Background
Research hub

Market structure research and fragmentation analysis

We publish what we learn from building systems that reconstruct scattered liquidity. Our research spans venue behavior, model architecture, and the structural shifts reshaping modern markets.

Original research

Studies drawn from live platform data and model development, not recycled commentary.

Technical depth

Methodology disclosed, assumptions named, limitations acknowledged.

Practitioner focus

Written for people who navigate fragmented markets daily.

Latest announcements and updates

March 2026

Venue coverage expansion

We added eleven new alternative trading systems to our fragmentation monitoring pipeline, deepening visibility into European and Asia-Pacific dark pool dynamics.
January 2026

Model transparency release

Practitioners can now inspect the full attribution chain behind any fragmentation output, including venue contributions and confidence intervals.

November 2025

Conference presentation

Our research team presented findings on adaptive venue behavior detection at the annual Australasian Market Structure Symposium in Melbourne.
September 2025

Partnership announcement

We established a data collaboration with a regional exchange group to improve normalization of non-standard order book feeds.
July 2025

Platform architecture update

The reconstruction engine was rebuilt for sub-millisecond inference latency, enabling real-time fragmentation alerts during high-volume trading periods.

May 2025

Research publication

Our paper on inferring hidden liquidity from observable market fragments was accepted for peer review by a leading quantitative finance journal.

Electronification of fixed income

Fixed income markets are following equities down the path of fragmentation — once concentrated in dealer networks, liquidity now disperses across electronic platforms, request-for-quote systems, and all-to-all venues. The shift creates familiar problems: practitioners cannot see the whole picture because no single platform captures it. Our research examines how AI reconstruction techniques developed for equity fragmentation can be adapted to bond markets, where trade frequency is lower, instrument diversity is higher, and the definition of liquidity itself is more nuanced. We document the methodological challenges — sparse data, heterogeneous instruments, and the absence of centralized order books — and propose adaptive approaches that learn venue behavior from what is observable rather than requiring complete visibility. The work is ongoing, and we publish findings as they mature.

Practical guidance for fragmented markets

Best practices drawn from our research and from practitioners who use our platform daily. These are starting points, not prescriptions — every desk and instrument has its own rhythm.

Monitoring

Monitor venue dominance over time

A venue that handled thirty percent of your flow last quarter may handle eight percent today. Track dominance trends, not just point-in-time snapshots, to understand how your execution landscape is shifting.

Monitor venue dominance over time

A venue that handled thirty percent of your flow last quarter may handle eight percent today. Track dominance trends, not just point-in-time snapshots, to understand how your execution landscape is shifting.

Monitoring
Analysis

Distinguish structural breaks from noise

Venues change matching engines, fee schedules, and routing logic without announcement. Learn to recognize when a behavioral shift is structural rather than temporary — our anomaly flags help, but your market intuition is the final filter.

Distinguish structural breaks from noise

Venues change matching engines, fee schedules, and routing logic without announcement. Learn to recognize when a behavioral shift is structural rather than temporary — our anomaly flags help, but your market intuition is the final filter.

Analysis
Validation

Cross-reference fragmentation with your own fills

Platform outputs suggest where liquidity resides, but your execution history reveals where you actually accessed it. The gap between the two is often the most instructive signal.

Cross-reference fragmentation with your own fills

Platform outputs suggest where liquidity resides, but your execution history reveals where you actually accessed it. The gap between the two is often the most instructive signal.

Validation
Risk

Question model confidence explicitly

Every output carries an uncertainty estimate. High-confidence predictions deserve different responses than low-confidence ones. Train yourself to look at the confidence band before acting on the point estimate.

Question model confidence explicitly

Every output carries an uncertainty estimate. High-confidence predictions deserve different responses than low-confidence ones. Train yourself to look at the confidence band before acting on the point estimate.

Risk
Execution

Adapt your routing logic incrementally

Fragmentation insights should inform routing gradually, not trigger wholesale strategy changes overnight. Test small adjustments, measure fill quality, and expand what works.

Adapt your routing logic incrementally

Fragmentation insights should inform routing gradually, not trigger wholesale strategy changes overnight. Test small adjustments, measure fill quality, and expand what works.

Execution
Collaboration

Share observations with your research team

The practitioners who spot model outputs that feel wrong provide the most valuable feedback. Our platform is designed to learn from your experience — flag discrepancies and explain what you observed.

Share observations with your research team

The practitioners who spot model outputs that feel wrong provide the most valuable feedback. Our platform is designed to learn from your experience — flag discrepancies and explain what you observed.

Collaboration

Reference

Key terms in fragmentation analytics

Definitions for concepts we use throughout our research and platform. Understanding these terms helps you interpret outputs and engage with our methodology on its own terms.
Core

Liquidity fragmentation

The dispersion of trading interest across multiple venues, platforms, and mechanisms, such that no single market provides a complete picture of available supply or demand.

Venue

Dark pool

A private trading venue where orders are not displayed publicly before execution, designed to minimize market impact for large transactions.

Method

Order book reconstruction

The process of assembling a coherent view of market depth and interest from multiple fragmented sources with different formats, latencies, and update frequencies.

AI

Venue behavior model

A statistical or machine learning model trained to recognize the characteristic patterns of a specific trading venue, including its response to order flow and stress conditions.

Analysis

Structural break

A persistent shift in the statistical properties of a venue's behavior, often indicating a change in matching logic, fee structure, or participant composition.

Execution

Fill probability

The estimated likelihood that an order placed at a given price and size will be executed, accounting for current fragmentation and competing flow.

Metric

Fragmentation index

A quantitative measure of how dispersed liquidity is across venues for a given instrument, with higher values indicating greater scattering and lower concentration.

AI

Adaptive inference

A modeling approach that updates beliefs continuously as new data arrives, rather than relying on static parameters trained at a single point in time.

Venue

Crossing network

A venue designed to match buy and sell orders internally at prices derived from public markets, often with minimal or no displayed order book.

Market

Latency arbitrage

A structural phenomenon where participants with faster data access exploit price discrepancies that exist only for milliseconds across fragmented venues.