WorldQuant · Spring Alphathon

WorldQuant Spring Alphathon

Writing alphas in an expression language, before I could really code. It turned into a two-year job.

  • Expression language
  • WebSim
  • Alpha research
  • Quant finance
  • Gold, Level 2

I got to know of WorldQuant through the Wall Street Club at BITS Goa. The coordinators of the club had already taken part in WorldQuant's challenges, and some of them had been offered research consultant positions off the back of it. I was studying economics and computer science, so it seemed like a natural thing to try.

Wall Street Club, BITS Goa Wall Street Club, BITS Goa India's first student-run investment club, and the reason a computer science student ended up writing trading signals at all

Writing alphas without really writing code

At that point I wasn't fluent in programming, so I worked in the expression language instead. The idea was to build strategies (alphas) that could be traded in the markets to make money.

You had different universes to work in, defined by market capitalisation: US top 200, Europe top 500, Asia top 1000. Very few strategies worked across all of them, because liquidity behaves differently in each. You could do time-series ranking, or compare similar firms against each other, or anything else you could think of. What you did was up to your imagination.

The 70% rule

To submit, your alpha had to have less than 70% correlation with the existing pool. That rule exists so nobody just resubmits a signal that already exists, and so the pool the firm actually trades from stays diverse.

At the start it was hard to find anything with a low enough correlation. Gradually, as I spent more time in the system, I started to see what worked and what didn't. I submitted a few. But I still wasn't sure what was really happening underneath, or why.

The book

Trying to understand it properly, I found the book by WorldQuant's founder, Igor Tulchinsky: Finding Alphas. Reading it gave me a far better picture of what was actually going on, and of how to move forward.

Finding Alphas: A Quantitative Approach to Building Trading Strategies, edited by Igor Tulchinsky
The difference between submitting alphas and understanding why any of them worked

That is what got me through the Alphathon. I finished with a gold certificate at level 2, ranked 23rd internationally and 2nd in India, with a leaderboard score of 32,091.

WorldQuant WebSim dashboard showing Spring Alphathon leaderboard rank 23 and score 32,091
The WebSim dashboard: rank 23, score 32,091

What came after

I was offered a research consultant position, which I held for about two years. I left it in the end because of a conflict of interest; by then I was interning at another firm.

WorldQuant Challenge certificate: Gold Level 2, 2017 Spring Alphathon, awarded to Shubhransh Jagota

Once the consulting money started coming in, I stopped having to ask my parents for BITS's notorious fees.

Make it rain

It also gave me enough money to go and blow myself up in the options market. Sigh. But that's a story for another day.

I've answered a few questions about WorldQuant and getting started with alphas on Quora, if that's useful to anyone going down the same road.