I am working on building an S&P trading model with an emphasis on risk optimization and quantitative pattern recognition. The general idea is to use price/volume factors based on historical backtests to increase the probability of making a profitable trade. Of course this is much easier said than done...but I'm not looking to be able to predict big movements, I am looking to predict the next few ticks with a probability that is very slightly over 50%. Then by applying this to a large number of trades in one day, a small profit can be realized. Finally, when convinced of the stability of the model, the *wise* use of margin can greatly increase returns while keeping risk at acceptable levels. Has anyone ever thought about something like this? Just wanted to throw out this idea to get some feedback. Thanks.
Sounds good. And obviously people thought about it. Now if you laid down ideas on the implementation, that would be interesting.