Tester giving identical results even with a factor of 10 difference in spread Why?
In my quest to assess the sensitivity of spread values during testing I can't quite get my head around what it is doing.
I created two custom symbols based on AUDCHF specification. I download bar data from tickstory and I modified the spread values in each file to 24 and 200, respectively. For each file I clicked the bar tab and imported M1 data, selecting the correct file for each symbol. The data appeared to import correctly see the screenshots below. I initially selected "Open Prices only" as the modelling method in strategy tester and ran the EA backtest for both custom symbols. The results were exactly the same. I repeated this for th 1 minute OHLC modelling and the two files with a factor of 10 difference in spread gave the same results. The question is Why? It doesn't seem right to me. Can anyone help?
The logs are pretty big (~110Mb). What's the best way of posting what you need?
The tester information are needed only. Not anything from your specific EA.
To avoid the trouble to cut the log in chunks, you could download the file in the cloud somewhere (Google drive or something similar) and give me the link. I will check.
In my quest to assess the sensitivity of spread values during testing I can't quite get my head around what it is doing.
I created two custom symbols based on AUDCHF specification. I download bar data from tickstory and I modified the spread values in each file to 24 and 200, respectively. For each file I clicked the bar tab and imported M1 data, selecting the correct file for each symbol. The data appeared to import correctly see the screenshots below. I initially selected "Open Prices only" as the modelling method in strategy tester and ran the EA backtest for both custom symbols. The results were exactly the same. I repeated this for th 1 minute OHLC modelling and the two files with a factor of 10 difference in spread gave the same results. The question is Why? It doesn't seem right to me. Can anyone help?
won't the tester alter the spread to the value chosen manually ? (unless real ticks are used) ?
won't the tester alter the spread to the value chosen manually ? (unless real ticks are used) ?
What value chosen manually ?
It's custom symbol, build from minutes data and test on Open Prices or 1 minute OHLC.
The tester information are needed only. Not anything from your specific EA.
To avoid the trouble to cut the log in chunks, you could download the file in the cloud somewhere (Google drive or something similar) and give me the link. I will check.
Sent you a link to the log files.
Many thanks
I tried to replicate it but i could not add bars to custom symbols via the manual method .
with the script method though the 200 spread is 200 and the 20 spread is 20
(i used the below to create 2 fakesymbols one with 200 one with 20)
#property version "1.00" input string name="FAKESYMBOL";//name to save input int manualSpread=200;//spread (-1 to use existing) //+------------------------------------------------------------------+ //| Expert initialization function | //+------------------------------------------------------------------+ int OnInit() { //--- if(_Period==PERIOD_M1){ EventSetMillisecondTimer(44); }else{ ChartSetSymbolPeriod(ChartID(),_Symbol,PERIOD_M1); } //--- return(INIT_SUCCEEDED); } void OnTimer(){ EventKillTimer(); bool loading=true; int load=0; while(loading) { Sleep(44); load=Bars(_Symbol,PERIOD_M1); if(load>10000){loading=false;} } //collect the rates MqlRates rates[]; load=10000; int copied=CopyRates(_Symbol,PERIOD_M1,0,load,rates); if(copied>0){ Print("Copied "+IntegerToString(copied)); Print("rate[0] "+TimeToString(rates[0].time,TIME_DATE|TIME_MINUTES|TIME_SECONDS)); Print("rate[last] "+TimeToString(rates[copied-1].time,TIME_DATE|TIME_MINUTES|TIME_SECONDS)); if(CustomSymbolCreate(name,NULL,_Symbol)){ if(manualSpread>-1){ for(int i=0;i<copied;i++){ rates[i].spread=manualSpread; } } datetime from=rates[0].time; datetime to=rates[copied-1].time; int ratesadded=CustomRatesReplace(name,from,to,rates,copied); Print("rates added "+IntegerToString(ratesadded)); }else{Print("Cannot create custom symbol");} }else{ Print("Cannot copy rates"); } Print("DONE"); ExpertRemove(); } //+------------------------------------------------------------------+ //| Expert deinitialization function | //+------------------------------------------------------------------+ void OnDeinit(const int reason) { //--- } //+------------------------------------------------------------------+ //| Expert tick function | //+------------------------------------------------------------------+ void OnTick() { //--- } //+------------------------------------------------------------------+
I tried to replicate it but i could not add bars to custom symbols via the manual method .
with the script method though the 200 spread is 200 and the 20 spread is 20
(i used the below to create 2 fakesymbols one with 200 one with 20)
Both tests use the spread specified in the rates yeah
I would suggest the op did a test with new symbol names as he may be stepping on another symbol possibly (i.e. new bars did not import or smth)
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
In my quest to assess the sensitivity of spread values during testing I can't quite get my head around what it is doing.
I created two custom symbols based on AUDCHF specification. I download bar data from tickstory and I modified the spread values in each file to 24 and 200, respectively. For each file I clicked the bar tab and imported M1 data, selecting the correct file for each symbol. The data appeared to import correctly see the screenshots below. I initially selected "Open Prices only" as the modelling method in strategy tester and ran the EA backtest for both custom symbols. The results were exactly the same. I repeated this for th 1 minute OHLC modelling and the two files with a factor of 10 difference in spread gave the same results. The question is Why? It doesn't seem right to me. Can anyone help?