Information

Username: astevani
Member since: 11 Mar 2019
Last login: 11 Mar 2019
Status: Active

Activity

Where Created Comments
Algorithms 3 8
Forum Topics 2 3
Jobs 0 0

Last Algorithm Comments

AS
astevani · 5 years ago

The skip days are disabled .... so no day is skipped .... you can try on tick data.....

AS
astevani · 5 years ago

thanks

AS
astevani · 5 years ago

Hi Luigi,

 you can use 2 years o 3-4 year until ending 2018 for optimization and all the 2019 for back testing ... in this way some systems that i have optimized will win...

i think that 50 of commision and 2 pips of spread are too much... but the systems are working again....

AS
astevani · 5 years ago

during the optimization some days that have spikes, big spikes, let's the optimization process to learn where are the spikes because in that days you earn more. But that spikes don't return in the futures.....

so all parameters are calculates only to get the spikes... 

for a good optimization need clean data... the best is clean the data knowing the history of exceptional facts

 

AS
astevani · 5 years ago

            skipdate = new List<DateTime>();
            skipdate.Add(new DateTime(2017, 1, 11));
            skipdate.Add(new DateTime(2017, 6, 26));
            skipdate.Add(new DateTime(2017, 7, 19));
            skipdate.Add(new DateTime(2017, 8, 3));
            skipdate.Add(new DateTime(2017, 8, 24));
            skipdate.Add(new DateTime(2017, 9, 19));
            skipdate.Add(new DateTime(2017, 10, 25));
            skipdate.Add(new DateTime(2018, 1, 12));
            skipdate.Add(new DateTime(2018, 1, 25));
            skipdate.Add(new DateTime(2018, 2, 14));
            skipdate.Add(new DateTime(2018, 2, 15));
            //not over
            skipdate.Add(new DateTime(2018, 2, 16));
            skipdate.Add(new DateTime(2018, 6, 13));
            skipdate.Add(new DateTime(2019, 1, 2));

This are the days that i skip in the first version of robot for EURUSD.....

For the robot without without source code this can be a real problem because you dont know if you buy a robot with the pair history inside..... so in backtesting and optimization this robots can win many trade only because they have the history of the pair inside and when you use the trading system on real market the robot will fail....

AS
astevani · 5 years ago

In the first version of the robot, the robot win easy trade in that days. So I try to skip that days in the optimization to make a better optimization. In the source code, now, is disable. So all the days are used. If you want you can check that days and you will see that in that days there is strange volatility.

AS
astevani · 5 years ago

This optimization work no so bad :)

 

[ChartParameters]
Symbol = EURUSD
Timeframe = m1

[cBotParameters]
RsiPeriods = 9
AdxPeriods = 30
Quantity = 0.5
QuantityRatio = 1.3
StopLoss = 50
TakeProfit = 40
RsiBuy = 32
RsiBand = 10
RsiSell = 72
NormalPips = 2
ReversePips = -0.2
RatioBreakout = 1
RatioSLTP = -0.6
ReverseRatio = 1.5
ADXthreshold = 30
FromHour = 3
ToHour = 15
AdxSkip = 10
StopLossSpan = 15
MaxPositions = 4
RangeMin = 0.1
RangeMax = 0.7

AS
astevani · 5 years ago

Have you some ideas for an improvement?