ATS programming is EXTREMELY platform-dependent. A programmer with prior experience on the exact platform would be the most desirable, but an experienced ATS programmer on any platform should know the pitfalls to be wary of. In any event, to obtain the niche skills needed.... bring money.
Find a good programmer and make him a partner, salary + % of profits. If the ATS is profitable this arrangement will work fine 90% of the time and your risk will be limited. But given that 80% of ATS fail, you should not worry about the risks involved with this type of arrangement. Otherwise, if you want to get a bit more sophisticated it will cost you more. Design your ATS software so it uses formulas you can input from a formula editor and never let the programmer see the ones you will use. Test the program with the "fake" formulas and then when you are ready input the actual ones yourself. Ron
A neighbour has recommended someone he employs on a consultancy basis at work.(national company) the chap will be calling me so we'll see what he sounds like.(I was hoping the neighbour could help (nice chap) but his programming isn't up to it,unfortunately.
Anyone with less then 10yr of systems programming experience is waste of time and they cost from $1k per day upwards . Bare in mind, they work weekends as well. So a full blown ATS designed from scratch, say in C++ on NT, with some serious back office features will make you short ¼ Mil USD easily. Good luck is your trading
I am software developer / consultant and have been for more years than I care to admit and for a 1/4 mil you can find someone to not only write your ATS but to also do odd jobs around your home. But seriously, depending on the complexity of your strategy, the coding could take any where from a few hours (TradeStation, Esignal) or several days to several months (VB, C#, C++). Some of the best strategies I have ever coded have been written using the esignal scripting language in a few hours.
I've been debating doing a little freelance programming work to help other retail traders automate their trading system inorder to increase the number of trades / profit for them. I figured a little extra money would help increase the base capital for my own automated testing. I already have the API work for IB tested and coded into my own c++ objects, which would cut down implementation time for IB users. I have a lot of high performance graphics opengl+gpu computation background so I am able to make extremely fast / efficient graphing UI so that I am control my automated system when needed. I figured I'd post on this thread just incase some retail IB traders are interested in looking into automation after reading the discussion. Feel free to contact me privately through ET messages.
freelancing for small traders or any small business is nightmare. they never know what they want and are never willing to pay what it costs. They are more hassle than anything. creating a some standalone software that people can download is a better bet.
What I consider as important points in ATS development: 1. Don't try and learn programming yourself - unless it is for basic tasks which can be implemented in VBA and the like. If your plumbing needed attention and it was a complex task most people would call a plumber. 2. In order for a programmer to get an ATS to do what you want - he needs to understand it. Most programmers dont have access to the markets, necessary funds, or the inclination to trade. Find an intelligent, recommended person you can trust to develop your system - and explain it to them well. 3. Be prepared to give constant input to a developer - and continually check that the developer understands what you need. If you are not willing to do this, you may end up with what the developer thinks is a fantastic system - but is completely useless for trading. 4. Never insist that all code is written in C++ or a specific language. Hire a developer with a good enough skills base to decide on the best programming language for the task at hand. A fast C++ system may take a year to develop, whereas a C#/.NET implementation may take 3 months. 5. If you a getting a contractor/contracting company to do the work for you, be sure to check if the contract includes rights to source-code, or if you need to purchase source code at an additional premium. 6. Always prefer off-the-shelf solutions than developing your own system. However, always check that the solution works and can be extended to meet all future requirements before purchasing. 7. Speed costs - above anything else, writing fast, scalable, multithreaded software is difficult, costly and should only be tackled by people who have the resources to tackle the problems that will occur.