Skip to content
All library documents

Estimating Intraday Option Open Interest for GEX Analysis

Article Quant Q&A · Author: skepticalforever
Read the original

Summary

Option open-interest figures published by financial websites are generally daily snapshots, which limits their direct use in intraday analysis. The post raises the question of whether intraday estimates or suitable vendor data are available, motivated by the desire to improve gamma exposure estimates during the trading day.

One proposed approach is to train a neural network, such as an LSTM or a conventional artificial neural network, on daily observations. The target would be next-day open interest by strike and days to expiration, with option prices, Greeks, bid and ask data, spreads, and underlying-market indicators as candidate features. The author presents this only as an idea and expresses uncertainty about its soundness. No vendor recommendations, model evaluation, data requirements, or evidence that the proposed features can predict intraday open interest are supplied.

Key ideas

  • Published option open interest is described as a daily snapshot, leaving intraday values uncertain.
  • Intraday open-interest estimates could support more timely gamma exposure analysis.
  • A proposed neural-network method predicts next-day open interest by strike and days to expiration from options and underlying-market features.
  • The suggestion is speculative and comes without validation or performance evidence.

Tags

OptionsMachine learningMarket microstructureVolatility
Full text
# Methods for tracking option open interest intraday


# Methods for tracking option open interest intraday












It is my understanding that open interest option values on financial websites are a reflection of a snapshot value each day. Is anyone aware of methods for estimating intraday open interest, or aware of any financial data vendors that offer their estimations on this?

## Answer by fede72bari (score 0)

https://quant.stackexchange.com/a/80876

I am looking for ideas for the same aim in order to make a better GEX estimation at the intraday level. A rough idea of mine is to train LSTM or ANN networks on a daily basis having as targets the OI for each strike and DTE released the day after. Using as features OHLC of the 1st and 2nd order greeks, bid, ask, bid/ask spreads, underlying daily indicators. Then, use it at the intraday level as an estimation. I don't know if it makes sense.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.