Synopsis

In today’s market environment, underpinned by elevated rates, recession fears, fundamental technology changes, and geo-political crises, it is important to discover how option trade flow is structured and classified across market participants (Institutional, Market Makers, Retail, et al.) and strategy types (Momentum, Mean-Reversion, Tail Hedge, etc.,) to make informed trading decisions.

Our current study employs unsupervised learning methods, KMeans clustering, and UMAP
dimensionality reduction to illustrate how the SpiderRock Option Print Set – containing intraday traded and theoretical prices, volume, implied volatility, etc. – is used to reveal order segmentation across various data metrics and dimensions.

We focused on SPX and VIX 0DTE options, of particular interest post-COVID, during two
periods as they exhibited interesting market patterns due to the interplay between monetary policy and market dynamics. December 19th and 20th, 2023, were two days of extremes from market close to all-time highs to subsequent sharp selloffs, and Feb 15th, 2024, was a day with new highs following the previous CPI data-induced selloff.

This paper highlights key findings from our clustering analysis and showcases insights into the order flow derived from SpiderRock’s options data that can be used to enhance market strategies and market participant decision-making. Note, we illustrate the clustering approaches at a more descriptive level and try to point to directions that can lead to meaningful results.

Similarly, users can implement data from SpiderRock’s Option Features Dataset, which is a curated collection of options analytics, to leverage option market information for overall market behavior for additional studies.

For more information, please visit www.spiderrock.net/data, follow us on Twitter at @SpiderRockChi, and visit our LinkedIn Page. 

Get The Full

White Paper