Research That Questions Its Own Assumptions

We publish what we find, including the failures

Our quantitative research team publishes working papers, conference presentations, and model governance reports on the methodology behind our resilience scoring engine. We believe the best way to build trust in a predictive system is to expose it to scrutiny, document its limitations, and invite correction from peers in market microstructure and machine learning.

Read latest papers

A working paper on why quoted depth collapses faster than spread widens, and what order book features predict that collapse before it becomes visible to conventional metrics

Market Depth Fragility During High-Volatility Regimes

Conventional liquidity metrics focus on observable quantities: quoted spread, displayed depth at the touch, and time-weighted averages. These measures describe the present state of the order book but offer no insight into its structural stability under stress.

Our research identifies three precursor signatures that historically precede rapid depth erosion: asymmetric cancellation clustering on one side of the book, increasing adverse selection in trade flow, and widening imbalance between displayed depth and hidden liquidity patterns.

We trained a gradient-boosted ensemble on six years of tick data across equities and futures markets, withholding the final eighteen months for out-of-sample validation. The model achieved meaningful discrimination between instruments that would experience depth degradation and those that would maintain stability.

Importantly, we document the regimes where the model performs poorly: during scheduled macroeconomic announcements, when structural breaks occur, and in instruments with persistent spoofing activity that distorts normal microstructure patterns.
The full paper includes error analysis, feature importance rankings, and a discussion of why explainability matters more than marginal accuracy gains when deploying predictive models in regulated trading environments.

Latest Research and Commentary

March 15, 2026

Resilience Score Validation Report

Quarterly assessment of model performance against realized depth degradation events, including false positive rates and regime-specific accuracy breakdowns.

February 28, 2026

Asia-Pacific Market Structure Review

Analysis of order book fragility patterns across regional equity markets during the January volatility episode, with cross-market comparison.

January 20, 2026

Model Governance Update

Documentation of retraining triggers, feature drift monitoring, and committee review outcomes for the fourth quarter of 2025.
December 10, 2025

Conference Presentation at QuantMinds

Our lead researcher presented findings on explainable fragility scoring to an audience of quantitative analysts and risk officers in London.
November 5, 2025

Academic Partnership Publication

Joint paper with university researchers on the theoretical foundations of order book stability metrics accepted for peer review.

Upcoming Events

Webinars, roundtables, and conference appearances where our team discusses market structure resilience and AI applications in trading

Join our researchers for live discussions on resilience scoring methodology, model validation practices, and the evolving landscape of AI in market microstructure analysis.

virtual webinar on resilience scoring methodology
Webinar

Resilience Scoring in Practice

A technical walkthrough of how our models generate scores, how to interpret feature attributions, and how to validate signals against your own execution data.

market structure roundtable discussion in London
Roundtable

AI and Market Structure Roundtable

An moderated discussion with quantitative researchers from three asset managers on the practical challenges of deploying machine learning in execution workflows.