Another issue that has poped up. I have increased the margin level. Then I have stopped the running Zorro with real money and when asked closing open trades, I clicked NO. And when asked to "Stop Trading" I also answered NO.

When installing again Zorro to the real account, now the color of the band is all red when before closing it was green and red. I am just wondering if it will resume the open trades and will continue monitoring them or if it will start from scratch. In case it does not monitor the open trades, does it mean I will have to monitor them manually? there are 9 open trades.

Now the slide of Margin has increased to what I have indicated in the MMax configuration.

The meaning of the red/green band is described here:

http://manual.zorro-trader.com/trading.htm

The left segment of the band is the closed profit so far. Which is of course zero directly after starting a strategy.

Thank you jcl, that is clear. And bytheway, Zorro undertook the monitoring of the open trades.

Regarding the re-test issue, does anyone have a glue why sometimes it works and some times not? It can not be to the UAC business because it is set to "off" since the beginning.

Can it be due to VPS constrains?

You know the word is 'clue', right? And you are just being quirky?

Thank you for the correction, DMB.

Coming back to the re-Test issue. I jus read in the forum somewhere that for the S version it gives this problem so it can not be because of the VPS. Could it be because of overloading the system of Zorro S at certain moments? I am just shooting...

Your system is always tested when you click "Test". But it is only compiled when you clicked "Edit" or when you changed something in the code. As long as the code is unchanged, Zorro has no need to compile it again.

If your system behaves differently every time you test it, the most likely reason is a bug of your script. You should then debug it and find out where the different behavior comes from.

Thank you for your patience, jcl. The script is a simple script which depending on the parameters of Stop Loss and Target gives around 100% Anual Return for EUCHF:


A1,A2,A3 Ar62% compiling...............
Error 055: EUR/CHF 2009 price history missing
Error 055: EUR/CHF 2010 price history missing
Error 055: EUR/CHF 2014 price history missing
BackTest: A1,A2,A3 Ar62% EUR/CHF 2009..2014
Profit 4013$ MI 118$ DD 1015$ Capital 3554$
Trades 4280 Win 42% Avg +11.6p Bars 56
AR 107% PF 1.55 SR 0.78 UI 9.3% Error 12%

The script is the following:

function run()
{
BarPeriod = 240;

vars A1 = series(priceClose(2));
vars A2 = series(priceClose(1));
vars A3 = series(price(0));
//Vlr = 10*PIP


if (A1 < A2 and A3 < A1) enterLong();
if (A1 > A2 and A3 > A1) enterShort();

Stop = 35*PIP;
TakeProfit = 80*PIP;
//ExitTime = 1;

}

I have been testing it with different values of stop loss and target, and after your clarification I understand that when Test doesn't give any result it is because some previous result was the same as present one. The trick I had to do for keep on giving results is to change slightly the values of Stop Loss and Target, like one more figure or so, and after some runs, I can put back the values which I was interested to know and then it picks them up and gives a result. Result which I forgot from the first run it gives it back again doing this trick.

Let me ask another question. Is there any way to know if the backtests really perform what is instructed in the script?

Let me clarify this question. In Metatrader, when doing backtests you could check if the trades performed where done according to the instructions. Most of the cases, even when the indicator upon which the EA was build up indicated perfectly the moments when to enter and exit, the trades were not performed according to the what one would expect from the indicator, making the backtest very unreliable.

In the case of Zorro, is there any place where one could look, even for some short period of tested time, to see if trades are performed according to the instructions of the script?

The bug was writing A1 instead of A1[0]. So your script trades random and can't produce a result. Using series is explained in Workshop 4. Also you're testing an asset with some history data missing.

I do not understand your second problem, though. MT4 backtests are not unreliable because MT4 would not perform trades. They are unreliable for other reasons related to the tick-based execution, to the MQL4 structure and to in-sample testing.

Allrigth, then I rewrite the script as follows and apply it to the EURUSD that has no problem of history data.


function run()
{
BarPeriod = 1440;

vars A1 = series(priceClose(2));
vars A2 = series(priceClose(1));
vars A3 = series(price(0));
//Vlr = 10*PIP


if (A1[0] < A2[0] and A3[0] < A1[0]) enterLong();
if (A1[0] > A2[0] and A3[0] > A1[0]) enterShort();

Stop = 30*PIP;
TakeProfit = 500*PIP;
//ExitTime = 1;

}

This gives the following results:


A1,A2,A3 Ar62% compiling...............
Error 055: EUR/USD 2014 price history missing
BackTest: A1,A2,A3 Ar62% EUR/USD 2009..2014
Profit 173$ MI 3$ DD 61$ Capital 60$
Trades 176 Win 18% Avg +12.9p Bars 3
AR 74% PF 1.48 SR 0.75 UI 13.3% Error 54%

That is, an annual return of 74%, which is not a big deal though the script can be refined and get better and more robust results. It is just an exemple for the sake of the discussion here. The main thing is that the re-Test problem seem to have been solved. I re-tested many times and now it gives results every time, so I assume that the script has now no bug, isn't it?