3 ms·

Moved from development to quant trading. Still involves writing code but that's not the focus, and the code is for yourself, not somebody else. Much more satisf

Moved from development to quant trading. Still involves writing code but that's not the focus, and the code is for yourself, not somebody else. Much more satisfying having such a direct feedback loop between your code/research and its impact on the markets in production. Also feels more intellectually stimulating in that many of the problems you solve don't have an existing solution (or at least not one in the public domain). And very satisfying if you're a competitive person, because your every success is in a sense the failure of somebody in another trading firm.

nouveaux 5y ago
One of the concerns about trading that I always have is being able to compete with the big traders. What are your trading speeds like? How long have you been doing this?
- logicchains 5y ago
  You don't have to do it all yourself; there are quite a few companies that will offer you trading capital and access to their infra in exchange for a 50%+ share of your profit. I myself am on a team at a smaller firm with a high degree of independence, around five years of experience in the industry. The infra (including FPGA/ASIC execution platform, microwave network etc.) is the big expense for latency. On the software side it's often easier for a single experienced C or C++ dev to write trading software with a hot loop of 2-3 micros than it is for a large team anyway, because it's harder to ensure memory allocation and layout coordination/discipline with multiple devs. The ideal however isn't to compete for those ultra low latency opportunities, it's to find a trade that nobody faster than you is doing. E.g. it's too capacity constrained or cumbersome for the large firms, or in a less developed market that they aren't yet in (e.g. somewhere like Nigeria or Indonesia, or some random blockchain), or it's so creative that nobody else has thought of it.
_benj 5y ago
I’ve been considering quant for a while too. What you mention about latency sounds like HFT, market making strategies? Instead of that I’ve been looking into momentum and mean reverting stuff, and my best friend being a profitable trader gives me the confidence (hopefully not misplaced) that we can automate him away and scale the strategy that he has already developed, or at the very least setup a system that receives weekly input and can operate by itself with low maintenance requirements. I’m curious what was your journey into quant from software development… did you studied statistics and data analysis? Saved to have a ton of capital before starting? Open a quantconnect account and profit? :) Thanks and congrats!

nouveaux 5y ago

One of the concerns about trading that I always have is being able to compete with the big traders. What are your trading speeds like? How long have you been doing this?

logicchains 5y ago
You don't have to do it all yourself; there are quite a few companies that will offer you trading capital and access to their infra in exchange for a 50%+ share of your profit. I myself am on a team at a smaller firm with a high degree of independence, around five years of experience in the industry. The infra (including FPGA/ASIC execution platform, microwave network etc.) is the big expense for latency. On the software side it's often easier for a single experienced C or C++ dev to write trading software with a hot loop of 2-3 micros than it is for a large team anyway, because it's harder to ensure memory allocation and layout coordination/discipline with multiple devs. The ideal however isn't to compete for those ultra low latency opportunities, it's to find a trade that nobody faster than you is doing. E.g. it's too capacity constrained or cumbersome for the large firms, or in a less developed market that they aren't yet in (e.g. somewhere like Nigeria or Indonesia, or some random blockchain), or it's so creative that nobody else has thought of it.

_benj 5y ago

I’ve been considering quant for a while too. What you mention about latency sounds like HFT, market making strategies? Instead of that I’ve been looking into momentum and mean reverting stuff, and my best friend being a profitable trader gives me the confidence (hopefully not misplaced) that we can automate him away and scale the strategy that he has already developed, or at the very least setup a system that receives weekly input and can operate by itself with low maintenance requirements. I’m curious what was your journey into quant from software development… did you studied statistics and data analysis? Saved to have a ton of capital before starting? Open a quantconnect account and profit? :) Thanks and congrats!