How do you come up with a (hopefully) profitable trading system (set of rules for executing strategy) from the initial idea? I would like to learn more about this process of creating trading system/plan, so I would be thankful if anybody is willing to share his methodology. Thanks
Like any software project: first planning and prototyping (ie. pretesting the system), then developing, then testing (using simulated data, and also real data), then integrating, ie. installing/going live. Keeping logs, analysing logs, and refining the program, ie. making it more robust after each found issue.
Of course the system logic is the most important part: that should be constantly under review with the experience made from the live sessions. Ie. a fulltime monitoring, analysis, refinement. At least in the inital phase, ie. some months long.
Usually one uses simulated market data (GBM with possibly some extensions like jump-diffusion), and an options pricing model like Black-Scholes-Merton (BSM) for most of the the simulation tests.
Strategy development begins with trading ideas. Those are tested in simulations. If it gives good results then one can integrate it into the program, or make a new program for it.
From my painful experience: it's a very time-consuming and frustrating hard job, and is of course also a dangerous job because of the possible liabilities should things go wrong... The risk is high, but the reward can also be high...
One should let develop and maintain the software only by experienced developers as it is a critial property.
From idea to a plan: first write a concept, then refine it to a detailed plan... It's like programming with if-then-else like constructs...
Here are some links:
http://www.futuresmag.com/2012/08/31/guide-trading-system-development
http://www.investopedia.com/university/tradingsystems/
http://www.investopedia.com/university/systemcoding/
https://www.wpi.edu/Pubs/E-project/...ed/Interactive_Qualifying_Project_Group_5.pdf
.
