70% Win EA but with huge amount of charts errors. How to fix it ? - page 2

New comment
 
Trademanfx:

lol yeah I guess so. Thanks :D


I don't have the code to share. I have only the ex4 file!

Then don't share either if it's a paid one, there's cracking tool to decode ex4 files someone told me recently !! 

 
Felipe Ponce Aragon:
Note: I didn't say that if the EA doesn't use those factors it will perform as in backtest, just that if it had them, it won't. 

If it doesn't have them, often the backtest with low quality is good enough. In that case I would attach it to demo account and let it trade for 1 month, then backtest and compare results only of that month. If you're ok with the comparison then I would attach to live account. 
You are right. I should first let it trade for 1 month and then compare the results. Thank you very much for your advice Felipe.
 
Icham Aidibe:

Then don't share either if it's a paid one, there's cracking tool to decode ex4 files someone told me recently !! 

It's paid and uses account number. So it would be of no use to share it.


Thanks for your advice.

 
Trademanfx:

Thank you for your response.

Why can't they be fixed ? I understand that these mismatched errors won't affect the performance of the EA in the future ?

The EA doesn't use any of the factors you mentioned as far as I know. It also doesn't make a big loss because it has a stop loss based on account percentage  (1% in the backtest).

Ups sorry I forgot to answer this question:  


"Why can't they be fixed ? I understand that these mismatched errors won't affect the performance of the EA in the future ?"


If the EA does not have any of the characteristics I previously mentioned, it's very probable that it will perform in real market 90% as good as in backtest. The mismateched errors are defects in the history of the pair you backtested, but it's imposible to have 100% perfect data, even from expensive backtesting software. The result you see is probably, maybe, what the EA managed to achieve even with the damaged history. So it's not curve fitted, it's a robust EA that will make a profit regardless of market conditions.


You can also try to find expensive history and backtest, but in my experience those backtests are not better. I have seen 99.99% backtests and the EA does not perform even 5% similar to the backtest. So attaching to DEMO and comparing backtest with DEMO result is better for me.


Have a blessed day!

 
Felipe Ponce Aragon:
Ups sorry I forgot to answer this question:  


"Why can't they be fixed ? I understand that these mismatched errors won't affect the performance of the EA in the future ?"


If the EA does not have any of the characteristics I previously mentioned, it's very probable that it will perform in real market 90% as good as in backtest. The mismateched errors are defects in the history of the pair you backtested, but it's imposible to have 100% perfect data, even from expensive backtesting software. The result you see is probably, maybe, what the EA managed to achieve even with the damaged history. So it's not curve fitted, it's a robust EA that will make a profit regardless of market conditions.


You can also try to find expensive history and backtest, but in my experience those backtests are not better. I have seen 99.99% backtests and the EA does not perform even 5% similar to the backtest. So attaching to DEMO and comparing backtest with DEMO result is better for me.


Have a blessed day!

That sounds great. I'm glad that those errors won't be a problem in trading. I'll have to test it on demo for at least a month to see how it will perform. Thank you Felipe for your insight.
12
New comment
Reason: