Which prefarable time period offers maximum variations of different trends and fluctuations for EA testing? - page 2

[Tonny Obare]

drbmgojanur:
Sir do you advice an EA optimized for years 08 09 10, can perform well in general or current period?

Just because 08-10 were volatile it doesnt mean these years must be included but as we are in 2013 if you go back say 5 years then they are covered in the test. In my view its usually best to optimise for say 5 years but then scruitinise the strategy year by year ensuring it would close at a profit say in 4 out of the 5 years if it were to trade that year alone. Some people will tell you "oh underwater" while ill tell you there is no strategy or investment that will forever just profit 100% some months or years will be closed in losses. Anyone who seeks 100% of the days/weeks/months/years to close in profit will never find such a strategy and its such kraken like ideology that leads many to failure/disappointment. A strategy can be in loss even 75% of the year but then it closses that year at a profit and that profit will remain a profit that year nonetheless.

[Clock]

drbmgojanur:
Sir do you advice an EA optimized for years 08 09 10, can perform well in general or current period?

I'd say as a very general advice:

if your EA performs well from 08-10 because you ran an optimization, you're pretty bound to lose your money very fast.

This of course is not true for ALL stretegies.... but chances are pretty high you were just "curvefitting" those three rather extreme years. That just means, that "by luck" (through optimization) you found those parameters that "accidently" worked a couple of times dureing that period. Long lookback period, short timeframes (10 minutes?) and many deals yre your friend...

[Stephen Njuki]

The primary function of an EA is not to lose money. Making money is secondary. Therefore you should be able to test a well written EA over any period and still have some money left over to place new trades as per your signal. (and ideally this should be done while having as many deals as possible)

[Simon Gniadkowski]

ssn:
The primary function of an EA is not to lose money. Making money is secondary. 

 You don't need an EA for that primary function,  simply don't trade and you will not lose money, no need for an EA.  

 

I see things a little different, the primary purpose of my EAs is to make money,  they don't have a secondary or tertiary purpose . . .  perhaps this is why I don't have an EA live trading right now and other people do    perhaps they are content to BE. 

[Stephen Njuki]

RaptorUK:

 You don't need an EA for that primary function,  simply don't trade and you will not lose money, no need for an EA.  

 

I see things a little different, the primary purpose of my EAs is to make money,  they don't have a secondary or tertiary purpose . . .  perhaps this is why I don't have an EA live trading right now and other people do    perhaps they are content to BE. 

in order for an EA to qualify as one that can make money it must not lose money.

[Simon Gniadkowski]

ssn:

in order for an EA to qualify as one that can make money it must not lose money.

Of course . . .  but you have no need of an EA to accomplish your primary function,  if you want an EA then just don't let it trade.  you can be happy that it is 100% successful in completing it's primary function,  it will never lose a bean.