Great EA in backtest! - page 63
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
here is an article that explains what are the numbers at backtesting report:
https://www.mql5.com/en/articles/1486
'Bars in test' field displays the depth of the history on which the modeling was based.
so i understand if we have both same data from alpari, and the same period is tested we sholud have at least equal number of bars. so why we have different? i'm tired.I agree with you. My second backtest was identical to the first one - like yours. I have attached two screenshots concerning the data in MT4. Please compare this with yours. My FXDD software is build 197.
I am tired too, but this problem has set "EnableSleeping=false". I have downloaded now IBFXmini software and have integrated the data. And just while writing this I can see that this backtester draws a curve like yours. It is the software. Do I need a new broker?
I have attached two screenshots concerning the data in MT4. Please compare this with yours.
mine is different :/ (attached). i'll try to import data once again.
Do I need a new broker?
i don't think so. we need some time to figure out why there are so many differences.
for the past 24 hours i have watched the live account and the demo account match each other pretty much trade for trade with only about a 1 pip difference between the two. as of present they are both down from the balance they started with.
I wonder how much to stress over the differences we discover. I wonder if the goal isn't just to find a level of performance we can find acceptable. And develop an understanding of what is producing the desired effects. It seems like there is a trade off between drawdown and profitability. If I reduce the lots to so small that there is hardly any drawdown there is also little profit.
Hi kalamari
At you in adjustments at backtest EnablePIPTimer = true?
Different results problem not solved yet but I have had some trades (demo) with this expert, 4 winner and 1 looser, since yesterday.
I am tired too, but this problem has set "EnableSleeping=false".
You have found such parameters in what version EA?
We need to be clear about this.
Let's observe the difference of our tests:
1. Broker platform.
IBFX has other labels for mini lots (EURUSDm, 1.0)
FXDD mini lot is 0.1
2. Reports.
Kalamari: 16418 bars
kat: 14219 bars
Kalamari: 1534666 ticks modelled
kat: 2980208 ticks modelled
Can this influence the results?
The data we have used is the same but in the strict sense we have to adjust the GMT shift. Alpari data is CET, which is now GMT + 2.
Any idea to clear up?
I will test again, too.Hi Kat & Kalamari,
Both of you are feeding Alpari data into different broker and the test result is not same. Possible explanation:
(1) The number of data is not same for M1, M5.. H1. It must be exactly the same.
(2) MetaTrader still has some bugs, sometimes, the result can be different if you tried on H1, then continue to H4, then continue to other currencies and go back to H1 etc.... However if you are facing this, you can quit the MT, then you must delete the catche files
C:\Program FIles\MetaTrader Alpari\tester\history
C:\Program FIles\MetaTrader Alpari\tester\log
(3) Then contract of the currencies is not the same. There is specific variables for each broker like: Spread, DIgits, Stop Level, Pendings, contract size, Profit calculation, swap type, swap long, swap short, margin calucaltion mode, margin hedge, etc. This can be seen on "Symbol Properties" of the Strategy Tester.
(4) And it still consider your setting of your active account used. For example, if you use "leverage 1:100" account will have result different from "leverage 1:500". Though it is testing mode, but most of the information used is taken from your active account.
So yes, different broker will behave differently.
You have found such parameters in what version EA?
In all experts except the Holy Grail.
Hi Kat & Kalamari,
Both of you are feeding Alpari data into different broker and the test result is not same. Possible explanation:
(1) The number of data is not same for M1, M5.. H1. It must be exactly the same.
(2) MetaTrader still has some bugs, sometimes, the result can be different if you tried on H1, then continue to H4, then continue to other currencies and go back to H1 etc.... However if you are facing this, you can quit the MT, then you must delete the catche files
C:\Program FIles\MetaTrader Alpari\tester\history
C:\Program FIles\MetaTrader Alpari\tester\log
(3) Then contract of the currencies is not the same. There is specific variables for each broker like: Spread, DIgits, Stop Level, Pendings, contract size, Profit calculation, swap type, swap long, swap short, margin calucaltion mode, margin hedge, etc. This can be seen on "Symbol Properties" of the Strategy Tester.
(4) And it still consider your setting of your active account used. For example, if you use "leverage 1:100" account will have result different from "leverage 1:500". Though it is testing mode, but most of the information used is taken from your active account.
So yes, different broker will behave differently.fikko, thanks, this make it clearer. I am now testing on a fresh IBFX paltform, maybe these results will be more reliable. I will keep at this because I think the effort is worth if we could confirm kalamari's results, which are really amaizing.
However we have to start once again but strict methodically.
bad news guys
i've downloaded data from alpari once again, cleared historical data in mt and imported fresh. results attached.
i don't know why i had such good results last time. maybe the reason was corrupted data in some way. i'm very sorry for that. i'm so dissapointed. i'll leave CT for some time. once again sorry guys.