EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3363. A Machine Learning Model for Options Traders that Predicts the Magnitude but not the Direction of Post-Earnings Stock Price Jumps
Invited abstract in session TC-63: Advanced Options Strategies Using O.R. and Machine Learning, stream OR in Banking, Finance and Insurance: New Tools for Risk Management.
Tuesday, 12:30-14:00Room: S14 (building: 101)
Authors (first author is the speaker)
| 1. | Ilkay Boduroglu
|
| R&D, Riskoptima Wealth Tech Corp. |
Abstract
US companies typically announce their earnings report after the stock market has closed. Their stock price usually experiences a significant move, up or down, the following day. Predicting the direction of this movement can be tricky, as the stock price may go down even after a good earnings report and vice versa.
As a result, market players focus on non-directional options strategies, such as Iron Condors, to capture a quick profit overnight. Even then, if the magnitude of the post-earnings price move (in either direction) is too large, this strategy ends up losing money. Therefore, it is crucial to predict the magnitude of the post-earnings price change successfully. A highly successful predictive model could make a massive difference in a hedge fund's performance.
This paper aims to develop a Machine Learning model that predicts the magnitude of the post-earnings stock price change. We first noticed that Estimize.com is an acclaimed data provider in the finance industry, and it publishes consensus predictions for earnings reports based on the predictions of its best competing analysts. Also, Thomson Reuters, one of the two major data providers in finance, publishes Estimize's consensus predictions the day before earnings releases. We then decided to record the variance of the most confident analysts' predictions at Estimize. We then analyzed the relationship between this variance and the magnitude of the post-earnings stock price change. We have promising results.
Keywords
- Financial Modelling
- Machine Learning
- Web-based Information Systems
Status: accepted
Back to the list of papers