5 ms·

Why are you ex algo trader? It still seems like a growing market, so it's interesting that you stopped doing it. I'm right now working on optimizing my long te

Why are you ex algo trader? It still seems like a growing market, so it's interesting that you stopped doing it. I'm right now working on optimizing my long term investment strategy, but the hardest thing is that there's nobody to talk to it about it (unlike when I have other coding problems), and I have no idea when I'm doing something totally stupid in my code base. Getting the right metrics to optimize is really hard, as mostly I'm just looking at numbers and graphs.

osrec 8y ago
It's a crowded market, but not necessarily a lucrative one. I worked for a large investment bank, was paid well, but was also rather unhappy. I'm not a fan of office politics and also hated the fact that what we were doing was actually quite unsophisticated, yet we dressed it up like the most complicated thing ever to clients. We were using wavelets for analysing market micro-structure and using any generated signals to influence our trades. Wavelets were the topic of my thesis, so I was a good fit for the team. I know it sounds complicated, but if you looked at the signals we generated, they looked awfully similar to a rather basic trend-following style regression model. I believe we found a very complicated way to compute signals that a simple trend follower would compute in far fewer CPU cycles. Anyway, I got a fairly decent bonus one year (because markets were trending nicely), and decided I want to set up my own tech company. I was always a nerdy guy, and I thought I could contribute more in the tech space. With hindsight it was a good move, and financially I believe I am now better off than most of my trading colleagues. My advice to new guys joining our desk (maybe it helps you a bit): focus on your exit, more than your entry, and if you can help it, be really, really, really lucky :P
- _cs2017_ 8y ago
  > what we were doing was actually quite unsophisticated, yet we dressed it up like the most complicated thing ever to clients. Doesn't this describe nearly all of the financial industry? Do you happen to know of any exceptions?
  - osrec 8y ago
    Yep, and nope.
- dbancajas 8y ago
  how long did you work in that space and how much income you got over that span?
inference 8y ago
I bet he/she.they either 1. made fuck-you-money and retired 2. burnt-out from the stress and decided to make slightly less large piles of money in return for less stress.
- osrec 8y ago
  Somewhere between 1 and 2, but very good guesses :)
rjf72 8y ago
Perhaps the most important paragraph in the article: "Everyone is competing against everyone else. If one firm succeeds in making the market more efficient through quantitative techniques, then there’s less money left over for other people to make exceptional investment returns. There will be one or two firms that are good at innovation and recognizing things that other people haven’t recognized. But everyone else will be fighting over scraps." The market is, in theory, not a zero sum game. Algorithmic systems, in practice, are. In theory the only way you can genuinely beat the market is when you find an inefficiency in it. And inefficiencies are generally based on a lack of knowledge. If everybody knows what you know, that inefficiency will no longer exist. And now pair this with the reality that most (and it's probably safe to say all) algorithmic systems are not really especially complex. That doesn't mean they're easy to make though. The 'secret sauce' is in exactly what data is used, how the pieces you have interoperate, the way inputs and outputs are massaged, and so on. But this drives a huge need for secrecy. Nobody who is successful is going to want to talk about how their system is designed largely because most would not be especially difficult to replicate if they provided more than very abstract details.
- awongh 8y ago
  Is this always true? One thing I wonder is, given that the amount of money needed to move the needle in a market could be actually quite large, how much money woyuld you need to make off a given inefficiency before it becomes a zero-sum game? Is it 10s of thousands? Hundreds of thusands? Millions? And given some metric like trading volume how much does that affect that number? I've heard of some hedge funds raising way less than they could have, maybe because their strategies would be killed much faster with 500mm than with 50mm ?
  - blackflame7000 8y ago
    It's not a zero-sum game because debt is limitless. Imagine an extreme example where you rigged the game and literally won all the money in the world and then kept winning, what would happen next? People would find another form of payment via services and the money is worthless if you are the only one who has it. Therefore you might lend them money in exchange for a service so that they will keep playing your rigged game now accruing a credit of services in addition to money.
benj111 8y ago
I'm trying to parse your comment. It sounds like you're basically trying to do part time what dozens of large companies spend large amounts of time and money trying to do, and then expecting them to discuss with you what they've done? You know those books that show you how to get rich doing X? Those writers make far more money from the book than they do from X. No doubt you can find corners of the investing internet that will discuss these things, but you'll probably find an awful lot of chaff in your wheat, because the successful ones won't be chatting about it on forums. Now on a more constructive note, optimising your long term investment strategy is probably best achieved by getting a pension with some cheap trackers, and a savings account. If you're already doing that then great because that's 95% of your optimisations there. I wont recommend any web sites, because the details will depend on the country where you live.

osrec 8y ago

It's a crowded market, but not necessarily a lucrative one. I worked for a large investment bank, was paid well, but was also rather unhappy. I'm not a fan of office politics and also hated the fact that what we were doing was actually quite unsophisticated, yet we dressed it up like the most complicated thing ever to clients. We were using wavelets for analysing market micro-structure and using any generated signals to influence our trades. Wavelets were the topic of my thesis, so I was a good fit for the team. I know it sounds complicated, but if you looked at the signals we generated, they looked awfully similar to a rather basic trend-following style regression model. I believe we found a very complicated way to compute signals that a simple trend follower would compute in far fewer CPU cycles. Anyway, I got a fairly decent bonus one year (because markets were trending nicely), and decided I want to set up my own tech company. I was always a nerdy guy, and I thought I could contribute more in the tech space. With hindsight it was a good move, and financially I believe I am now better off than most of my trading colleagues. My advice to new guys joining our desk (maybe it helps you a bit): focus on your exit, more than your entry, and if you can help it, be really, really, really lucky :P

_cs2017_ 8y ago
> what we were doing was actually quite unsophisticated, yet we dressed it up like the most complicated thing ever to clients. Doesn't this describe nearly all of the financial industry? Do you happen to know of any exceptions?
- osrec 8y ago
  Yep, and nope.
dbancajas 8y ago
how long did you work in that space and how much income you got over that span?

inference 8y ago

I bet he/she.they either 1. made fuck-you-money and retired 2. burnt-out from the stress and decided to make slightly less large piles of money in return for less stress.

osrec 8y ago
Somewhere between 1 and 2, but very good guesses :)

rjf72 8y ago

Perhaps the most important paragraph in the article: "Everyone is competing against everyone else. If one firm succeeds in making the market more efficient through quantitative techniques, then there’s less money left over for other people to make exceptional investment returns. There will be one or two firms that are good at innovation and recognizing things that other people haven’t recognized. But everyone else will be fighting over scraps." The market is, in theory, not a zero sum game. Algorithmic systems, in practice, are. In theory the only way you can genuinely beat the market is when you find an inefficiency in it. And inefficiencies are generally based on a lack of knowledge. If everybody knows what you know, that inefficiency will no longer exist. And now pair this with the reality that most (and it's probably safe to say all) algorithmic systems are not really especially complex. That doesn't mean they're easy to make though. The 'secret sauce' is in exactly what data is used, how the pieces you have interoperate, the way inputs and outputs are massaged, and so on. But this drives a huge need for secrecy. Nobody who is successful is going to want to talk about how their system is designed largely because most would not be especially difficult to replicate if they provided more than very abstract details.

awongh 8y ago
Is this always true? One thing I wonder is, given that the amount of money needed to move the needle in a market could be actually quite large, how much money woyuld you need to make off a given inefficiency before it becomes a zero-sum game? Is it 10s of thousands? Hundreds of thusands? Millions? And given some metric like trading volume how much does that affect that number? I've heard of some hedge funds raising way less than they could have, maybe because their strategies would be killed much faster with 500mm than with 50mm ?
- blackflame7000 8y ago
  It's not a zero-sum game because debt is limitless. Imagine an extreme example where you rigged the game and literally won all the money in the world and then kept winning, what would happen next? People would find another form of payment via services and the money is worthless if you are the only one who has it. Therefore you might lend them money in exchange for a service so that they will keep playing your rigged game now accruing a credit of services in addition to money.

benj111 8y ago

I'm trying to parse your comment. It sounds like you're basically trying to do part time what dozens of large companies spend large amounts of time and money trying to do, and then expecting them to discuss with you what they've done? You know those books that show you how to get rich doing X? Those writers make far more money from the book than they do from X. No doubt you can find corners of the investing internet that will discuss these things, but you'll probably find an awful lot of chaff in your wheat, because the successful ones won't be chatting about it on forums. Now on a more constructive note, optimising your long term investment strategy is probably best achieved by getting a pension with some cheap trackers, and a savings account. If you're already doing that then great because that's 95% of your optimisations there. I wont recommend any web sites, because the details will depend on the country where you live.