EA: Awesome Backtestingresults - Crappy Livetradingresults (demo)
I do not think the problem is backtesting. I think the market is always unique. There is a so called "butterfly" effect in this market. If your EA does not have the capability to handle all market conditions, a small change in your EA settings can give you totally different result, like to select different date. For the same reason, the forward testing has no guaranty to give you the same results as the backtesting. It is all about the EA strategy - how to build an EA to handle all the market conditions and win out in the long run. Very few have succussed.
Well the MT4 calcs its Ticks Ask and Bid itself with a fixed spread of the moment you start the backstest - my experience.
It seems that you have optimized your EA against the way MT4 cals it Ticks.
I guess you need to change you quotes history in order to get 99% model quality(send me p.m.). I know there are two options.
Another point could be your entries and exits, but for this one need to the code.
Good luck Gooly
Starting with Build 509 you get to choose
Hey everyone,
I have an MQL4-EA which has awesome Backtestingresults for the last 6 month. It works on M1 DE_30, it's based on scalping and these are the results of 6 month:
When I developed it, I only looked at one month and these are my first results on 6 month (so I think it's not overfitted for this period!). However, I tried this EA on livetrading (demo) this week, and it had really crappy results (-3.500$). It could be just bad luck (the half year had a long downward-period at the end...) or my backtestresults are just crappy and unrealistic!
- I know these results are unrealistic high, but how can I backtest my EA to get a "real" result?
- I backtested the period of this week and the results were very close to what the livetrading did (what is very good in my opinion, it means this week was just bad luck for the EA).
There is no such thing as "luck" . . . there is always a reason, you just need to find and understand it.
6 months is not long enough, test for longer and on other symbols.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
Hey everyone,
I have an MQL4-EA which has awesome Backtestingresults for the last 6 month. It works on M1 DE_30, it's based on scalping and these are the results of 6 month:
When I developed it, I only looked at one month and these are my first results on 6 month (so I think it's not overfitted for this period!). However, I tried this EA on livetrading (demo) this week, and it had really crappy results (-3.500$). It could be just bad luck (the half year had a long downward-period at the end...) or my backtestresults are just crappy and unrealistic!
- I know these results are unrealistic high, but how can I backtest my EA to get a "real" result?
- I backtested the period of this week and the results were very close to what the livetrading did (what is very good in my opinion, it means this week was just bad luck for the EA). However, the number of trades and the times of the trades were complety different!? Backtesting yesterday: 56 Trades whereas the livetrading did 39 Trades -.- Why? :'(
- Now I've seen that the modelquality is only 24% (also visible in the image above). Does it has any influence? How can I make this better?
- I would like to backtest it on more data, from 2012. However, the MT4 just doesn't have any more M1-Data. When I give a period in 2012 he just doesn't do anything at all. Even if I click on the historydata and tell him to download the M1, I just get the respond "No more recent data available" (translation from german).
To sum it up: How can I backtest my EA realisticly?
Thanks for your help :)