Finding hypothesis to work in your favor could be a struggle if you looking in wrong direction.
 

I have MS in Electronics (but I work as Data Scientist in an IT firm and have mad skills on AWS) and many folks over here from non-finance (chemistry, aeronautics, etc) background have found success in pushing their Alpha to the market. The folks from non-finance background are the reason why I keep pushing myself again and again on different hypothesis testing, etc. I barely spend anytime coding as Python is childs play for me and 2 years on QC have given me enough grip on how to use the framework/components - bummer that I haven't completed boot camp - I like to learn it the hard way and you guys always help via this forum. :)

I recently had luck stealing bits and pieces from Vladimir Peter Guenther Leandro Maia and others, and amalgamating with few concepts from my background in DSP (digital signal processing) and got very satisfying results for 2016-2021

PSR: 99%

Drawdown: 17%

Sharpe: 2.278

Beta: 0.028

CAGR: 49.3%

Net Profit: 710.36%

and yet I'm not happy with it because its not mine and I didnt work it from scratch and IN/OUT strategy could have strong bias on my algo - trust me when I say bias, I dont meant skewed distribution in Tech sector. My portfolio has about 37% in tech and is well distributed in cyclical, defensive and other sectors. By bias I mean use of GLD, SLV, XLI, XLU to signal 'stay away' from the market - such correlation may or may not hold true in the future considering current events of 2020. It could be good or bad but let that be for a seperate discussion.

For me, what makes me interested in qaunt trading is experimentation in pushing the limits on the metrics (drawdown, sharpe, PSR, beta) and over the time I have seen weird hypothesis holding out true in Competition/Market submissions, etc - It just fascinates me. I cannot stress enough but folks from non-quant background are my super heroes and true inspiration to hang around here and keep experimenting. We also can't ignore the fact that there is enough noise in the market to leverage it to our advantage - how and when is the question.

Sorry for the click bait but here is my golden question - What is your Alpha?
I'm more interested in learning what thought process you guys go through before pulling stats in jupyter notebook or start scribbling your main.py

Mine looks something like this:

1. Explore weird concepts from different background (for eg: Fourier Transformation)

2. Try it in research notebook with available libraries. Basically block testing before putting everything in one piece

3. Stitch everything in main.py and experiment on different universe and params.

4. Fail, Rinse and Repeat

What the actual crap goes in your head before you have that eureka moment that says 'Aaahha putting this in main.py will definetely work' and few param optimization closes the deal?

Do you guys pick concepts from old books? read lot of research papers? build lag-plots, correlations until you find a golden nugget? Is scikit-learn your friend? Analyse technical indicators (I personally feel they all are crap)? Recreate existing practices and add your own flavours (mean reversion, momentum, pair trading, etc)? 

I'm willing to colaborate with you'll brainy people as long as I learn the process - I promise you will benefit from my Python/QC framework skills, NodeJS and AWS if you want to deviate from QC and have a billion dollar idea.

Jared Broad Do you think we can have "Official" Slack or Discord enabled for the community? quick brain borrowing would be extremely easy and efficient. Thanks a lot for creating such an amazing platform. Happy to work for QC if the pay is decent (wink wink). Long Live QC !!

PS: Sorry for venting out my frustration. I know for sure its not a math problem. Its an ART and I want to learn it.