Thoughts on the validity of building blocks and strategy criteria

Hi all,

Delighted to have discovered Strategyquant… although it has derailed me from learning Python as it has so much more than I could ever dream of creating.

I’ve immersed myself in lots of useful content, built a solid workflow and working towards building some strategies.

I have a question, that I suppose is subjective. I’ve tried a few methods, using a large number of building blocks vs. tailoring to just a couple based on previous strategies (e.g. just MACD and EMA).

Using a large number of building blocks seems to create “better” strategies but to some degree even with some parameter constraints the criteria seem quite random.

How do you guys go about it, do you just go with a large number of blocks and use more esoteric setups based on what it kicks out, or do you prefer a smaller set of blocks based on some underlying logic?

I imagine there is no right or wrong answers but it’s an interesting topic as Strategyquant has opened up so many opportunities (although as per above… maybe too many!).

Hi,

you can start with the simplest approach. Just explore pure randomness for various markets better for those well-know and explore as many block as possible. You will soon discover some interesting blocks. The next step you can do is to continue your exploration using templates https://strategyquant.com/?s=templates. We have a great tutorial on mean-reversion strategies https://strategyquant.com/blog/building-mean-reversion-strategies-using-templates-and-limit-orders/

Great. Thanks Tomas. That tutorial is super useful.

tomas262 wrote:

Hi, you can start with the simplest approach. Just explore pure randomness for various markets better for those well-know and explore as many block as possible. You will soon discover some interesting blocks. The next step you can do is to continue your exploration using templates https://strategyquant.com/?s=templates. We have a great tutorial on mean-reversion strategies https://strategyquant.com/blog/building-mean-reversion-strategies-using-templates-and-limit-orders/

Hi Tomas,

Exploring templates which is enabling me to fix some rules that I wanted to. The one thing I can’t work out… I’d like long and short symmetry for the entry and exit rules but I can’t see where/how I can set that on the AlgoWizard template builder (and obvs the strategy template overrides default settings in builder). How can I do this?

Hi,
you need to use ‘negated conditions’ to achieve long/short symmetry. Check the screenshot attached. Also make sure you set the builder correctly.

The builder -> What to build -> Trading directions -> Long / Short symmetry