Better strategies, better models, and of course, better profits.
Exactly 1 year ago, I ventured out to fill the gap between useful, entertaining quantitative finance content and those passionate about the beautiful, dynamic field.
What originally started as a passion project has led to discovery of proprietary models and strategies that I can confidently say are exclusive to us.
So, today, we’ll be going over some of our latest and most significant developments.
- Update: All Medium posts are now un-paywalled and fully accessible.
The goal of this experiment was simple and noble: predict the overnight direction of the S&P 500. We wanted to use machine learning to help us, but first, we needed to find useful features that would be helpful in prediction.
To do this, we got some help from the paper: The Role of Leveraged ETFs and Option Market Imbalances on End-of-Day Price Dynamics. In sum, the paper theorizes that the net position of market makers and dealers had a significant and capturable effect on market prices. The paper focuses on the end-of-day effect, but we wanted to see if it could be exploited on an overnight basis.
Doing this meant that we needed to build a full model of the entire option market. Luckily, we had quite an interesting tool:
By the end, we were able to build this into both a powerful model and a surprisingly profitable strategy:
Remember Brownfield Fund from “The Big Short”?
Yeah, those guys.
Before they made their big trade in the 2008 financial crisis, they had claimed to grow their fund from $100,000 to $30,000,000 through options trading. Naturally, we needed to both verify that the claim was true, and after, see if we could replicate the strategy for ourselves.
“People hate to think about bad things happening, so they underestimate their likelihood”
One of their most interesting early trades was back in 2002, during a period of volatility for Capital One. In sum, there were fears and investigations surrounding Capital One’s health, and accordingly, the stock price traded at an extremely low price:
Despite the fears, Capital One continued to post record earnings in the following quarter, but the stock price did not move budge. This led the two to believe there was one of 2 outcomes:
- The company was indeed fraudulent, the earnings were false, and in the coming months, the investigations would deliver a terrible verdict to reflect that.
- The company’s earnings were true and it was in a significantly better position than the stock price and pending investigations would indicate.
Because of this, the pair theorized that after the investigations, the stock price would either be worthless, or at least double. This meant that the options market needed to price-in a bi-modal distribution:
In reality, options generally price in a normal distribution where out-of-the-money strikes are almost always sold extremely cheap. Noticing this, the pair bought 8,000 LEAP contracts for a total of $26,000. After the investigations vindicated the firm, the share price returned to its appropriate price and the position increased to more than $500,000.
As fascinating as this early trade was, it is just one of many we covered — furthermore, there are many trades just like this available even up to now.
While we’ve done well in traditional option markets, they aren’t the only game around. Prediction markets like Kalshi and PolyMarket offer event contracts that are similar to options, and some of then are even tied to indices like the S&P 500:
We theorized that taking a wager of the S&P 500 closing price within a specified price band was akin to shorting volatility — fortunately, we developed an extensive framework for that exact style of trading: Building a Volatility Trading Empire.
This is actually a strategy currently in production, so we won’t delve too deeply here, but by extending our volatility trading framework to these markets, we were able to pull out an actually solid edge:
So yeah, we’ve been up to quite a bit of mischief.
It’s been an exciting journey learning about new corners of the market, as well as getting better at the tools (ML, Python) needed to really squeeze out an edge. We’ve come this far in just year 1 — and we’re only just getting started.
Many of our strategies are still in production, so even if you’re looking for something more hands-on, we’ve got you covered. So, if you’re passionate about quant finance, there’s only one playbook to follow — The Quant’s Playbook.
See you there & happy trading! 😄