Independent quantitative research

Efficiency is built.
We study
its limits.

We investigate how capital, competition and implementation constraints shape financial markets—and the scale at which an opportunity can exist.

Explore our research Our approach

FIELD NOTES / 01

The geography of market efficiency

INTERACTIVE CONCEPTUAL MODEL · SYNTHETIC INPUTS

The geography of efficiency

A price gap travels through routes with different depth, delay and cost. Change the assumptions to see how gross discrepancy and implementation frictions move together.

More connected routeConstrained routeConceptual flow

A highly connected reference market in this hypothetical network. 3/6 connected routes · relative flow index 0.41

Network topology and flow are authored scenarios. They are not exchange quotes, measured relationships or proprietary signals.

Trade size $25,000Illustrative notional; USDLarger size consumes a greater share of assumed available depth.Round-trip fees 8.0 bpBoth legs combined; basis pointsA user-selected scenario input, not a venue fee schedule.Fragmentation 58%0 connected ↔ 100 fragmentedHigher fragmentation widens the illustrative gap while reducing routes and depth and increasing latency.Market activity 54%0 quiet ↔ 100 activeThis scenario assumes activity raises the gap and depth, with more congestion and latency.

Illustrative gross gap19.7 bp

Assumed total friction21.0 bp

Net feasibility estimate-1.3 bp

Assumed effective depth$111,469

Modeled transmission delay126 ms

Model assumptions and size sensitivity Illustrative only · no observed quotes or capacity

Net estimate = illustrative gross gap − round-trip fees − modeled impact − modeled latency loss. The gross gap is an exogenous scenario input, not a price observation. This deterministic teaching scenario is not risk-adjusted P&L. Financing, credit, inventory, funding, settlement, queue uncertainty and endogenous price impact are excluded.

The artificial equations set gross gap to 10 + 12 × fragmentation + 5 × activity basis points; effective depth to $180,000 × (1 − 0.58 × fragmentation) × (0.62 + 0.58 × activity); impact to 0.35 × (size ÷ effective depth)^0.72 × 100 basis points; delay to 24 + 115 × fragmentation + 65 × activity milliseconds; and latency loss to (delay ÷ 100) × (0.65 + 0.35 × activity) basis points. Fragmentation and activity are scaled from 0 to 1. Connected routes are max(1, round(6 − 4.5 × fragmentation)); the relative flow index is max(0.08, (1 − 0.72 × fragmentation) × (0.35 + 0.65 × activity)). These choices are pedagogical, not fitted to market data.

Same assumptions, different sizes

Model output · USD notional and basis points

Gross gap, fees, impact, latency loss and net feasibility for six illustrative trade sizes under the selected scenario.
Size	Gross	Fees	Impact	Latency	Net estimate
$5,000	19.7 bp	8.0 bp	3.7 bp	1.1 bp	6.9 bp
$15,000	19.7 bp	8.0 bp	8.3 bp	1.1 bp	2.3 bp
$30,000	19.7 bp	8.0 bp	13.6 bp	1.1 bp	-3.0 bp
$60,000	19.7 bp	8.0 bp	22.4 bp	1.1 bp	-11.8 bp
$120,000	19.7 bp	8.0 bp	36.9 bp	1.1 bp	-26.3 bp
$200,000	19.7 bp	8.0 bp	53.3 bp	1.1 bp	-42.7 bp

01 / Our premise

Capital connects prices.
Constraints shape
the route.

Markets become efficient through work: discovering information, bearing risk, financing positions and executing trades.

That work consumes resources. We study where its constraints change price formation—and whether a discrepancy survives the costs of acting on it.

The size of a market is a starting question, never a verdict on its efficiency.

02 / What we investigateSix lenses. One connected question.

A map of the mechanisms.

Research program

Efficiency & capacity

How does feasible scale change a discrepancy?

Capital occupancy

Cross-market price discovery

Which venue informs the others?

Information transmission

Liquidity & microstructure

What does a received quote change reveal?

Adverse selection

Derivatives & financing

What keeps capital from converging prices?

Fragmented collateral

Nonlinear market dynamics

When does the same mechanism change regime?

Liquidity feedback

Computational discovery

Can AI improve verified research throughput?

Controlled hypothesis search

03 / From pattern to opportunity

A visible difference
is only the beginning.

A RESEARCH STANDARD

From pattern to bounded claim

Each stage asks for stronger evidence. A result advances only when the next economic question has an answer.

01 / 05

What happened, where, and on which clock?

Method

Define the instrument universe, venue, event time, receipt time and eligible observations before measuring a pattern. Preserve missing, stale and rejected records in the accounting.

What could falsify it

The pattern disappears under valid timestamp, instrument-mapping or data-quality checks.

Reproducible measurement; no claim of prediction yet.

This is a methodological framework, not a report of five completed Side Street studies.

04 / Scientific directionResearch programs · proposed

Three frontiers.
A common economic test.

01 /

The Efficiency–Capacity Frontier

How much capital can an opportunity accommodate?

Explore the question 02 /

The Anatomy of Price Discovery

Where does information enter, and how does it travel?

Explore the question 03 /

The Economics of Execution

Where does a forecast lose its economic value?

Explore the question

05 / How we work

Make the question falsifiable.

Inside the approach

01Hypothesis
02Measurement
03Controls
04Replication
05Economics
06Decision

A negative result can be the right result. We retain failures, challenge complex models with simple controls, and keep statistical evidence separate from execution value.

06 / Selected researchDocumented work