As I stated in an earlier post in this thread I see science overachievers (Phds) playing a role in HFT limited to defining super efficient hardware and software infrastructures (another thing that comes to mind is the interest of the financial world when it seemed neutrinos travelled marginally faster than light speed). I don't see the trading algos logic requiring higher mathematics or anything beyond graduate level science skills. I am a computer science undergrad myself. I developed some MATLAB software for my own use in the past (exploring the relation between different factors in the trading games with nice 3d surfaces). I appreciate how computer skills would benefit certain aspects of trading and what could be done that is useful and well beyond anything commercial software does, but graduate level would be sufficient. (someday I hope to employ somebody to help me put my ideas in practice. I estimate a couple of hundred thou to do things properly) Either I am wrong about trading or we differ in our view of what value a Phd can add. ras72
Basically it is big players AI having competition with each other in light speed. It is all about who is in another level first to rule the market for some period of time. Sometimes few big players would join together to rule it all but they are not ally forever. Even programmer for HFT can earn US million per year, and there is a reason for that because it is just that complicated so they cannot just find a cheap programmer from India or China to do the job. Simplifying saying there is just no way to be "just simple" when we are talking about more than half of the trading volume is done by HFT in USA and the algorithm is about how to get in and out in nano second and have profit, while there are like nine other competitors doing the same thing and the algorithm needs to beat at least half of them. If we are talking about longer time frame like holding for months, then the strategy can be simple or extremely complicated, but not the HFT case.
Even bringing into the equation the increased complexity of AI, the utility of which I somewhat doubt (but then I don't know enough about HFT), my take is that in between superstar Phds and a âcheap programmer from China or Indiaâ there is a realm of good, smart programmers who could get the job done; but the viewpoint distinction is becoming a punctilious one. Also, I hold the perspective that the value added is in the concept or observation whereas you assign crucial importance to the algorithmic implementation itself. It may be so in HFT. I doubt it is so in anything even only slightly longer frame (from a few seconds upward). ras72
-Engineering -Mathematics/Statistics -Physics -Operations Research -Computer Science (good idea to combine this with one of the above) Throw in some kind of trading experience, competitive sports, prior risk taking, etc. and you should do alright. It varies within the industry as well, high end prop shops hire differently than i-banks.
it is true that trader job is one of the most difficult jobs to get, especially with pretty bad econ and gov has more restriction on what banks can do with their money.