TimeFrame vs "M" Chart Selection

New comment
 
When an EA sets iMA(NULL,timeframe,period,0,MODE_SMA,PRICE_HIGH,0), can a different "M" Chart selectiion, such as M1, M5 or M30 override this timeframe statement?

Thanks in advance.
Wackena
 
Could you, please, explain what you mean?
Perhaps, you can provide us with the section of your code with iMA function.
 
Could you, please, explain what you mean?
Perhaps, you can provide us with the section of your code with iMA function.


Since this EA is not my creation, I do not feel it is right to disclose programming information.

My question is this, if a TimeFrame (period) is programme within the EA, can this be changed by using a different period selection on a pair chart?

Wackena
 
My question is this, if a TimeFrame (period) is programme within the EA, can this be changed by using a different period selection on a pair chart?


If the TimeFrame is strictly specified within EA, the change of timeframe on a pair chart will not interfere the working of the expert advisor.
 
I have an EA that has TF-15 specified within the programme. When backtested on M1 period, I get very good results. When backtested at period M15, I get very poor results. Is there something in the strategy tester that allows period selection ot override or at least effect test results. If this is is happening in strategy tester, does it also happen in forward testing or Live.

Or, could it be that TF=15 is specified within the programme, but not correctly implemented as the programme runs.

Wackena
 
Testing of experts on a timeframe other than that specified within the code may be incorrect. It would be better either to test it on the same timeframe, which is given in the iMA function, or not to specify any timeframe in the code at all:
iMA(NULL,0,period,0,MODE_SMA,PRICE_HIGH,0)
 
Most opinions that I have heard state that it doesn't matter which Chart period you select, if the TimeFrame is hardcoded into the EA.

I started 3 side-by-side tests to compare using the same settings on Firebird v63g EA that has TimeFrame hardcoded to 15 within the programme. I used 3 different Chart periods - M1, M15 & M30.. Just to note, that I used 2 forward test Demos and 1 Live account for this test. Here are the results from June 14 (start at 1800 GMT) to June 16 (end at 2400 GMT).

M1 (Live)
eur/usd - 10 trades (10 wins, 0 loss)
gbp/usd - 19 trades (18 wins, 1 loss)
usd/chf - 15 trades (14 wins, 1 loss)
usd/jpy - 6 trades (5 wins, 1 loss)

M15 (Demo)
eur/usd - 14 trades (14 wins, 0 loss)
gbp/usd - 4 trades (4 wins, 0 loss)
usd/chf - 5 trades (5 wins, 0 loss)
usd/jpy - 5 trades (5 wins, 0 loss)

M30 (Demo)
eur/usd - 1 trades (1 wins, 0 loss)
gbp/usd - 10 trades (10 wins, 0 loss)
usd/chf - 3 trades (3 wins, 0l oss)
usd/jpy - 2 trades (2 wins, 0 loss)

If there is supposedly no difference between Live and Demo resuts, there seems to be a significant difference in trading activity. Also, it seems M15 and M30 periods, traded less times and "maybe" handles trending spikes better. I say "maybe", because this is suxh a short test period. Also, there are some indications here that different currency pairs may perform better on different Chart periods on ths Firebird v63g EA.

Wackena
New comment
Reason: