Strategy tester

New comment
 

hey does anyone know where i can get a decent strategy tester? the one that comes with the interbank package only has about a months worth of data.

 

Hi caldolegare

The problem is not with your software. You must download historical data and try again.

This link explains all about strategy tester(including link to databank):

http://www.metatrader.info/node/67

Bye.

 

thanks that worked perfectly

 

Hi ,

I have just read a good article in mql4 official site MQL4 as a Trader's Tool, or The Advanced Technical Analysis - MQL4 Articles It says:

if a black candle has been closed, what is the probability of forming a white one? This, of course, depends on a great variety of factors, but let us refer to statistics.

//EXAMPLE 2 //Calculate seqences of 1st order //BW means after black going white candledouble BW=0; double WB=0; double BB=0; double WW=0; for(i=Bars;i>0;i--) { if(isBlack(i) &&isWhite(i-1)) BW++; if(isWhite(i) &&isBlack(i-1)) WB++; if(isBlack(i) &&isBlack(i-1)) BB++; if(isWhite(i) &&isWhite(i-1)) WW++; }''We will use this function to draw the candle color distribution according to values of indicators WPR and RSI''...........

My question is lets say I have RSI indicator producing buy-sell signal can I do a test to find most profitable consecutive white or black or black and white candle or white and black candle occured after buy or sell signal ?How can I extentend this little further:

How should I examine the relationship of the consecutive black candles or white candles or blackand white candles close to high or close to low and open to high or open to low ?Your feed backs are welcome in particular Mladen and MrTool?

 
nevar:
Hi ,

I have just read a good article in mql4 official site MQL4 as a Trader's Tool, or The Advanced Technical Analysis - MQL4 Articles It says:

if a black candle has been closed, what is the probability of forming a white one? This, of course, depends on a great variety of factors, but let us refer to statistics.

//EXAMPLE 2 //Calculate seqences of 1st order //BW means after black going white candledouble BW=0; double WB=0; double BB=0; double WW=0; for(i=Bars;i>0;i--) { if(isBlack(i) &&isWhite(i-1)) BW++; if(isWhite(i) &&isBlack(i-1)) WB++; if(isBlack(i) &&isBlack(i-1)) BB++; if(isWhite(i) &&isWhite(i-1)) WW++; }''We will use this function to draw the candle color distribution according to values of indicators WPR and RSI''...........

My question is lets say I have RSI indicator producing buy-sell signal can I do a test to find most profitable consecutive white or black or black and white candle or white and black candle occured after buy or sell signal ?How can I extentend this little further:

How should I examine the relationship of the consecutive black candles or white candles or blackand white candles close to high or close to low and open to high or open to low ?Your feed backs are welcome in particular Mladen and MrTool?

nevar

As far as I see it can not be used meaningfully. The model that is used is giving expected results for a sequence of 2 bars sample (all at around 25% - if the number of tested bars was greater the result would be even closer to those 25% - leaving out the "no result" ones is not justified on a 5 digit broker since it will almost never happen with 5 digit prices) and as it is even more expected, it gives an expected overall result of 50-50. It only means that the trending period lengths are roughly the same over the longer sequence of data. Of course, the criteria that close is open as a "trending criteria" is also a questionable criteria (I rather look at those changes as a normal change of price within a predictable range of prices, but that is my personal opinion)

In my opinion looking for patterns in a 2 bar sample is not justified - the sample is too small for any conclusive results and patterns on such a small sample can not be used in any kind of estimation

____________________

PS: I do not know if that was done on purpose or not in the code (if it was simply a case when a coder did not know the way how metatrader seeks data, for example), but I fail to see what is the purpose of using current and future value in pattern determination. Same results would be achieved with a simple, straightforward current to first previous comparison

 

Thank you Mladen for your valuable feedback.

I explain more clearly:

I want to separate my trades in two parts: those signals created by 50% to 55% win/loss probability (as an example) and those created above 60% on historical data .But when I want to seperate the trades on historical data how should I statistically define the specific data (example 50% to 55% probability that has occured in historical period)?By ocurrence of consecutive black and white candles or white and black candles etc..in 50% to 55% probability historical period?And/or by high low open close relationship of candles occured in 50% to 55% probability historical periodThat is why I thought this article can help or it can be starting point.

New comment
Reason: