Hi Guys! I'm a software programmer with lots of experience developing automated trading solutions for the Interactive Brokers platform. I'm looking to monetize my experience and software development skill by building custom automated trading platforms for other day traders with "systems". Does anyone have suggestions as to how I can approach this market - I really don't have much of any idea. thanks in advance, Will email: website@dudziak.com
Will, I have a question... I'm thinking of "copying" IB's TWS platform. Well, at least some minor parts of it so that I could use it as an order entry platform for a different broker. How hard a task might that be and what might I need to spend to accomplish this? Thanks!
I'd expect to charge on a per-project basis. I'm very familiar with the "discovery" or 'quoting' phase of software projects - it was an integral role of my last job. I am am usually able to assess fairly well the ammt. of effort a given development task will take before writing the first line of code, thus I'd prefer (and most clients do as well) to go the route of project-based pricing.
Risktaker, For the remainder of this post, "copy" is meant to mean "build with inspiration from". with that understanding, I will continue this post... My major familiarity with IB is via the APIs they provide (application programming interfaces). The APIs allow a 3rd party program (mine) to interact with TWS, gathering real-time data, making decisions, and executing real-time trades automatically. If you are interested in copying the actual TWS interface (the part that a casual user interacts with) to use with another broker, you have a whole host of software issues. primary among them are: #1 Software isn't friendly for copying from one program to another (especially without source code), so you'd essentially be forced to build a brand new program. #2 if the broker you are trying to do this for does not provide APIs I'd say you're just plain out of luck. Short of explicit help from the broker's software development team, there's no way you will be able to interface with their system without APIs. if this isn't what you meant by your post, please reply and clarify.
Will, you understood correctly. So it's not enough to build a "platform" for a broker simply knowing they use/follow the "FIX" protocol? An API is necessary, then? Thanks! 'risk
Constructing a trading platform from scratch, is quite an undertaking - unless you want a very small feature set. as a heads up, I'd say that the Interactive Brokers TWS interface represents about 20,000 - 50,000 man hours (maybe more - I tend to underestimate how slow other developers are). That includes all the bells and whistles, redundancy, and security. Put simply, making something really easy and intuitive to use, while being highly functional is very difficult. IB has done well with this. as for FIX integration, without knowing more about what broker you are talking about, I'd say at first blush I wouldn't consider developing a platform unless they provided APIs. Either APIs or a whole ton of documentation on how to integrate securely with them via FIX. I'm trying to focus on developing on developing automated software solutions to complex trading strategies - on the order of 50-1000 hrs of development time. If this sounds like you, send me an email. Hope that helps, risktaker! Will website@dudziak.com
1. You should be able to look up on any brokers website fairly quickly and see if they provide a API or not... or call them... 2. If they have an API then you would want to get opinions on THAT API... so you know what you are getting into... is it a high quality API per others opinions who have used it... ! That is very important... 3. If you have or developed an edge... you are usually fine with a non complex system... You can use a Real Time Data Feed to Excel (RTD) + a FIX scripted execution system executed thru Excel with a good API's... to execute even moderately complex trading systems... Excel 2007 has multi processor capabilities and much higher usage of RAM then earlier versions and has new multiâthreaded workbook abilities... The fact that it can receive a high quality market data feed directly into it makes it a platform.. already ready to go for you to tie into an existing API... Why re invent the wheel... You can do in Excel for a few thousand dollars what you might also try to do in a shiny NEW custom created platform for 50to 150K... 4. Complex Custom Software Development is outrageously hard to get done right... factored properly... and created to be flexible for easy maintenance and changes over time... it is very time consuming and very expensive... be very very sure you want to go down that path... http://www.elitetrader.com/vb/showthread.php?s=&threadid=163555
Thanks Edge, I agree for the most part. The process is very time consuming, and can be expensive - but "expensive" is of course relative to the scale and profitability of the system. If your system is some simple numbers crunching, like if the stock goes up 5 minutes in a row, buy... etc. then something simple like that can be handled in excel with vbscripting (usually). however, if your system is more complex, like: 1. automatically download Jim Cramer's picks from the previous night's broadcast 2. scan for high P/E recommendations that have had two sequential up days 3. the stock must open higher 4. then executing a short trade at exactly 9:45am A type of system like this needs custom software implementation for any kind of complete automation; and that's what I've done for myself in the past. Apparently I'm a better programmer and statistician than stock picker. Will website@dudziak.com