[A Question About EA Testing]Test Results are different each time, is it common? - page 2

New comment
 
1105:

dabbler：

Thanks for your immediate feedback. Could u please explain more about "non-repeating" test? what I did was handling test for twice, each time I set same start date and end date, then press button "start", any special thing I should pay attention to?

Well you might say something like 

if( MathRand() > 10000 ){
    // exit trades
}

I wouldn't recommend such a thing, but some people seem to think trading randomly is successful. (I don't).

In order for the code above to simulate differently each time you would need to seed the random number generator with the time of day or some other changing value.

As far as the spread goes I have simulated one day and had an EA work and the next day the archived identical EA didn't work (profitably). This was simply because the spread at the weekend was much worse than during the week!

 
1105:

Hi Buddies,

I just wonder whether I can find an agent who provides the lowest spread in this world, maybe 20? or even lower?


In UK I have seen Adverts on Bloomburg by a Broker saying they have zero spreads - Sorry can't remember the name - Maybe they tax you on profits instead.
 
Ickyrus:

In UK I have seen Adverts on Bloomburg by a Broker saying they have zero spreads - Sorry can't remember the name - Maybe they tax you on profits instead.


That's really an exciting news. Seems I should visit website of FSA to have a screen check))

Thanks a lot!

12
New comment