Differences in Backtest Results
Forums › ProRealTime English forum › ProOrder support › Differences in Backtest Results
- This topic has 7 replies, 3 voices, and was last updated 3 years ago by Jiankyr82.
-
-
11/15/2018 at 11:48 AM #84909
Hi there,
Lately I am having serious doubts about the confidence of the backtest results in PRT. Mainly because I found significant differences in backtests that shouldn´t exist. For instance between real and demo results, in backtest and also in proorder trading. I understand that it´s normal to have some differences but I am talking about serious differences in different systems. Also when I share some systems with a friend that lives in a different timezone, even changing the hours accordingly the difference in results are too high.
Do you experience same kind of things? Because if I can´t belive the backtest results then my systems are worthless
11/15/2018 at 12:11 PM #84911For instance between real and demo results, in backtest
Do you mean differences between Backtest on your Live Platform and Backtest on your Demo Platform? Are both Platforms same settings, time zone and spread etc?
Why not post screen shots of the differences and (better still) post the code so one of us could also backtest on our Live and Demo Platforms and compare with yours?
Just ideas to move your issue forward a bit?
11/15/2018 at 12:48 PM #84918Hi Grahal,
Yes, differences between Live and Demo Platforms. I haven´t checked all my systems, but some of them have small differences and others biggers, with everything the same of course (system, period, spread…)
For instance these are the results in Live and Demo for the same system,timezoneperiod/spread…
This is the code and my timezone is UTC+00:00 Europe/London
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132///Sistema ME (Market Edge)//EURUSD 1H v2.2//Por TempusFugit//Ultima revisión: 14-11-2018////////////////////////////////////////////////// System parametersDEFPARAM CumulateOrders = FalseDEFPARAM PRELOADBARS = 2000//VARIABLESHL = 16 //Hora Entrada LargosHLS = 20 //Hora Salida LargosHS = 7 //Hora Entrada CortosHSS= 13//Hora Salida CortosLongStopPercentage = 0.65LongProfitPercentage = 1.1ShortStopPercentage = 1.9ShortProfitPercentage = 0.65RSL = 85//LIMITE PARA FILTRO RSI-MOMENTUM LARGOSRSS = 70//LIMITE PARA FILTRO RSI-MOMENTUM CORTOSm = 7 //PRIMERA CIFRA DEL MACD PARA FIJAR TENDENCIAXMACD = -0.25 //LIMITE VALOR MACD PARA FILTRAR TENDENCIA// GESTION POSICIONPOSITIONSIZEINCREASE = 1.4//MULTIPLICADOR PARA TODAS LAS POSICIONES.RSIMAXSIZE = 1.6//Max incremento posicion por RSI-MomentumSHORTINCREASE = 1//Multiplicador para posiciones cortasLIMITEPERDIDAS = 99999 //SI LLEGA A ESTA CANTIDAD DE PERDIDAS EL SISTEMA SE PARAREDUCIRPOSICION = 0.75 //MULTIPLICADOR DE LA POSICION SI EL BENEFICIO ESTA POR DEBAJO DE SU MEDIANMAPS = 500 //PERIODOS DE LA MEDIA PARA CALCULAR SI SE REDUCE POSICION. MAYOR O MENOR SEGUN LO ACTIVO QUE SEA EL SISTEMA Y DEL TIMEFRAMEPERCENTCROSSMA = 0 //PORCENTAJE QUE TIENE QUE SUPERAR EL SPROFIT A SU MEDIA PARA REDUCIR LA POSICIONIF STRATEGYPROFIT<-LIMITEPERDIDAS THENQUITENDIF//CALCULO SOBRECORMPRA/SOBREVENTAOVERBOUGHT = RSI[2](close)>RSLOVERSOLD = RSI[2](close)<100-RSS//CALCULO TENDENCIA ALCISTA/BAJISTAb = round(m*26/12)c = round(m*9/12)NOBEAR = MACDline[m,b,c](close)>XMACD/100NOBULL = MACDline[m,b,c](close)<-XMACD/100//////////////////POSITION SIZE\\\\\\\\\\\\\\\\\\\\\\//Position Size by WeekDay.Media de 1.5 por díaIF CurrentDayOfWeek=1 THENLPDAY=0SPDAY=1ELSIF CurrentDayOfWeek=2 THENLPDAY=1SPDAY=0ELSIF CurrentDayOfWeek=3 THENLPDAY=1.9SPDAY=1.9ELSIF CurrentDayOfWeek=4 THENLPDAY=1.8SPDAY=1.8ELSIF CurrentDayOfWeek=5 THENLPDAY=1.8SPDAY=1.8ENDIF//Position Size by RSI2XRSI = RSI[2](close)IF 100-XRSI>=80 THENLPSRSI = RSIMAXSIZEELSIF 100-XRSI>=70 THENLPSRSI = RSIMAXSIZE*0.75ELSIF 100-XRSI>=60 THENLPSRSI = RSIMAXSIZE*0.50ELSELPSRSI=0ENDIFIF XRSI>=80 THENSPSRSI = RSIMAXSIZEELSIF XRSI>=70 THENSPSRSI = RSIMAXSIZE*0.75ELSIF XRSI>=60 THENSPSRSI = RSIMAXSIZE*0.50ELSESPSRSI=0ENDIF//Cálculo de tamaño de posiciónsp = strategyprofitIF strategyprofit*(1+(PERCENTCROSSMA/100))<Average[NMAPS](sp) THENLPOSITIONSIZE = REDUCIRPOSICION*POSITIONSIZEINCREASE*(LPDAY+LPSRSI)SPOSITIONSIZE = SHORTINCREASE*REDUCIRPOSICION*POSITIONSIZEINCREASE*(SPDAY+SPSRSI)ELSELPOSITIONSIZE = POSITIONSIZEINCREASE*(LPDAY+LPSRSI)SPOSITIONSIZE = SHORTINCREASE*POSITIONSIZEINCREASE*(SPDAY+SPSRSI)ENDIF//////////////CONDICIONES ENTRADA/SALIDA/////////////LongEntry = Hour=HLLongEntry = LongEntry AND NOT OVERBOUGHTLongEntry = LongEntry AND NOBEARLongEntry = LongEntry AND DAYOFWEEK<>1//LongEntry = LongEntry AND close>99999999999ShortEntry = Hour=HS OR (BARINDEX-TRADEINDEX>4 AND Hour=HS+2)ShortEntry = ShortEntry AND NOT OVERSOLDShortEntry = ShortEntry AND NOBULLShortEntry = ShortEntry AND DAYOFWEEK<>2ShortEntry = ShortEntry AND DAY<>1 AND DAY<>2 AND DAY<>24 AND DAY<>29 AND DAY<>31//ShortEntry = ShortEntry AND close>99999999999IF LongEntry THENBUY LPOSITIONSIZE CONTRACTS AT close LIMITSET STOP LOSS LongStopPercentage*CLOSE/100SET TARGET PROFIT LongProfitPercentage*CLOSE/100ENDIFIf LongOnMarket AND Hour>=HLS THENSELL AT close LIMITENDIFIF ShortEntry THENSELLSHORT SPOSITIONSIZE CONTRACTS AT close LIMITSET STOP LOSS ShortStopPercentage*CLOSE/100SET TARGET PROFIT ShortProfitPercentage*CLOSE/100ENDIFIF ShortOnMarket AND Hour>=HSS THENEXITSHORT AT close LIMITENDIF11/15/2018 at 1:01 PM #84922I forgot… for EUR/USD 1h, spread 0,9
11/15/2018 at 9:01 PM #84953On my Demo Platform on EURUSD 1H with spread = 0.9 100k bars I get results attached … number of trades < half what you get!
The plot thickens, I am thinking what might be the reason.
I cannot test on Live sorry as I have my Live Platform disabled.
11/20/2018 at 4:34 PM #85342Because of these inconsistencies now I am running some systems both in Demo and Live at the same time and as expected there are some differences in the results of the trades (no exactly the same prices of entry/exit –> normal/expected) but also a few trades that doesn´t match between Live/Demo (WHY???). Very few days running both systems so I can tell yet how often or if it occurs only in some systems but I can tell that it happens.
Am I the only one with these inconsistencies?
11/20/2018 at 5:33 PM #85345Am I the only one with these inconsistencies?
I doubt it very much, but others are not checking the same inconsistency that you are?
We are all trying to deal with other inconsistencies … roll on Version 11.3 and this will be all behind us (I bet NOT!? 🙁 )
09/01/2020 at 10:07 PM #143155 -
AuthorPosts
Find exclusive trading pro-tools on