November 17, 2009 | Posted by andrew Well finally I’m getting some useful results as a result of adding the mini-tournaments with the integral validation stage. Here are two strategies trained on 5 minute GBP/USD dollar data for 2006 – 2008 that also show a profit on 2009 data :-
;; Candidate 1
long entry = (- (+ (* (min hour [...]
Categories: Forex, Genetic Programming | 
Tags: | 
9 Comments »
November 11, 2009 | Posted by andrew I’ve had an interesting problem since I started my mini-tournaments with validation on unseen data in that small runs completed OK but larger ones have either locked up the computer (unresponsive with the hard disk thrashing) or the process had been killed.
At first I suspected that my laptop had been hacked into and the hacker [...]
Categories: C, Genetic Programming | Tags: | No Comments »
November 5, 2009 | Posted by andrew Well after several weeks of running I ended up with two contender algorithms :-
Generation 511: Best = 4962.700000
long entry = (- (+ (* (min hour bar) -30) (vol (trunc bid))) (high (- (abs -40) (abs (* bar hour)))))
long exit = (- (close (min (abs (+ bar minute)) (* (abs [...]
Categories: Forex, Genetic Programming | Tags: | No Comments »
September 30, 2009 | Posted by andrew Although the run was 2 days in, I had noticed that the profitability was not increasing particularly from a level of around 11,000 pips (over 21 months) and that the programs were becoming bloated which was slowing the generations.
Even though I had reinstated the simplification process, the mutation scheme generally encourages increasing complexity over time [...]
Categories: Forex, Genetic Programming | Tags: | 2 Comments »
September 27, 2009 | Posted by andrew I tracked down the last remaining bug in the simplification process and did a mini run for 2 days which failed to produce any more bugs and produced a modestly profitable strategy even without stop losses and the full set of indicators.
I had previously disabled the simplification process as it contained a bug and thought [...]
Categories: Forex, Genetic Programming | Tags: | No Comments »
September 25, 2009 | Posted by andrew Last night I tracked down the segmentation violation that was causing my programs to bomb out – it was an error in the simplification process. I’ve been having second thoughts on simplification so I’ve disabled it for the time being – more on this in a later post.
I also enriched the terminal set to include [...]
Categories: Forex, Genetic Programming | Tags: | No Comments »
September 24, 2009 | Posted by andrew It’s been a while since I’ve done any serious work in C so I’d forgotten about some of the disadvantages: segmentation violations.
The good news is that I’ve now implemented everything that I need in C to start my first run and performance is OK – I’ve estimated that each generation will take about 2 hours [...]
Categories: C, Genetic Programming, Scheme | Tags: | No Comments »
September 23, 2009 | Posted by andrew I’ve had a very busy series of late nights rewriting all of my Scheme code in C. I love Scheme as a language and it has really enabled me to get into this whole genetic programming adventure but alas I’ve reached the performance limits of the language.
The problem came when I tried to load my [...]
Categories: C, Genetic Programming, Scheme | Tags: | No Comments »
September 14, 2009 | Posted by andrew So here’s the plan for creating at least one successful trading strategy for forex :-
As you can see, I will be hosting the trading robots as Expert Advisors in MetaTrader. This makes a lot of sense because many brokers offer MetaTrader based integration so no further work is required to use these for demo and [...]
Categories: Forex, Genetic Programming, MetaTrader, Scheme | Tags: | No Comments »
Recent comments