Quant student is inspired to tackle academic research

Zaid Annigeri, a Master of Quantitative Finance student, wanted to know more about the models that are used to develop strategies for pricing options, so he set out to test them. 

Using several years of options market data, he performed an empirical comparison on two models that are heavily used to predict stock movements. He presented his research at the Future Alpha 2026 Conference in Brooklyn. His full paper is available here.

“In industry, if you question a method that everyone uses, you are usually seen as difficult not curious,” he said. “In research, you’re allowed to ask does this actually work and report whatever the data says, even when the answer may not be what people want to hear. 

Annigeri, a former aeronautical engineer, begins his second year of the MQF Program in September. He answered some questions about his research and why he decided to do it. 

Q: Why are you interested in this topic and what motivated the research?
A: “Options are contracts people use to bet on or hedge against future price moves. Think of them as insurance policies on stocks. Everyone uses them, from Wall Street desks to pension funds. To price one correctly, you need a picture of how much the market expects a stock to move at every future date and every possible target price. That picture is called a volatility surface. Traders read it the same way a forecaster reads a weather map. The whole industry treats this map as the right answer. I got interested because that landscape is surprisingly hard to draw. The market only gives you a few points, and you have to sketch a smooth surface through them. If you sketch it carelessly, the shape can accidentally imply trades that shouldn't exist in real markets, like a portfolio that costs nothing to build but is guaranteed to make money. I wanted to work on the version of this problem where the landscape is never allowed to imply the impossible.”

Q: How did you do the research?
A: “I worked with several years of real options market data. On any given day, the prices of all the options on a stock trace out a shape called a volatility surface. My first step was to redraw that surface cleanly using the SVI model, but with a strict rule built in: It was never allowed to imply a free lunch, or an arbitrage. With that clean surface in hand, I tested whether it actually helped in practice. I built a simulation where I acted as a trader hedging each option day by day, once using the sophisticated surface (built with the SVI model, short for Stochastic Volatility Inspired) and once using a simpler, older method based on the stock's recent movement. For every option I tracked how much money was left unexplained at the end of the day, which is the hedging error, and I ran that comparison across both calm and turbulent markets to see exactly when the extra rigor paid off and when it did not. Statistical tests then confirmed the differences were real and not just luck, and that is how the pattern emerged, with the surface earning its keep for some options and the simpler method holding its own for others.

Q: Were you surprised by what you discovered?
A: “Honestly, yes. I expected the sophisticated tool to reliably beat the simpler one, and what I found was more nuanced. It wins some of the time, loses some of the time. The surprising part was not that it lost sometimes, but that the losses had a pattern. For options with prices close to where the stock is currently trading, the sophisticated tool earns its keep. For options with prices far from where the stock is trading, a much older and simpler method does just as well or better. That split was not obvious going in, and it reframed the problem for me as less academic and more like fixing a small bug that leaks money every day.”

Q: How will this research impact the industry and practitioners?
A: “There are three distinct ways: Options trading desks get an infrastructure piece that catches a class of hidden pricing bugs before those bugs cost money on real trades. Risk and model validation teams get a clean reference they can compare their firm's internal tools against, to catch drift and mistakes. And finally, anyone building signals or strategies on top of options data (which is where a lot of new quant research is heading) gets cleaner inputs.”