Topic: Best way to optimize M5 or M1 strategies

Hello guys,
I am trying to find the best way to optimize M5 or M1 strategies, only price opening, so that they are consistent.
Due to the large number of full history bars, this becomes a rather complex task.
Some alternatives:
1) Optimize for the entire period using genetic algorithm and few variations in each parameter, until filtering the main intervals of setups;
2) Divide the large period into several smaller periods and optimize each of them, taking the average of the optimizations to define the best setup;
Do you use any of the above methods? Any other suggestions?
Thank you in advance for your opinion.

Test Analyst of Softwares and Aplications.
Computer Science

Re: Best way to optimize M5 or M1 strategies

Can I ask. how many bars are you using for the whole test?

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Best way to optimize M5 or M1 strategies

After build good M5/M1 strategies here.
In Metatrader, year 2000 until now, to confirm the consistence and optimize.

Test Analyst of Softwares and Aplications.
Computer Science

Re: Best way to optimize M5 or M1 strategies

Can I ask how you can know that that huge amount of data is accurate?

Gaps and bad bars are going to hurt your results, how did you verify that data?

Can you not use somerthing like 100,000 bars or even less.

I feel that you will harm your purpose by using all of that data......

I suggest that you use less than 100,000 bars and then do a careful 'walk forward'

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Best way to optimize M5 or M1 strategies

Blaiserboy wrote:

Can I ask how you can know that that huge amount of data is accurate?

Gaps and bad bars are going to hurt your results, how did you verify that data?

In MT5, I'm using Metatrader Demo server with full data (Real Tick downloaded).
In MT4, I'm using Dukascopy server with full data too.
99% quality history in tests of both cases smile
I believe this data is reliable, no?


Blaiserboy wrote:

I suggest that you use less than 100,000 bars and then do a careful 'walk forward'

It's a good suggestion, I'll analyse and compare.

More suggestions?
I'm open for more suggestions. smile

Thanks a lot!

Test Analyst of Softwares and Aplications.
Computer Science

Re: Best way to optimize M5 or M1 strategies

If you are using Dukas data, that is not your broker data so the bars will probably be different.

Do you know that the MT5 data is the same as your broker MT4 data.?

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

7 (edited by srsantana86 2017-11-26 23:29:57)

Re: Best way to optimize M5 or M1 strategies

Blaiserboy wrote:

If you are using Dukas data, that is not your broker data so the bars will probably be different.

Do you know that the MT5 data is the same as your broker MT4 data.?

Dukascopy is my broker, using third part FTT server for MT4.
MT5 is only to approach the best setup more fast, because the parallel tests.
99% quality history isn't a reliable data?

Test Analyst of Softwares and Aplications.
Computer Science

Re: Best way to optimize M5 or M1 strategies

sounds like you have it all together

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....