Optimization with genetic algorithm
The preliminary version of the client terminal with genetic optimization is released.
The genetic algorithm can be enabled in properties of the tested expert, the tab "Testing". The genetic algorithm works only in case if the total of passes exceeds 2000.
Client terminal can be downloaded using https://download.mql5.com/cdn/web/metaquotes.software.corp/mt5/mt5setup.exe?utm_campaign=MQL5.community
The genetic algorithm can be enabled in properties of the tested expert, the tab "Testing". The genetic algorithm works only in case if the total of passes exceeds 2000.
Client terminal can be downloaded using https://download.mql5.com/cdn/web/metaquotes.software.corp/mt5/mt5setup.exe?utm_campaign=MQL5.community
What does exactly 'genetic optimization' means? Thank you.
1. Randomly collect population
2. Crossing over pairs of chromosomes
3. Random mutations
4. Random inversions
5. Selecting best chromosomes
6. goto 2
2. Crossing over pairs of chromosomes
3. Random mutations
4. Random inversions
5. Selecting best chromosomes
6. goto 2
Well done MetaQuotes! I've been using genetic optimisation now for some time and have found that it's THE best way of finding the best parameters in complex trading systems. For those of you who aren't familiar with genetic algorithms in general, have a look at http://en.wikipedia.org/wiki/Genetic_algorithms.
BUT, getting a trading system to work on historical data doesn't mean that it'll work satisfactorily in the future. My software has produced trading systems showing 200+% profit on historical data, only to fail when I then run that trading system on data not included in the training data. The ever present problem you have with Neural networks (i.e. over optimisation) also occurs when using genetic optimisation techniques......
By the way, is anyone else out there using genetic computing techniques in trading? If so, have you had any real successes or are you also coming up with supposedly fantastic trading systems that in a live situation, only produce 5-10% returns?
Trading is a HARD way to make a living... Still, it's a hell of a lot more exciting than some boring desk job!
BUT, getting a trading system to work on historical data doesn't mean that it'll work satisfactorily in the future. My software has produced trading systems showing 200+% profit on historical data, only to fail when I then run that trading system on data not included in the training data. The ever present problem you have with Neural networks (i.e. over optimisation) also occurs when using genetic optimisation techniques......
By the way, is anyone else out there using genetic computing techniques in trading? If so, have you had any real successes or are you also coming up with supposedly fantastic trading systems that in a live situation, only produce 5-10% returns?
Trading is a HARD way to make a living... Still, it's a hell of a lot more exciting than some boring desk job!
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
The genetic algorithm can be enabled in properties of the tested expert, the tab "Testing". The genetic algorithm works only in case if the total of passes exceeds 2000.
Client terminal can be downloaded using https://download.mql5.com/cdn/web/metaquotes.software.corp/mt5/mt5setup.exe?utm_campaign=MQL5.community