13 ms·

Math Whiz Trades Without Humans

99 points by monsieurpng 7y ago

badfrog 7y ago
What's the actual story here? This can't be the first trading firm that doesn't have human traders?
- abbadadda 7y ago
  I agree there is not much of a story here. On their website they list that they're recruiting for "Operations". This means there are humans operating the trading strategies, even if they don't call them "traders." Most firms have "operational traders" and calling them something other than traders is a little bit disingenuous. I have email correspondence with XTX regarding a "Trading Operations" job posting, so to write an article about how there are no humans involved in the process is not only not really a story, but also probably inaccurate. With all that said, they've done well so kudos to them. But is it worth a story? Probably not.
- EsssM7QVMehFPAs 7y ago
  Look at their revenue vs net profit chart. I bet scaling this kind of investment model that reliably is something out of the ordinary. They print money.
  - qeternity 7y ago
    Look at some other hft shops like Virtu for instance, this is fairly common. What is difficult is scaling the activity. These guys take on very little risk, which makes it difficult to deploy lots of capital. The guys who manage to crack that balance are the more money than God types like RenTech.
- dcolkitt 7y ago
  The real story is that XTX is primarily a currency trader. Unlike stocks, futures or options, currencies aren't tied to a single exchange. If you want to go out and change your dollars to yen, there's not a single venue or even federated system for doing that. If you go out and build an amazing strategy for, say, index futures, all you have to do is set up an account at the futures exchange. If you make a bunch of money the exchange doesn't care. The exchange isn't your counterparty, so your profits don't come at their expense. They're just happy to execute the volume. Whereas with currency, most trades are done through over-the-counter "liquidity providers" (LP), traditionally major international banks. If you make too much money, the LPs get pissed off. Since they're acting as your direct counterparty, your profit comes at their expense. Consequently, a great trading system alone isn't worth that much in currency space. If you make too much money, you'll just get banned from all the LPs. Like many others who tried to break into the space, XTX already had great HFT trading models. But their real innovation become an LP themselves. Rather than beg the major banking franchises to give them a seat at their table, they went out and built their own. That required creating business relationships with upstream sources of order flow, like retail forex brokers. That customer base gave them the capability to capture market share without worrying about counterparty risk. Combined with their superior HFT models, it gives them a major edge over traditional banks and LPs. They can offer tighter spreads and still make a profit at a price point that their competitors can't.
  - avn2109 7y ago
    This is much better than the article and should be the top comment.
  - lonelappde 7y ago
    XTX did all that without humans?
b0b10101 7y ago
> XTX -- the name refers to a mathematical formula used in its trading algorithms Wow so they do linear regressions?? Incredible. But in all seriousness, for how long are we going to dress up statistics as some new cool and exciting methods? Are VC's and bosses really this easily fooled?
- deepnotderp 7y ago
  Some (not all!) of the techniques at places like RenTec are fairly sophisticated. Hence their returns ;)
  - BubRoss 7y ago
    Like what?
    - larnmar 7y ago
      Nobody really knows, except for a few people who aren’t telling. There’s some clues in the mathematical backgrounds of the kind of people they hire though. Fifteen years ago they were snapping up experts in stochastic calculus. Nowadays, I’m not sure whom they’re hiring.
      - throwawaymath 7y ago
        They're not exactly forthcoming, but it's not the hyperbolic mystery the mythos would have you believe. You can piece together what matters to them - in the abstract - by searching MathSciNet for papers published by their researchers prior to joining. For example, it's pretty easy to see they hired an information theorist away from Princeton in 2017. But obviously that's not enough to reproduce everything they do, or else they wouldn't still be legendary. As it happens a lot of what makes them successful is not the sophistication of their trading algorithms and research, but also the sophistication of the execution and reliability. As an aside, from friends there Renaissance hires researchers in three primary ways: 1. They have a small network of professors who they solicit for promising new PhDs willing to leave academia each year. 2. They watch professors and postdocs in specific specializations, and reach out to those whose research meaningfully interacts with a thesis they're interested in internally. 3. They send small groups to conferences to poach people working elsewhere in industry (particularly tech) whose work is applicable to their own. They also do hire people who directly apply of course, but most hires are reactive. They especially like to hire people whose work or research looks like it might begin to encroach on their own, or is just notable and impressive. The math is certainly important to them, but that's just one dimension of it.
tobmlt 7y ago
X transpose X inverse, X Transpose Y, it's the best linear unbiased estimator (B.L.U.E. eh hem), and let me tell you whyyy, Because ... ah crap it's been to long, did anybody else learn this song in grad school for econometrics (or similar) when studying Gauss Markov theory? Here's to Professor Jackson at Auburn for this memory I can never quite remember the second line of, and yet I can never quite forget the whole thing...
hikefast32 7y ago
‘Member long term capital management? I ‘member.
XTX_UK 7y ago
https://twitter.com/AlexanderGerko/status/1203765678379012096 https://twitter.com/AlexanderGerko/status/120376567837901209...
agerko 7y ago
What goes for journalism these days: clickbait headline, "research" based on out of context 4 year old interviews + expert opinion of people who similarly have zero understanding of the business, literally not get a single thing about the company right, but, hey, clicks!

99 points by monsieurpng 7y ago

badfrog 7y ago

What's the actual story here? This can't be the first trading firm that doesn't have human traders?

abbadadda 7y ago
I agree there is not much of a story here. On their website they list that they're recruiting for "Operations". This means there are humans operating the trading strategies, even if they don't call them "traders." Most firms have "operational traders" and calling them something other than traders is a little bit disingenuous. I have email correspondence with XTX regarding a "Trading Operations" job posting, so to write an article about how there are no humans involved in the process is not only not really a story, but also probably inaccurate. With all that said, they've done well so kudos to them. But is it worth a story? Probably not.
EsssM7QVMehFPAs 7y ago
Look at their revenue vs net profit chart. I bet scaling this kind of investment model that reliably is something out of the ordinary. They print money.
- qeternity 7y ago
  Look at some other hft shops like Virtu for instance, this is fairly common. What is difficult is scaling the activity. These guys take on very little risk, which makes it difficult to deploy lots of capital. The guys who manage to crack that balance are the more money than God types like RenTech.
dcolkitt 7y ago
The real story is that XTX is primarily a currency trader. Unlike stocks, futures or options, currencies aren't tied to a single exchange. If you want to go out and change your dollars to yen, there's not a single venue or even federated system for doing that. If you go out and build an amazing strategy for, say, index futures, all you have to do is set up an account at the futures exchange. If you make a bunch of money the exchange doesn't care. The exchange isn't your counterparty, so your profits don't come at their expense. They're just happy to execute the volume. Whereas with currency, most trades are done through over-the-counter "liquidity providers" (LP), traditionally major international banks. If you make too much money, the LPs get pissed off. Since they're acting as your direct counterparty, your profit comes at their expense. Consequently, a great trading system alone isn't worth that much in currency space. If you make too much money, you'll just get banned from all the LPs. Like many others who tried to break into the space, XTX already had great HFT trading models. But their real innovation become an LP themselves. Rather than beg the major banking franchises to give them a seat at their table, they went out and built their own. That required creating business relationships with upstream sources of order flow, like retail forex brokers. That customer base gave them the capability to capture market share without worrying about counterparty risk. Combined with their superior HFT models, it gives them a major edge over traditional banks and LPs. They can offer tighter spreads and still make a profit at a price point that their competitors can't.
- avn2109 7y ago
  This is much better than the article and should be the top comment.
- lonelappde 7y ago
  XTX did all that without humans?

b0b10101 7y ago

> XTX -- the name refers to a mathematical formula used in its trading algorithms Wow so they do linear regressions?? Incredible. But in all seriousness, for how long are we going to dress up statistics as some new cool and exciting methods? Are VC's and bosses really this easily fooled?

deepnotderp 7y ago
Some (not all!) of the techniques at places like RenTec are fairly sophisticated. Hence their returns ;)
- BubRoss 7y ago
  Like what?
  - larnmar 7y ago
    Nobody really knows, except for a few people who aren’t telling. There’s some clues in the mathematical backgrounds of the kind of people they hire though. Fifteen years ago they were snapping up experts in stochastic calculus. Nowadays, I’m not sure whom they’re hiring.
    - throwawaymath 7y ago
      They're not exactly forthcoming, but it's not the hyperbolic mystery the mythos would have you believe. You can piece together what matters to them - in the abstract - by searching MathSciNet for papers published by their researchers prior to joining. For example, it's pretty easy to see they hired an information theorist away from Princeton in 2017. But obviously that's not enough to reproduce everything they do, or else they wouldn't still be legendary. As it happens a lot of what makes them successful is not the sophistication of their trading algorithms and research, but also the sophistication of the execution and reliability. As an aside, from friends there Renaissance hires researchers in three primary ways: 1. They have a small network of professors who they solicit for promising new PhDs willing to leave academia each year. 2. They watch professors and postdocs in specific specializations, and reach out to those whose research meaningfully interacts with a thesis they're interested in internally. 3. They send small groups to conferences to poach people working elsewhere in industry (particularly tech) whose work is applicable to their own. They also do hire people who directly apply of course, but most hires are reactive. They especially like to hire people whose work or research looks like it might begin to encroach on their own, or is just notable and impressive. The math is certainly important to them, but that's just one dimension of it.

tobmlt 7y ago

X transpose X inverse, X Transpose Y, it's the best linear unbiased estimator (B.L.U.E. eh hem), and let me tell you whyyy, Because ... ah crap it's been to long, did anybody else learn this song in grad school for econometrics (or similar) when studying Gauss Markov theory? Here's to Professor Jackson at Auburn for this memory I can never quite remember the second line of, and yet I can never quite forget the whole thing...

hikefast32 7y ago

‘Member long term capital management? I ‘member.

XTX_UK 7y ago

https://twitter.com/AlexanderGerko/status/1203765678379012096 https://twitter.com/AlexanderGerko/status/120376567837901209...

agerko 7y ago

What goes for journalism these days: clickbait headline, "research" based on out of context 4 year old interviews + expert opinion of people who similarly have zero understanding of the business, literally not get a single thing about the company right, but, hey, clicks!