New EA based on the !xMeter - page 20
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
Forward testing
Very good results on EURUSD last week
THIS WEEK:
This week's results for the EurUsd pair is nothing to write home about.
The market for this pair was very flat.
statement_1xmeter_2.htm
.
There is a bug which make the backtests useless; in two words, the EA knows the future when it uses the daily values High, Low and Close : those values are those of the closed daily bar. This means that at 8 AM it already knows what will happen. But there is no meaning to use the previous bar.
Another problem is to define a broker independent day start.
Another bug is to assign a zero value if the range is null (ie the first tick of the day); it should be 50. Anyway, the values aren't very meaningfull at the begining of the day, when the range is not yet wide. Maybe a trailing time range would be better.
Also, it seems to me that the values should be normalized and maybe wheigted by the number of pairs giving an information.
I am currently working on all that and will keep you informed.
Nevertheless, forward testing should be interesting...
What is this statement for?
as i see you set the HighLevel and LowLevel to 3 !!! why???
identical properties of the expert. identical time. different results.
identical properties of the expert. identical time. different results.
As i told before the backtest won`t work with this EA and all the backtests of this EA are just a beauty of metatrader`s bug.(If someone disagree with the knowledge of mql then welcome to discus it.).
BTW- what was the reason of setting HighLevel and LowLevel both to 3?
any special reason?
Very good results on EURUSD last week
THIS WEEK:
This week's results for the EurUsd pair is nothing to write home about.
The market for this pair was very flat.
statement_1xmeter_2.htm
.bend through a tester, for this week. сompare result.
BTW- what was the reason of setting HighLevel and LowLevel both to 3?
any special reason?оptimization.
...beauty of metatrader`s bug...
100% agree
There is a bug which make the backtests useless; in two words, the EA knows the future when it uses the daily values High, Low and Close : those values are those of the closed daily bar. This means that at 8 AM it already knows what will happen. But there is no meaning to use the previous bar.
Another problem is to define a broker independent day start.
Another bug is to assign a zero value if the range is null (ie the first tick of the day); it should be 50. Anyway, the values aren't very meaningfull at the begining of the day, when the range is not yet wide. Maybe a trailing time range would be better.
Also, it seems to me that the values should be normalized and maybe wheigted by the number of pairs giving an information.
I am currently working on all that and will keep you informed.
Nevertheless, forward testing should be interesting...Backtest is useless??? because they know the future from daily values High, Low and Close ???
. Are you trying to say that this ea was only good in backtesting???
The backtesting for EU from june 2004 to jan 2005 shown a very good smooth climbing curve. The bigest drawdowns was about 60% from your ac ac shown in circle.
The only doubt to me was why some of the situation this backtest seems like in idle mode(no trade open) sometime for 1 day even though from the candle it shown quit heavy movement. You can see it from post # 403 to 404 below graf.