StrategyQuant X coensio workflow EURUSD H1 V1
NOTE: THIS IS NOT THE FINAL WORKFLOW ONLY AN EXAMPLE (TEMPLATE) WORKFLOW = A GOOD PLACE TO START...
This automated workflow generates and validates strategies on EURUSD H1 Timeframe, using Dukascopy tick data (UTC+02, DST No Weekends). Workflow is mainly based on free courses provided by quastic.com, which we fully automated using 'custom projects' feature of new SQ-X.
Note that strategy generation (first workflow block) and all validation tests (thereafter) are very time consuming. A good practice is to close all other applications on your PC and let it run until the workflow is finished. Note that the workflow is configured to remove all strategies that do not pass specified PASS filters/conditions at each workflow step. After strategy generation you will have >1000 strategies in the databank. This number will decrease systematically after each new workflow validation step.
The workflow consists of the following steps:
- Strategy generation step: in this step strategies are generated. At least 1000 needed to proceed to the next step.
- OOS1: This test will run all strategies on an Out-Of-Sample period (5 years of data 2005 to 2010)
- Slippage at 3: Test at higher slippage of 3pips.
- GBPUSD market: Test on a different market.
- A: TF_M30: Test on a lower TimeFrame.
B TF_H4: Test on a higher TimeFrame.
- MC random trades: Monte-Carlo randomized trades test.
- MC skipping trades: Monte-Carlo random skipping of trades test.
- MC random parameters: Monte-Carlo randomized strategy parameters test.
- MC random volatility ATR: Monte-Carlo randomized market volatility test.
- MC random slippage: Monte-Carlo randomized slippage test.
- MC random spread: Monte-Carlo randomized spread test.
- Last OOS2 test: Final test using 1 year 'unseen' market data. This test simulates forward-test results.
- Strategy selection: Selection of the top best strategies (highest Ret/DD)
- Portfolio correlation check: Removing strongly correlated strategies
- WFM: Walk-Forward matrix optimization and strategy selection
- Forward trading
To be clear, this automated stuff is also very fresh for me, none of this has been tested on a real account yet.
Any feedback, improvement suggestions are welcome.
Link to the workflow package:
UPDATE: See complete step by step installation guide HERE:
Note1: Some of the validation steps in this workflow are based on statistic/probabilities, if you eventually will filter out some good strategies, a good practice is to go back to test 1 (OOS1) and run those strategies through the whole process again.
Looks good - I'll go thru it.
very nice. I will post my workflow later. At the moment I have only a description in a text document. I will post this later. I tested so many workflows. No one works 100%. At all workflows some curvefitted strategies are left.
Hi Mr Coensio,
I just bought SQX and create first Strategy with all default from SQ ,but some reason I let it run couples hours ,nothing result ,I knew some thing from my setting .How can I have right setting or ranking .Thanks
Hi Coensio ,
This is Munchie again ,after I read all your workflow with audit one form SQ ,and redo ,it still came out no result ,I want to attach file but some reason ,can not .I would like you can help me to find out what is the problem with my first strategy.
oh i did attached the test ,and the other why it was no result .
This is indeed very strange and should not happen...I see you have a PC that is fast enough in order to generate at least 1 strategy / minute...
Ok I see your problem! You need to first download sufficient amount of tick data. Now you are trying to generate strategies only using few weeks of data, that is not enough for H1 timeframe.
Go to Data Manager and download all available tick data from SQ servers.
You also need to clone the data to your broker's timezone and correct it for DST shifts, but this is second step...
I just registered on this forum and I tried to use the wokflow you kindly mentioned above in a pdf file( https://www.coensio.com/forum/wp-content/uploads/wpforo/default_attachments/1549454920-CoensioWorkFlow-EURUSD-H1-V1-automated-workflow-for-SQ.pdf)
but when I tried to acces the settings for fast PC I gor an error: HTTP ERROR 500
I purchsed StrategyQuant several months ago but I have difficulties with generating strategies (My PC is I7 4790, 16GB RAM, SSD, so I believe it is not very bad)
Why cannot I acces that file settings? (are they for contributors?) They would help me a lot.
Not intersted in free strategies (this would be unfair and rude).
Just a sample work flow (the link in pdf file does not work).
Try to use those links:
and WorkFlow big:
First, I am not challenging anything. I am asking to learn. (seen how intentions can be misinterpreted online sometimes).
I reviewed the Ranking criteria of your work flow.
- Why is the Profit Factor IS 1.1 and OSS 1.3?
- Why is Ret/DD IS 3.9 and OSS 3?
- in the MC, why is there no ranking criteria for the MC data? How do you filter MC results
- Why is the 1 OSS1 test include 2010-2016 that was already tested in Build Strategies? Same with Final OSS? Why does it include 2005-2016?
I appreciate your guidance.
Yes you need to understand that this is workflow version 1 (currently I am working on my version 6), so yes there is room for improvement...
- Why is the Profit Factor IS 1.1 and OSS 1.3? -> You can see that OOS check is not checked (check is disabled), the whole data range is tested as IS in this case.
- Why is Ret/DD IS 3.9 and OSS 3? --> the same applies see which check is enabled and which is disabled. I tried multiple things and left only things that work for me enabled.
- in the MC, why is there no ranking criteria for the MC data? How do you filter MC results --> yes you are correct, you can disable/remove all checks under 'Ranking' tab and only use checks under 'Cross checks (robustness)' tab, under 'Filters'...
- Why is the 1 OSS1 test include 2010-2016 that was already tested in Build Strategies? Same with Final OSS? Why does it include 2005-2016? --> in the end it does not matter much, every type of strategy selection using any type of OOS will introduce a data mining bias which is never good...but we need to do something to select our strategies.
Feel free to create a copy of the workflow change it, improve it, test it and share it with us 😉
Thank you for the reply. I just wanted to learn from you, and I did. I did not realize some where checked and unchecked and that you had MC criteria in the filter tab.
I do appreciate you putting the workflow together and sharing it as a place for us to start and customize for ourselves. I know it is just a guide, and that is why I wanted to inquire about the details as maybe there is a lesson there.
I will start this week, and hopefully can start getting some strategies to post, and experience to further the discussion on this forum.