we could and then one day I remember the FIR

They'd be launched we were just kind of like you know what screw it let's just let's just try it let's just do it you know we got to take her we were very risk-averse at the time too and we're scared hey you know to do something

We've never done and we're just like you know what let's just let's just try it turn it on see what happens and hope for the best and that's literally you know what we did and thankfully things didn't blow up on day one thankfully nothing

You know crazy happened and we were like okay this is great actually we could you know we figured out you know we needed that kind of feedback for the markets and were able to fine-tune the strategies even more and just continue

Going from there so but yeah it was definitely a relief to finally have launched and to be able to tell people okay we finally launched we're no longer you know felt like I was a fraud all the way up until we imposed and so we're

Like oh we're finally a real hedge fund you know and we can have a real track record to show people which was fantastic you're in business yeah just recapping those first few years looking back on it what would you say was some

Of the you know the breakthrough moments which she had and also some of the unexpected challenges which she came up against there's so many I don't even know where to start I would say like when we first started the business we

Wrote down you know some of the things that we wanted to save money on and some of the things that we wanted to spend more money on before getting to launch because we had such a limited budget but there were some things that were worth

Hopefully you know we thought were worth spending some money on like a good lawyer was a good example or maybe finding a good broker you know they someone with reputation as well auditors etc we could splurge on and then you

Know things that we valued like we want it to be a flat organization back then everyone was called partner and you know see and then it actually did not go too well and I have so many reasons for that and we wanted to have perks almost like

We I don't know we're a bunch of naive kids right so we literally thought about like let's make this like Google like a Silicon Valley company with like free food and unlimited vacation and all these great perks and we quickly

Realized like those are both really bad ideas on the organizational side you know you don't want to call someone we hire like literally like head of hft from firms like Citadel and you know get Co back then and other firms and a kcg

Back in the day as well and they would come in and be like wait a second you know why do I have the same title as this kid who graduate not you know not necessarily me but like what about this kid you just hired right who

Just came out of school you know why is he or she called a partner we were like oh gosh you're right you know we realized that people actually felt less equal because of that rather than more equal which was pretty bad and then in

Terms of perks as well just didn't work out you know people abused the perks like crazy and it was kind of unfair made it more unfair and gave people kind of almost giving people too much choice sometimes is a bad thing so things like

That and then you know also splurging on things like lawyers and brokers and stuff we realized quickly that um it also wasn't worth it to hire just the biggest companies out there you know that it's okay to go for a smaller as a

Small hedge fund that you know it's okay your investors aren't gonna look down on you for hiring a smaller lawyer you know so long as they are still qualified to do their job right so just things like that we made a lot of mistakes on to

Anything of what else we we did so much wrong that it's just hard to even begin to imagine so I'm just trying to think of what else there was oh I mean we had a settle we started off in my apartment actually after graduation we worked in

My apartment for some time until we were kicked out we were evicted and the reason why we were evicted was actually kind of a random read all the cops knocked on my door late at night but I'm what happened was the we were we had

Built our server cabinets literally in the closet of one of the like the master bedroom basically and we rigged our own electric you know basically we rewired the electric system and built our own box and stuff and I think Jay weighing

One of my co-founders he actually electrocuted himself in the process so we were clearly not experienced at all we should have hired a you know a unionized I don't know person right to do this and we were doing some pretty

Sketchy things we ended up you know we're trading so much we're having we had so much going on on the server-side that we were literally using up all the electricity in the building and you could also you know people thought that

We were like we had a marijuana farm or something they like literally people we're running some kind of marijuana farm and so that's why the cops knocked on or thinking like you're expecting to

Find us growing pot or something or you know harvesting drugs whatever it was and they came in they're like wait what is this they saw there's computers and monitors everywhere and they're like what are guys some genius hackers like

If you thought were hackers it's like no no no this is you know this is a small business here and you know I did live in the other bedroom so technically it's my whole my home business right and they're like no this is not going to work you

Gotta you gotta get out of there and get a get a real office and then the next day we literally packed up and moved to we work back when we worked first open in Boston at the time our rent shot up we were used to pay 4000 per month for

Rent and then we work it was 12,000 per month and this is three times as much and the room we were stuck in you know we work they pack you like sardines in these tiny little rooms and literally it was the size of the server closet that

We were that we had at the apartment so it was just what a crazy just a turn of events like that that really changed Jeannot changed the environment and stuff but lots of adventures like that that we were willing to go through and

You know finally we're very lucky to finally get to a point where we could launch I guess what a story that's hilarious if we look at Dome yard where it is

Today I'd love to just ask you a few questions around kind of some trading metrics if you will so can you tell me how many trades do you average per day currently of course it varies a bit just on

Average yeah recently 2500 ish a little under 2500 trades per day which recently has not been a lot we've done more more in the past you know depending on the market environment and stuff like that but in

General it's in the up in the thousands you know and completely automated and people are you really high frequency tray you know this is high frequency trading you know they're all intraday and holding periods anywhere between I

Mean I know it sounds like I'm not marketing and we're not fundraising just gonna clarify that right now so that people don't you know as CC doesn't come after me again as I'm just saying we're not

Fundraising but uh you know we holding period wise anywhere between you know let's the split second you know like less than a second microseconds all the way up to hours these days and we consider the hourly you know holding

Period so we caught we call the midterm strategies but to every single one of our investors that's still high frequency to to them because mid frequency is like weeks so days you know it's a different we live in kind of

Different universes it feels like a little bit but but yeah happy to answer any questions as much as I can without getting in trouble okay so are you flat at the end of each day like you don't carry anything overnight we are we used

To do some overnight stuff but um you know realize that our nation was too in the the high frequency you know me I guess intraday type of strategies and we also figured that the additional edge we would get from holding overnight

Wouldn't it was mainly a constraint that we put on ourselves to be flat but um it's just something we've gotten used to over time I would say there's really these days we could technically do overnight strategies again but it's just

Something that we haven't really done yet for various reasons of you know just profitability risk you know all those different factors that we need to keep in mind if we did that okay that number you gave before two and a half thousand

Right she for some reason thought that number would be a bit higher oh yeah it could be up to I don't know how much ten thousand you know twenty five thousand it really depends on what we're trading and how

The markets are reacting and how much you know what the liquidity profiles look like and all the other things like that too so you know sometimes it's just actually during the virus you know the the volumes did shoot up quite a lot

Because there's you know lots of speculators lots of day traders and and a lot of institutions everyone's just very active right now so it did shoot up a little bit but for now you know it's down to the 2000s the number of trades

America is that somewhat correlated to the the volatility of the market questioned us kind of yes to know depends again on what kind of strategy were implementing you know usually during very high volatile you know

Periods like like this basically we tend to have a lot higher volume and the hope is that we perform better and can maintain some kind of neutrality to the markets so that's a that's definitely the case for I would say of actually

Most high frequency trading firms as well as just during periods of you know higher volatility there tends to be just more naturally more opportunities out there for folks to to see and utilize right and are you able to share you'll

Win right oh man I actually did not that's one thing I did not ask the lawyers about but in general I will say you know we were really happy with our performance this year we did have like two very particularly disappointing

Months in general but um overall you know we were definitely still by a huge margin beating you know all of our benchmarks so it's it's been a great time for us too and also because it's oh this is also a first recession we've

Been through – because we didn't go through oh wait I was still in school so so the other thing is because of this we were able to prove ourselves and be like ha you see we can finally show we've been through you know the wringer we've

Been through this bad these crazy times and that week you can show that we can you know go through those and more so you know we've had some disappointing months but in general it's it's been alright okay and how about average daily

Turnover of course that would vary with the number of trades but just broadly speaking oh man just turning over so much it can really add up to quite a lot I mean you know we would trade like recently again like the most was 71

Billion I think per day this year so we've gotten quite high and substantial in terms of the volume and turnover rates and it'll probably continue like that for hopefully for sometime right and

Given that you're trading so much and I don't know you win right but I presume it's it's you know above 50% are you profitable at the end of each day like I presumed losing day would be pretty rare so we are not necessarily

Profitable at the end of each day and I think that's something you know unfortunately for us and maybe even for other firms as well it's become a lot more common like I know we read articles about virtu you know not having a single

Losing day in years right and you're like wow is that is that standard for the industry I'll say probably not most hft firms you know we can end the day losing you know having a negative you know much less money than like the day

Before that does happen as well so you know the goal is though that we do aim for a hundred percent winning days nonetheless and you know when a losing day happens that's that's okay you know we try to deal with it usually what

Happens for on our end is like we might have twenty nine winning days and then one big losing day that just wipes out everything we made during those twenty nine days sometimes that kind of stuff you know does happen to all companies

All firms but um it's its life right and we've learned to how to deal with those types of days how to mitigate those risks and learn from those you know crazy experiences okay and now I haven't asked you what products you're actually

Trading like what markets are you most active in pretty much almost everything we're not as active in like we're active in primarily US markets right now but um you know US equities is probably the one big market that were not as active in at

This point in time we've we've traded before in the past but I'm at this point in time it's just not the market for her for us most of our strategies are concentrated in in futures and we have a lot of FX at

This point in time as well we do have some option strategies that we're utilizing at this point too so yeah a lot of I guess you know delta-1 strategies and not a lot of equities although it's something that

You know we've every year every month we consider like oh maybe we should go back into equities again and try it out it really just depends on the conditions and you know what what we're testing we're constantly testing things behind

The scenes and figuring out whether or not it's the right time to pursue something and is that all future's in the US or not all of them but is it or are you more more focused on like indices

Futures or commodity futures or pretty much the the whole mix it's the whole mix yeah we're pretty agnostic within the future space at this point in time they do have to be products that are you know relatively frequently traded or

Else it would you know not be worth it for us to deploy strategy within so they're they're heavily traded products that we can hopefully take up you know chunk of the adv not a lot maybe like you know aim for one to three percent of

The average daily volume of that product but um yeah so you know we do look for things like that but in general we try to be agnostic in terms of sector okay now this might be a bit of a broad question but just on your strategies in

Such a highly competitive area of the market that you compete in how do you seek to generate alpha in terms of generating alpha there's a lot of different things to kind of keep in mind on the parameter side so I think the

Biggest thing is figuring out what are some niche areas that we can really take advantage of what are some products that you know we we can kind of be able to map out and be able to make predictions on so a lot of its based on you know

Small predictions maybe a few seconds ahead of time on where different trades and products will go so so that's a big portion of it but then also looking at other I think of other things we we do on this side let's see just mapping out

In general you know hundreds of different signals and you're like finding the different signals that we can trade mapping those all out and then we actually do use a layer of machine learning as well to do some a signal

Selection throughout the day as well so figuring out you know what those signals are and what's gonna make the most money during this certain environment and then when the vibrant changes you know what what kind strategies we want to pull

Back and would want to deploy so that's always a really tough question but it's surprising because that most hft firms you know I would never say that we're an AI driven machine learning based firm but we do use just a

A bit to do some you know basic strategy selection type processing is on our end what data drives the majority of your trading signals is it audible cologne like looking at the market depth or is there additional input from other data

Sources – it's mainly market depth yeah so just order book data is extremely helpful for us we do have a little bit of what we call metadata you know you might call alternative data in different types of areas – whether that's usually

It's like you know things like calendar data right like every high-frequency trading firm needs to know when bank holidays are and things like that or maybe things like SCC filings or big economic you know announcements or you

Know Fed meetings things like that you know that kind of stuff does need to be factored into you know the strategy as well so we can't just be completely focused on the market data but just also keeping in mind all the other components

Of data as well so yeah in general though I think the biggest source is still you know just market data that's the bread and butter we need to kind of survive and then from there you know it's almost like a bonus if we can get

Other sources of alternative data that could help us out here and there as well okay can you speak to almost like this strategy survival rate or the signal survival rate like what's the edge decay like on these signals like are you

Continually you know innovating and revising your signals or do most of them kind of last for a long period of time I mean I perhaps didn't really word that question very well but I think you probably understand what I'm asking here

Yeah absolutely I mean the half-life of these strategies you know can be relatively short it could be a month one and a half months you know and then within three months the strategy is you know pretty much almost worthless at

That point in time and so it's a process of constantly revising the strategy use the process of constantly finding new strategies as well and modeling those strategies and different environments to see you know how they would do

And and then also figuring out why do the strategy die and they could die for so many different reasons you know a lot of it's because competition has caught up or maybe because the environments changed in the you know signals that

You're looking at just are no longer valid for whatever reason that happens too so there's there's all kinds of different reasons but um it's just important to be able to figure out when it happens and why and be able to move

On quickly and not get too attached to one or two you know strategies but you know we have like hundreds and hundreds of different types of signals to potentially deploy and take a look at so so I'm just making sure we're careful

There I know there's stories about like I don't know if you've heard those stories of like Renaissance for instance and other types of funds where you know when people leave their those types of firms that are doing really well one of

The biggest reasons why it's hard to spin out of certain you know successful hft firms and to just do your own thing and be more successful is a lot of it's because they're constantly revising their strategies over at Renaissance as

Well I'm sure and you know when you steal a strategy by the time you get to launch three months later that strategy is already dead and it's just no longer effective so that's definitely something that we do as well on our end is just

You know it's not on purpose but just the nature of high frequency trading it's just that the opportunities you find may not necessarily last for a very long period of time and they might be really niche so once we find it and once

That niche kind of goes away then we're on to you know hopefully the next thing during the trading day what level of human import or decision making is there obviously the trades are all automated but is there any human intervention

During the trading session oh it's a really interesting question well first I will say like a lot of people have this vision of – to trading being you know a bunch of robots taking over the world or you know machines and you know that are

Out of control with you know humans not being able to really understand what's going on in this black box and that's really not the reality at all you know the these trades are being heavily monitored and you know by humans

Throughout the day when we're trading and you know really just gathering all kinds of data about what we're doing and stuff like that there's a huge human you know obviously we're not having human sit there and click to make those

Trades you know there's no it's just being placed automatically but there's always a human behind the scenes to tweak the strategies or to tweak some parameters or you know change some of the risk parameters things like that and

A lot of it's in code but just kind of changing that coding of what the strategy might look like so so yeah there's a surprisingly large human element in terms of what we do and you know the strategies also there are only

As dumb as the humans who are you know creating them hopefully as smart but you know I would say that you're they're as dumb as the person who created it so absolutely you know we need that human intervention constantly to make sure

That nothing goes rogue or you know crazy the other thing is like I'm trying to think of you know the markets as well right because this isn't like a closed-loop game it's not like a game of go or I don't know chess where there's

Rules and everyone's going for the same goal you know the financial markets things are constantly changing to the point where even today right you know you see a lot of things like oil going negative you know things that

Economically it's like in your textbook you know you would never it just makes no sense you know and that happens all the time and a lot of that might be just it's just driven by human very human behaviors and so you need to have a

Human behind the scenes to understand those behaviors in order for your strategy to do well you can't just have a strategy that you know you can't you know even if it's a cut some kind of deep learning model or you know if it's

Learning by itself from its own inputs and stuff like that like that's that's great but again you can't quite model the financial markets and we've seen startup after startup you know including Watson as well you know IBM Watson

Trying to model the markets with their ETF and you look at how they're performing and you know something's just not the best start to the ETF ai driven you know stuff yeah absolutely my my personal opinion as a high-frequency you

Know ironically has a high affinity trader is that there is an important human element that um you know we can't forget and behind the scenes now as you are high-frequency trading fund obviously Spade and reducing agency as a

Big factor in what you do what measures do you take to reduce latency but internally and externally you know the biggest thing people always imagine is like oh you're doing you know what they didn't flash boys but we don't do any of

That you know biggest thing is on our end it's purely pretty much you know just internally making sure okay let's you know gather data the fast as we can let's make sure that once the data is in place you know between

Getting raw data and processing that and turning it into an actionable strategy that we can reduce the amount of time it takes to be able to do that and then also be able to run that strategy you know on live data and be able to you

Know place those traits as fast as you can and and you know for latency it's basically you know as soon as we discover a decision we want to make the time between that decision and actually implementing it you know that kind of

That time we do try and reduce as much as we can you know whether that's just through direct market access you know with the exchange or making sure that you know our order management system is structured in a way that is as fast and

Efficient and can process you know a bunch of orders at the same time without having you know too much lag I know we've had I'm trying to give we've had some I know on the development side like we'll maybe consider like you

Know even just things like shortening names of stuff might help a little bit too so I'm just making sure that our names are a little bit shorter so that computer wise you know it takes a little less time to process so stuff like that

Might might work as well and then making sure that we optimize for different strategies on the latency side you know the strategies themselves so making sure that we're optimizing on that front so that they can be you know code wise

Right to be a little bit faster and take a little less processing time in order to complete so so that's definitely important as well I will say one other thing those you know I know everyone obviously talks about speed and thinks

About speed when it comes to high frequency trading but um on our end I think we found our niche in terms of you know where we're not trying to aim for being like the top fastest high-frequency firm out there and I

Think most firms are like that these days where you know we don't play the hardware game actually so we used to we used to like you know assemble our own Hardware buy the best thing out there and then we realized

Year they're gonna release a better I don't know FPGA a better you know piece of hardware somewhere out there and do we really want to pay an extra million dollars is it really worth it you just realized it just wasn't at some point

And they were lower hanging fruit in different areas a big reason by the way it's also because at the exchange level if their lien C's are slow and there's a lot of jitter happening at the exchange level then even if we did place an order

Really fast you know the the decay there's still a little bit of a decay in terms of the timing it takes and the alpha that we end up getting and so we're like well why don't we just figure out you know we call lowering fruit in

Terms of maybe better signals or looking at better products out there that you know we could take more advantage of without having to compete at the you know microsecond level all the time so that's you know going for strategies

That are have longer holding periods as well and may not be you know what you would typically expect of high-frequency trading just trying to be fast fast fast now one of the things which I understand about hft strategies is that they have

Capacity constraints and that's what's kind of interesting about what you guys do is you are fund where most traders who are running hft strategies mostly prop model whereas your fund and I guess that's one of the reasons is because hft

Strategies typically have constraint capacities how do you deal with that aspect of it now it's a great question it's one of the biggest things that keeps me up at night actually is some you know hft is although people always

Assume it's this menacing thing you know we're not a scalable business the industry is just not scalable given the nature of what hft is you know and and that's that's fine I think that could be a fair thing and you know it's just how

Things are the capacity constraints you know the reason why we're a hedge fund to begin with was we just didn't have we didn't even have you know a million dollars or a hundred thousand dollars you know we were literally negative

Money because we weren't debt so we didn't have any money at the time to trade and we supposed well you know even if we could start a small fund in the mean time it would be better than nothing and at least give us enough

Abus'd in our early-stage careers to you know how have a to know the industry better and also to hopefully you know learn a thing or two so so that was the big thing there but um you know if we could like let's see

If we could start today yeah we would probably be a prop shop instead of a hedge fund structure because the hedge fund structure you know it's great in that yeah you can you have all these external investors coming in you know

And and that's good for us to get to know the industry better and to see what's happening out there and to meet some great investors the bad part is obviously it's just it's a lot of extra work you know to have to do the weekly

You know sending newsletters out or sending emails and you know answering questions and stuff but also a regulatory wise to you know registering with all the different uh you know SEC NFA whatever other countries and

Jurisdictions that we want to be involved with so that is you know just a lot of administrative work cost you time and money having a director and the fund you know thinking about the legal setup of the

Fund structure all those other kind of stuff like that costs surprisingly a lot of money even before getting to launch and so so yeah that was kind of a trade-off that we had to make I guess in order to start the hedge fund structure

Right now Christina just one last question before we close it out here is high-frequency trading as profitable as everyone understands it to be I'm not sure if this is public information or if you're able to share it but if if so are

You able to share some performance numbers yeah yeah absolutely so I would say it depends on who you are because some people imagine that hft is just this money making you know cranking machine that you know never loses and

Stuff like that and that's definitely not the case at all it's a very you know it's fallible it has issues and a lot that's the reason why a lot of hft firms you know that existed ten years ago are no longer around today right it's

Because they still died for one reason or another or they got acquired or they just were no longer the profitability that they used to be I think a good analogy by the way is like in the crypto space as well you know where if you

Started a crypto hedge fund I know back in 2012 or 13 yeah you were great you know you you're good Congrats on being one of the you know earlier funds right in that space but if you started to crypt of fun today you

Know most of the Alpha is gonna be pretty much gone you know so it's kind of a similar thing where it gets harder and harder to compete as more people flood the industry like that but I will say like it was really surprising for me

Because we had spent two and a half years literally making no money at all which is you know for any startup that's extremely frustrating and you know almost like heartbreaking to see and then the day that we started launching

Strategies and stuff you know we were super nervous I remember turning it on and literally was nothing like we weren't trading at all and just nothing was happening and we're just waiting around the table you know our in our

Office just huddled there being like what's going on you know why isn't anything happening and then we just kind of gave up might in a day just went to go grab I think I went to go you know eat some ramen or something and then

Someone called me over they like Christina no comeback is finding the starting to train I was like oh my gosh you know the markets right about to close and we literally started trading at the clothes and we checked our

Balance and we had made $70,000 you know in like a minute and we were like oh my gosh that's that's more money than my parents make combined in a year you know and that feeling of like oh my gosh the half the

Tables turned you know and hopefully we won't lose you know um hopefully we won't lose the money but uh you know that that kind of feeling of oh my gosh and then thankfully that you know continued for for some time and then and

You know we we felt very just like holy crap this the wait was worth it and what about like annualized returns that type of thing are you able to share that info unfortunately I can't cuz we've gotten in trouble for sharing it in the past

For like it's considered marketing and it just does not go well with we were audited quite extensively in the past couple of years or so and this was always one of the things that they always get us on is they'll listen to

The podcast and be like oh you know you were why did you say that and why did you why why did you have this tone when you said it you know and like this could be construed as you know marketing in one way or another so unfortunately

We're not allowed to mark I would love too though because I I'm so proud of you know what we've done unfortunately not the you know we're not the type of fun that can market that kind of information understood that makes sense I mean the

Last thing you want is is legal hassles so totally understand well Christina just such an amazing job it's really inspiring to see what you and your your co-founders have done and what you've achieved over the past I think it's

Eight or nine years now so you know especially having started from scratch like built everything from the ground up it's just incredible awesome talking with you I'm very grateful for your time and I know we've been trying to set this

Up for a little while so I'm glad we could make it happen if somebody wants to find out more about you I know you've given a few talks and that type of thing which are available on YouTube you're also on Twitter can you please share

Your Twitter handle sure it's just my name it's at Cristina Qi which is spelled Qi yeah please feel free to check it out I'm a little crazy on Twitter which a lot of people don't like but I think there needs to be more loud

Vocal voices in this industry so and hopefully hopefully I have some reason in terms of mostly stuff I say and I'm not just completely crazy like some other folks in this industry which is fine but you know everyone has a

Different personality so feel free to you know check that out and then also recently I've been working on a data venture as well so if you guys are interested in market data and infrastructure colocation that kind of

Stuff you know definitely feel free to reach out as well because we do some really cool stuff on that front too I did mean to ask you about that actually I mean if you have time do you want to just give a quick summary of what you're

Working on there sure sure basically you know one of the biggest challenges we faced when we were launching the hedge fund was we had spent over a million full amount of money over a million dollars on data and

Infrastructure and and we feel like that barrier to entry is way too high we want to lower that barrier for other folks out there and just so you know this has nothing to do with Dome yard we're completely unaffiliated you know that

We're not front running in a dome here it's not taking that data at all from anybody's that's a completely different venture but basically just offering you know data that's a lot it's the quality that

We expect from you know institution quality data but we sell it all apart so it's broken down you can buy literally just one security if you wanted to or you know the entire thing if you want to do but the goal is like we make it a lot

Cheaper for folks and lower that barrier to entry so it's called data bento if anyone's ever curious like bento as in the bento box and the goal is like you know people can purchase data from various different sources and it's we do

Live in historical so you can have live data feeds from various different sources but in smaller chunks than what we typically see in the market so yeah that's a it's been fun to work on that it for me it's meaningful because that

You know I've been spending people always ask me questions about morality like you know you're an hft right like how a moral can you be and I've always wanted to do something that hopefully would bring some value to the industry

And have some meaning and so for me data bento has been fantastic in terms of bringing that kind of value and I'm very lucky to you know be able to have you know to have the hedge fund reach a stage where I can be more hands-off now

And that we have a team and that you know the team is fantastic to kind of take over a lot of the day-to-day you know operations and things that I didn't want to do as much so so that that's been very lucky and then you know to be

Able to start this data venture and hopefully help people out that's been a dream come true for me yeah that's cool that's really cool I'll make sure to include a link to that in the show notes and if someone wants to find out more

About dome yard what's the website it's just do me a comm if people know we were named after the MIT dome in Harvard Yard back then we didn't want to name the fund after ourselves and then by coincidence the comm was you know pretty

Much free so we were able to snag that and quit pretty quickly and got pretty lucky with the name nice nice well Christina very grateful for your time really appreciate your doing this thank you very much and we'll talk again soon

Thank You Aaron it's been an honor you've reached the end of this episode of chat with traders but rest assured there are more episodes loaded with real market insight and zero hype on the way soon so to stay updated with each great

New release subscribe to the podcast and iTunes and we'd love it if you'd leave a rating and review we'll catch you next time on chat with traders you

Related Videos

Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield hi folks what's good welcome to episode 187 of chat with traders so ...
Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield traders earn five field here welcome to one ninety this is an episode to...
Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield traders welcome to the show this is episode 192 similar to the last epis...
Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield what's up and welcome ladies and gents this is episode 195 and I&#39...
Markets speculation and risk this is the chat with traders podcast hosted by Erin Fifield hi team welcome to episode 196 this is the third in the series of my f...
Markets speculation and risk this is the chaplet traders podcast hosted by Erin Fifield boys and girls what is up welcome to episode 198 on this episode is a sp...
Markets speculation and risk this is the chat with traders podcast hosted by aaron field hi traders and welcome to episode 190 this here is another compilation ...
down below the wards the comings and goings are constant there are meals meds and masks the quantities swelling as the NHS is mobilized the phone's been ri...
So your first choice James is the sort of collection of surgeons in the hospital it's a Barbara Hepworth piece owned actually by the Royal College of Surgeo...
Usually a recession closes in slowly confidence ebbs away investment dries up jobs are lost this is different economic activity has simply stopped vanished unli...
when we heard the schools were closing that is the first time in my life that I went home and I actually was sick some families are really struggling to manage...
Sadly of those in hospital 12100 and seven people have now died another day another politician standing at this lectern with bad news to tell but today away fro...
Entering this phase in your life is difficult enough isn't it contemplating your death and saying goodbyes and making preparations but doing that in a plann...
the questions the barbs the beads of sweat are worried ministerial Brad all gone cease to be all silence in darkest hours past the familiar sounds of the green...
day 24 of lockdown giving away today 25 will live done at least 45 days of total lockdown before this is over the government confirmed today it was inevitable ...
It's about the houses nobody's listening to me nobody's taking my voice nobody's literally caring yeah that's how I feel Alice was a carer w...
what's always remarkable about America is its scale this is a place used to being the biggest the strongest the richest right now though what startling is ...
how covert nineteen impacts on the NHS is a key determinant of how we respond next to the outbreak there are various figures from different parts of the NHS th...
dr. Unwin is guessing his bi-weekly covert test we're testing today you can either be tested on a Monday or a Tuesday morning the lab arranged to collect t...
When health workers started dying from covert 19 they seem to be disproportionately from black Asian and minority ethnic backgrounds now politicians have starte...