The relationship between backtest results with forward test at ATC 2012
- The dominant factors to be the winner at ATC 2012
- ATC 2012 participants: Do you want to share you EA in future?
- Can you fix this code!
So far, no one choice linear relationship. What do you think about that? Does it mean no body believe to back test results?
For me the big mistake in using backtesting is choose the wrong methodology.
It is easy to find the ideal simulation from the past, hyper optimized, that will never happen again.
Poorly designed tests serve only to leverage dreams ;-)
For me the big mistake in using backtesting is choose the wrong methodology.
It is easy to find the ideal simulation from the past, hyper optimized, that will never happen again.
Poorly designed tests serve only to leverage dreams ;-)
Hyper optimized .... hmmm I know that. If so, is hyper optimization not advisable to obtain a suitable parameter values? What do you recommended to do optimization?
Good question!
In my opinion, market future is allways unpredictable and there is no magic solutions, but change and adjust the strategy and good tests methodologies can make it more predictable, if we can say it.
For example, supose a simple strategy of two moving averages crossing. Backtesting can be exciting, but hardly the performance will repeat, and the ideal is to use other filters and forward testing to double check.
Another way (that I like too much) is use random factors to several parameters, like Execution Random Delay option in Strategy Tester. In our company we use some libs to generate similar random factors to filter "hyper optimized" setups.
"The Expert Advisors that were successful in the Strategy Tester, showed good results during the competition"
I personally agree with that.
So let's wait and see, whether it also will happen in the ATC 2012.
- championship.mql5.com
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use