I am currently building a workflow for successful forex strategies. Join me!
62 replies
AlgotradingDE
4 years ago #277576
I’ve been using StrategyQuant for more than a decade, but believe it or not, I haven’t even used its full capabilities until now.
My process is to create thousands of strategies in the builder and then run them through a very selective robustness check. The (few) surviving strategies are then activated on an MT4 demo account, where they must execute at least 25 trades before I consider them for use on a live account.
So far, I am very happy with this process and would like to automate more of the production of successful systems. That’s why I’ve been diving into the custom project features that can be used to create custom workflows.
While I am building some sample workflows for successful forex strategies, I would be happy if anyone on this forum is interested in the same and willing to share their experiences.
Especially, I would be interested to know if anyone has ever started such a project on their own?
Gerhard Frischholz
https://Algotrading.de
1
Massimo Scapini
3 years ago #280394
Hello,
I am using the attached workflow, which has a quite extensive set of tests. The final success is highly variable depending on the market (e.g cor A
It also used external config files that you can either change either manually or randomly with a very simple external
0
Massimo Scapini
3 years ago #280396
Hello,
I am using the attached workflow, which has a quite extensive set of tests. The final success is highly variable depending on the market (e.g for EURUSD just one workflow cycle can be enough, while for certain Indexes or currencies it can take tenths or even a hundred of cycles)
It also uses an external config file that you can change either manually or randomly with a very simple external .bat file, in order to change the building config at every building cycle.
For the strategies the pass the whole workflow, I follow this approach:
0
kurt
3 years ago #280552
Just a thought slighty off subject , but instead of allowing the app to run generating lots of random strategies with various long short entry crietera
Would it first not be a good idea to say like i am going to build breakout only stratgies, then i will go into the algowizard and construct a base to work off based around the breakout style
so for instance if we are building on XAUUSD , where it tends to be baised towards breakouts, and has big moves we could capture on say 30m, 1 hour TF
we are then using the software to generate thousands of combinations of a specfic trading strategies based around the style , for a specfic symbol
0
Matias Casas Lopez
2 years ago #286497
Hi everyone!
I really learn a lot from your comments!
I would attach my workflow for GBPJPY.
For now it has developed some good strategies, please feel free to comment or advise me for improvement’s!
It would be a good idea of having a group (maybe telegram or discord) where we can share our knowledge more easily and have a more fluent communications.
Thanks a lot.
PD: Sorry for my english, I’m Uruguayan.
0
Chad Lin
1 year ago #290050
Hi, I’m in.
Thank you for sharing. I downloaded your workflow and found it quite similar to mine. However, I’m currently facing extremely slow output speed, it takes me about 1 hour to generate a qualified strategy, which means I need 3000 hours to get 3000 strategies, and this is only during the Build phase. I believe the key issue lies in the Build rules. My approach is simple: random generation with all modules checked, but this might be too broad and is affecting the generation speed. I wonder if there are ways to improve this?
I switched to genetic algorithms, but the speed is still very slow. My workflow is almost the same as Conmariin’s, even the filter settings are similar. Presumably we both took the same course.
0
Massimo Scapini
1 year ago #290051
<p dir=”ltr”>Hello,</p>
<p dir=”ltr”>In my experience the generation speed is depending on several factors:</p>
<p dir=”ltr”>1. The asset traded
2. The template used
3. The method (genetic should be much faster than random)
4. The computer you use</p>
<p dir=”ltr”>1. Try with different instruments
2. I suggest to reduce the number of blocks and maybe use a loop in the workflow chosing between different config files with diffferent sets of blocks and/or different templates (e.g. see my comments to the other related question you posted)
3. Use genetic method with standard templates, random if you build yours and want to keep better control on what is generated.
4. The processing power is very important. I use a 16 core/24 thread PC with 64Gb RAM and I get an average of 40 filtered strategies in the build phase (from 400K up to 2M generated raw strategies per hour depending on 1.2 and the Timeframe)</p>
<p dir=”ltr”>Hope this helps</p>
<p dir=”ltr” style=”text-align: left;”>M</p>
1
Conmariin
1 year ago #290052
Yes, exactly. What Massimo has already said. I agree on all his points. And as an addition: I googled which indicators and signals are mainly used with Eurusd, for example. The more blocks you add, the slower it gets.
Automatisches Handeln mit Expert Advisor
https://www.rabenesche.de
Active also in:
https://forums.babypips.com/u/ravenash
0
Chad Lin
1 year ago #290064
Thank you for your reply!
I believe the reason for my slow strategy development lies in the selection of Blocks. In fact, the Build phase should be the most critical stage coz it determines the strategy’s quality. However, the challenge is that many people, including me, don’t know how to select the appropriate Blocks.
For example, I thought Bollinger Bands might be useful but also wanted to try MACD, so I just selected all Blocks as well as all take-profit and stop-loss options. I was afraid that it might be slower if I select less Blocks.
Which Blocks do you find more effective for EURUSD and XAUUSD?
0
Massimo Scapini
1 year ago #290065
It really depends on the kind of template you use.
For a reversal template all channels are good (Bollinger, Donkian, Keitner, …)
For a breakout template you could even skip all signals and select only indicators and levels, chosing those you like most.
Anyhow, I am currently using only self-built templates and not the standard ones, so my experience of best indicators to use for each instrument is not really extensive.
0
Conmariin
1 year ago #290069
Eurusd and Xauusd are the fastest pairs to get strategies/bots in short time. I googled signals and indicators discretionary trader would use. Works good for me.
Automatisches Handeln mit Expert Advisor
https://www.rabenesche.de
Active also in:
https://forums.babypips.com/u/ravenash
0
andres gil
1 year ago #290092
Hola a todos, quisiera preguntar cuándo se debe parar el work flow o si el para por si solo, ya que lleva más de dos días trabajando y lleva como 5 ciclos y no para de buscar estrategias.
0
Chad Lin
1 year ago #290113
Hi Conmariin,
My own EUR/USD workflow just finished running and did not generate any valid strategies yet, EUR and XAU seems difficult to me.
I’ll now test yours and share the results with you.
Thanks for the workflow you shared.
1
Chad Lin
1 year ago #290163
Hello everyone,
I use JForex as data engine instead of MT5, guess what?
WOW, it’s faster!
I got 1036 strategies in 10 mins, lol.
But I haven’t test them yet. I trade on MT5 so the data engine of Builder is JForex but remains MT5 of robust tests.
0
Chad Lin
1 year ago #290180
The strategies generated by JForex are not good. Although I got over 1,000 strategies in just 10 minutes, not a single one passed the OOS test.
0
Massimo Scapini
1 year ago #290194
Hello,
here is my latest workflow, together with the building blocks and the custom templates I use.
The workflow loops between different templates using external scripts, so if you want to use it without the scripts you must remove steps 1,2,4 and select the template you want to use in the Builder config.
The more effective Templates are the BOUTs. You can try them with Gold and EURUSD H1
Anyhow, if you are interested I can provide you with the scripts, too.
1