Holy Grail of the Strategy Tester

Back to topics list  | 1 2 To post a new topic, please log in or register
avatar
1
alharam88 2013.07.12 12:49 #
 
wackena:

This is just an example to demonstrate the ability of a MT4 Expert Advisor (non-scalping EA) to produce great results on the MT4 Strategy Tester. These results, with same MT4 EA and settings, are not producible on a forward Demo or Live Trading account.

 Why? Lack of Strategy Tester to have full tick simulation of Bar Bid/Ask INTERPOLATION. I personally believe that without this INTERPOLATION, majority of EAs with entry/exit points and Indicators based on Bid/Ask will not have Strategy Tester results match up with demo and Live trading. 

 

Strategy Tester Report
Bogie-StrategyTester_HolyGrail-1
(Build 445)

SymbolEURUSD (Euro vs. United States Dollar)
Period5 Minutes (M5) 2010.07.02 14:36 - 2012.11.13 15:08
ModelEvery tick (the most precise method based on all available least timeframes)
ParametersMarginControlOption=true; TradeOnNewBar=false; UseMM=true; Note="if UseMM=false, set Lots (lot size) to trade"; Lots=0.01; Risk=10; SLF=3; MA_Period2=2; MA_Shift2=0; Gp1=20; EATest=true; BoxSize=10; RecordData=false; AccountProtect="-- Protecting Account Profit ---"; ProfitProtect=false; ProfitPercentCloseAll=10; LossPercentCloseAll=1; MA_Period=5; MA_Shift=0; PipGap=10; MagicNumber=54321; Mail=false;
Bars in test3845Ticks modelled1283877Modelling quality90%
Mismatched charts errors
Initial deposit250.00
Total net profit20070855.39Gross profit24606303.78Gross loss-4535448.40
Profit factor5.43Expected payoff27013.26
Absolute drawdown10.38Maximal drawdown192700.00 (4.10%)Relative drawdown27.51% (96.00)
Total trades743Short positions (won %)365 (58.08%)Long positions (won %)378 (62.70%)
Profit trades (% of total)449 (60.43%)Loss trades (% of total)294 (39.57%)
Largestprofit trade346900.00loss trade-53342.25
Averageprofit trade54802.46loss trade-15426.70
Maximumconsecutive wins (profit in money)11 (41223.32)consecutive losses (loss in money)6 (-140742.25)
Maximalconsecutive profit (count of wins)861273.24 (5)consecutive loss (count of losses)-140742.25 (6)
Averageconsecutive wins3consecutive losses2


i think u must forward test .
avatar
1
RichardFX 2015.02.07 18:13 #
 
wackena:

This is just an example to demonstrate the ability of a MT4 Expert Advisor (non-scalping EA) to produce great results on the MT4 Strategy Tester. These results, with same MT4 EA and settings, are not producible on a forward Demo or Live Trading account.

 Why? Lack of Strategy Tester to have full tick simulation of Bar Bid/Ask INTERPOLATION. I personally believe that without this INTERPOLATION, majority of EAs with entry/exit points and Indicators based on Bid/Ask will not have Strategy Tester results match up with demo and Live trading. 

 

Strategy Tester Report
Bogie-StrategyTester_HolyGrail-1
(Build 445)

SymbolEURUSD (Euro vs. United States Dollar)
Period5 Minutes (M5) 2010.07.02 14:36 - 2012.11.13 15:08
ModelEvery tick (the most precise method based on all available least timeframes)
ParametersMarginControlOption=true; TradeOnNewBar=false; UseMM=true; Note="if UseMM=false, set Lots (lot size) to trade"; Lots=0.01; Risk=10; SLF=3; MA_Period2=2; MA_Shift2=0; Gp1=20; EATest=true; BoxSize=10; RecordData=false; AccountProtect="-- Protecting Account Profit ---"; ProfitProtect=false; ProfitPercentCloseAll=10; LossPercentCloseAll=1; MA_Period=5; MA_Shift=0; PipGap=10; MagicNumber=54321; Mail=false;
Bars in test3845Ticks modelled1283877Modelling quality90%
Mismatched charts errors
Initial deposit250.00
Total net profit20070855.39Gross profit24606303.78Gross loss-4535448.40
Profit factor5.43Expected payoff27013.26
Absolute drawdown10.38Maximal drawdown192700.00 (4.10%)Relative drawdown27.51% (96.00)
Total trades743Short positions (won %)365 (58.08%)Long positions (won %)378 (62.70%)
Profit trades (% of total)449 (60.43%)Loss trades (% of total)294 (39.57%)
Largestprofit trade346900.00loss trade-53342.25
Averageprofit trade54802.46loss trade-15426.70
Maximumconsecutive wins (profit in money)11 (41223.32)consecutive losses (loss in money)6 (-140742.25)
Maximalconsecutive profit (count of wins)861273.24 (5)consecutive loss (count of losses)-140742.25 (6)
Averageconsecutive wins3consecutive losses2


I have come here as I was curious about an idea that I was exploring myself... and I wondered why a good backtest won't 

necessarily forward test well... albeit my EA was more of a scalper that the original poster's EA.

This idea when backtested for 3 months, went from 10k to 18 million, and only stopped exponential growth 3 month into it as it hit MAXLOTS half way through the backtest....

Yet when I tried it FORWARD TESTING on a LIVE or DEMO account, it didn't perform at all like this.

I didn't mean to revive a 3 year old post however I wanted to add to the body of knowledge.

In this case, I had worked on an idea just as they added spread into backtester, and it was before I reaslied spread was in points... silly mistake, and a funny sidetrack... but interested results...

This idea does work until spread is 10 points (1 pip), however I looked into brokers who have zero spread and commissions only, or fixed spread, and the commission exceeded the profits, as this EA takes small profits...

My profit factor was only 617.80 and it was a scalper (and low modelling quality), however it was a 3 month test, not 3 year.

Left wondering if INTERPOLATION is the problem too.

So I guess this is another example of a holy grail backtest, that is not usable live :(

www.richardfx.com











Back to topics list   | 1 2  

To add comments, please log in or register