Tester in MT4 Build 200 - page 3
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
I also have the same problem as I was back testing a system that was doing great then I got totaly diffrent results after upgrading from 198 to 200 ..
I am using alpari 1M data on both (198 and 200) and I dont login to any server with the testing terminal to prevent any change from outside (I feed it manually using alpari data only).
I didnt change anything except upgrading to build 200 ..
Now I dont know which results to trust given that both results are 90% .
I also have the same problem as I was back testing a system that was doing great then I got totaly diffrent results after upgrading from 198 to 200 ..
I am using alpari 1M data on both (198 and 200) and I dont login to any server with the testing terminal to prevent any change from outside (I feed it manually using alpari data only).
I didnt change anything except upgrading to build 200 ..
Now I dont know which results to trust given that both results are 90% .First of all, what do you mean whe saying: "I got totaly diffrent results"? Do you mean the graphical representation (Tester > GRAPH tab) or the actual statement (profits/losses/etc)?
Well, what I can tell you on this one is this: if you are talking about graphical representation and using the same data and not using History Center and the result looks different than 198, than its probably due to this change/ improvement in the MT:
In that case, I would trust build 200, rather than 198, since what they've done is they made it better, not worse. So theorhetically 200 should be more "real"/ precise.
Cheers,
Diam0nd
I LOVE
The system is still profitable, it is just that I got diffrent results regarding profits (Less by half), losses (more), drawdown (24% instead of 12%)
As I said I didnt change anything other than the MT4 bulid to 200.
If I trust the new results then I will have to revise the system to lower the drawdown which is more important to me than increasing profits .
Thats why I was asking if I can trust the new results and I was looking for feedback from simillar situation.
The system is still profitable, it is just that I got diffrent results regarding profits (Less by half), losses (more), drawdown (24% instead of 12%)
As I said I didnt change anything other than the MT4 bulid to 200.
If I trust the new results then I will have to revise the system to lower the drawdown which is more important to me than increasing profits .
Thats why I was asking if I can trust the new results and I was looking for feedback from simillar situation.Aha, so not only the picture but the actual statement has changed, right? Well, as I said, if I were you I'd trust the new build, since MetaQuotes are working on improving the tester and make it closer to real trading and not vise versa, of course, so I'd trust it
Cheers,
Diam0nd
I LOVE
Thank you Diam0nd
If the new tester gives worst results then it's probably closer to the reality.
It could also mean that the EA tested is very sensitive to small changes and that is maybe not very good.
Thank you Diam0nd
My pleasure. Hope the info helped
If the new tester gives worst results then it's probably closer to the reality. It could also mean that the EA tested is very sensitive to small changes and that is maybe not very good.
Good point. An excessively sensitive expert IS a bad thing. You might not wanna make your expert too sensitive. This is especially "grabs attention" when you try to test your EA on different brokers data, since each and every brokers filters/smoothers his market data differently.
Aha, so not only the picture but the actual statement has changed, right? Well, as I said, if I were you I'd trust the new build, since MetaQuotes are working on improving the tester and make it closer to real trading and not vise versa, of course, so I'd trust it
Ok. But what about the data you can download using the button? I get very different results on build 200 with Alpari 1M data versus the download data. Both build 200. This is really making me question the tester.
And the strange thing is, I only trade on the opening of new H4 bars!
The only tick changes I make is a very large trailing stop (300-400 pips). Based on the optimizer, the stop loss rarely gets invoked (ie the optimization becomes flat at about 300-350 pips).
Can someone else do an Alpari versus download test using build 200? I"d be interested in seeing if I'm the only one seeing differences.
Ok. But what about the data you can download using the button? I get very different results on build 200 with Alpari 1M data versus the download data. Both build 200. This is really making me question the tester.
And the strange thing is, I only trade on the opening of new H4 bars!
The only tick changes I make is a very large trailing stop (300-400 pips). Based on the optimizer, the stop loss rarely gets invoked (ie the optimization becomes flat at about 300-350 pips).
Can someone else do an Alpari versus download test using build 200? I"d be interested in seeing if I'm the only one seeing differences.Is it me or do you sound like you expect similar perfomance from 2 different data sources? I mean, each and every broker will give you different perfomance.
Ok. But what about the data you can download using the button? I get very different results on build 200 with Alpari 1M data versus the download data. Both build 200. This is really making me question the tester.
Fine. Let me ask you this then: WHY do you question the data from the History Center? I guess since you ASSUME that data from Alpari is better. Why is it better? Because it gives better results?
Cheers,
Diam0nd
I LOVE
Is it me or do you sound like you expect similar perfomance from 2 different data sources? I mean, each and every broker will give you different perfomance.
Fine. Let me ask you this then: WHY do you question the data from the History Center? I guess since you ASSUME that data from Alpari is better. Why is it better? Because it gives better results?
Cheers,
Diam0nd
I LOVE
Yes, I do get better results with Alpari.
But since the results are too different, I am just questioning the integrity. If I was off by a few hundred bucks, no big deal. But the difference is substantial.
I've also heard that Alpari is better. Maybe because since people get 90% (before build 200), so they assert it's simply better for that reason. But now that 200 is here and everybody is getting 90%, I think we even have to question if 90% is good enough. Again, findings are questioned simply due to largedifference in result.
I'm just asking other individuals to do the same test I"ve done: Alpari data versus non-Alpari data. Disregard small differences in your testing, since you are dealing with different brokers. But do question largedifferences.