is there any such software out there I can buy ..........that .......

 
is there any such software out there I can buy ..........that .......
<> has a short list of Expert advisors on the LHS . Within each EA you can specify different setting combinations.
<> has a list of currency pairs on the RHS

and it will just automatically ( feed in strategy tester ) test ( in serial or parallel ) each currency pair and each EA combination (within each EA it can use different preset and test out those combination too )
e.g. EA1 Preset1 USD CHF
EA1 Preset2 USD CHF
EA1 Preset1 EUR USD
etc

I would like to buy any such
 
crankshaftman wrote >>
is there any such software out there I can buy ..........that .......
<> has a short list of Expert advisors on the LHS . Within each EA you can specify different setting combinations.
<> has a list of currency pairs on the RHS

and it will just automatically ( feed in strategy tester ) test ( in serial or parallel ) each currency pair and each EA combination (within each EA it can use different preset and test out those combination too )
e.g. EA1 Preset1 USD CHF
EA1 Preset2 USD CHF
EA1 Preset1 EUR USD
etc

I would like to buy any such


I've never seen something like that.

Not quite what you're looking for, but you may be interested in FXDD Auto (Google it).

My only other comment is that searching through a list of EAs and parameters and currency pairs is greatly increasing the degrees of freedom of your optimisation. In orther words you will be in far more danger of over-optimisation, delivering a curve-fitted backtest that will never be repeatable.

Paul

 
Install the EAs. Open the tester, select one EA, select one symbol, select one period, test and repeat.
 

Questions tho phampton: If the system generator discovers something coded by someone else by hand are both systems 'curve fitted'? Are all purely technical systems 'curve fits'? Is there such a thing as a system that works on all time frame and currency pairs? If your system tests well for the past ten years, does this mean that it will work longer in the future as compared to a system that tests well over the last month? (Honest questions for debate... :D)

 
arachnode.net wrote >>

Questions tho phampton: If the system generator discovers something coded by someone else by hand are both systems 'curve fitted'? Are all purely technical systems 'curve fits'? Is there such a thing as a system that works on all time frame and currency pairs? If your system tests well for the past ten years, does this mean that it will work longer in the future as compared to a system that tests well over the last month? (Honest questions for debate... :D)


I'm just saying that when you have more choice, then there is more chance of curve-fitting. Using FXDD Auto, for example, one can create a near-perfect composite backtest by chosing systems and currency pairs. But it will almost certainly be useless in the future. I try to avoid systems that work well on one currency pair and poorly on all others because they are a lot more likely to be curve-fitted.

Trying to overcome the "past performance is not indicative of future gains" mantra is the one major challenge of system development. Even your statement here is not necesarilly true: "If your system tests well for the past ten years, does this mean that it will work longer in the future as compared to a system that tests well over the last month?".

After all, I could create a "system" which buys at the bottom of every dip, by literally recording the exact time that these occurred. Ridiculous? Not really - just an extreme example of curve-fitting.

Paul
Reason: