Quote from amazingIndustry:
you are a total dreamer!!! Coding up a trading system on couple Saturdays??? Are you a joker? Why do you think investment banks, high frequency trading houses and hedge funds hire hordes of programmers and pay them very handsomely? Because it takes a few days to come up with a system to manage risk properly and have all modules communicate efficiently with each other?
The hordes of programmers are the problem, not the solution.
I coded up my own systematic architecture and it took me alone 3 months to implement a messaging bus that can communicate in and out of process between different modules. I tested a lot of different open source solutions, coded parts up on my own, run extensive tests and ended up shipping 6.5 million price ticks per seconds over the bus on a 4 core machine (involving task async and task distributed work in .Net). So, don't suggest you are that much better that you come up with a full fledged systems in a few days of work, hilarious!!!
Every time you post, you sound more like a back office IT guy than a trader.
There's nothing wrong with being that, but, perhaps you just haven't seen strategies in markets/products that are still wide enough to exploit from a trader's desktop. No need to handle millions-of-ticks per whatever, and no tech department involvement, thankfully.
200 lines of code, you seriously start to piss me off, LOL. You are talking with your head in your ass, to say the least.
Meh, whatever mate. I built a bigger algo for an i-bank while dealing with silly internal IT "processes", then built it again for myself, and, yes, only took about 200 lines of code for the trading strategy. The IB API handles much of the other grunt work.
If you're into grunt IT work, exchange connectors, or fix message engines, etc, then yes, you or some poor bastard in Bangalore will have to write thousands of lines of cryptic exchange specific slop.
But that's just infrastructure.
It doesn't make the P/L.
No upside in getting involved in that sort of thing, unless "like" it, and enjoy working the India shift.
Newbies, be warned coding up a full fledged systematic trading platform takes months at the very least and is definitely not recommended for a programming newbie. Even Tradelink does not implement everything, it would probably take several weeks - 2 or 3 months to get everything implemented to operate smoothly with IB and to code up custom pieces.
It's not easy, and there is no real shortcut through open source platforms.
But, if you know where you're going, it can come down to a small amount of good code to build a strategy.
Back to my main point - Software architecture does not solve the "profitability" problem in any way.