HFT drives the future of finance
After struggling for the last 30 days or so, I’ve come to the conclusion that designing trading software is no different than architecting it for HFT. In fact, I have come to realize that the fewer dependences you have, the faster your system should be if implemented right.
From a market perspective, most US trading volume is driven by high-frequency trading. This is the future which cannot be denied. It is growing more and more as time passes. The problem is that regulators are not doing their job to protect the little retail trader. In actual fact, it is automation that is killing off this type of trader. As more more software becomes automated, I’m finding it dumber and dumber as better architected systems will reign supreme. It’s like the next generation of retail trader versus a well established institutional or professional trader. What am I saying? I think your coding skills will become the way to differentiate your performance as a trader in the future. The only difference is how you apply your analysis to it. Do you approach your trading as a trader, computer scientist, or mathematician?
Here is an interesting video debating how high-frequency trading effects everyone else
As I mentioned above, it is quite clear that your multithreading/concurrency/
Is Parallel Programming Hard? What to do about it?
This link includes some advanced resource tips and advise
C++ Concurrency book learning without TBB?
Note that the above link includes useful code snippets for developing high speed trading systems
I’ll post more as I get to it.
Thanks for reading,
Bryan