ok, but as a newbie I'll wait for someone more experienced to give high and low limits to different parameters.

I'm currently testing 2.4 on presets from Holyguy7 and from 1:00 AM I only made 3 trades (1:10 sell usdcad on daily BAD; 7:29 and 14:29 gbpusd on H1 first 28 pips and second actually loser). It seems to be normal, anyone have same results?

Also, to allow multiple timeframes tests, I changed three lines on 2.6 version:
1 -
extern int ID;
becomes
extern int ID_BASE=100000; to give a number to PG_2.6 (here 1)

2 - I added next line just under "int Bar;"
int ID;

3 - I added init() function
int init(){
ID=ID_BASE+Period(); return(0);
}

So on a Daily chart magicnumber will be 101440 and will appear on trade's comment. It will be easier to analyse results for each timeframe and kind of parameters. I didn't test it yet but it should work.
and for a specific chart you will have a fixed ID and not a random number.

Also, maybe we could add a initial takeprofit using superclose function. Just in case of computer crash.

PS: I made other changes that I must delete before posting. tell me if interested.

I have not recived any signals on majors ( Daily and Weekly)

It could have been great on all JPY if it had comes up a little more.

Maybe we could start by pending orders based on lower timeframe bars (M1?) to look for few pips more (covering spread for example). A kind of trailing start.

Great idea! The only thing to keep in mind is if you are going to run different tests on the same pair and TF to change the Base ID. I added those changes, but made one small alteration. ...because the Period can can change from the variable "period" (different from chart), I did this instead:ID=ID_BASE+P(); return(0);
The P() returns the actual period in use by the program. Great idea...what other changes have you played with?
Here is 2.6.2.

__________________
"Anyone who has never made a mistake has never tried anything new." -Albert Einstein

Sure P() is better. other changes were about comments but uninteresting.

Who code the superclose function? I am trying to well understand it but to had a initial takeprofit (which should be simple, a value like 2 times takeprofit) maybe the initial coder will do it faster. It's just in case of crash and not the most important while testing. can wait and I will have time to do it when tests are launched.

About tests plan, you seems to be the right person to first define limits to parameters. And after we'll see how many we are and which tests we should do first.

Cool. How about something simple like if the main signal is ready to go long and stochastics on M1 (K<D), then wait till the cross, then get in? ...or did you have something else in mind?

__________________
"Anyone who has never made a mistake has never tried anything new." -Albert Einstein

I'm really a newbie (only 3 months that I read a lot and backtest) but I have read only good things about stochs (not enought pratice to have a real advice).
But to keep this EA free of indicators (the initial idea), whe could just use trailing bars on M1? or M5? to be a the best place to trade and avoid trading in the bad way.

But let make the first steps in a real forward testing plan and based on results we'll see better if a system like trailing bars or stochs could have helped in a better entry, no?

Can you restate your question? I don't quite understand it. Sorry.

__________________
"Anyone who has never made a mistake has never tried anything new." -Albert Einstein

I don't like to have orders with no SL or no TP and when using superclose(), there is no TP untill Trailing be activated. So I ask why not set an initial takeprofit (before superclose set one) to be secured in case of computercrash. That's why I proposed two times the value of takeprofit variable.

Also to initiate on the test plan, parameters are:

Stoploss: 10 to 30?
Takeprofit: 20 to 100?
TimeFrames: M1 M15 H4 Daily Weekly..
Currencies: 12 from Holyguy7 selection? only those with a little spread (<=5)?

those are just suggestions.

Nich,

I think using any kind of indicator is a bad idea. The strength of the system is because it depends on price action and adding any derivatives and degrees of freedom will only degrade the performance.

Just my $0.02.

Maji