Hampson Strategies — Capability Brief

CDLA Research Architecture

Constraint-Deliverability Lag Alpha — Historically Audited Systematic Research

Andrew C. Hampson II·hscai.org·@drampson11·May 2026

“Standard systematic strategies find stress. CDLA finds what stress has not yet priced. The architecture is complete, frozen, and audited. The track record exists. This brief is an invitation to discuss the institutional context in which it is applied.”

01

What Conventional Approaches Miss

02

The CDLA Architecture — Four Gates

03

Carrier Cartography

04

The Audit Result

05

The Research Product

06

Claim Boundary & Proposition

The Problem

The Wrong Object

Systematic strategies test whether stress moved. CDLA tests whether the market has mapped that stress to the correct outlet.

Systematic research strategies are built on a foundational assumption that is rarely examined: that the alpha object is stress itself. When a constraint tightens, the model buys the asset that historically responded to that constraint. When volatility rises, the model buys defensive exposure. When spreads widen, the model shorts credit.

This is the wrong object.

The asset that responds historically to a constraint category is not the same as the asset the market is currently mispricing. The market prices the visible variable — crude oil when Hormuz closes, generic high-yield when credit spreads widen, broad utilities when power demand rises. It does not price the specific downstream outlet where the structural transmission actually clears. That lag — between when constraint pressure changes and when the market recognizes the correct expression — is where systematic alpha lives.

The framework was built to test the correct object: not whether the constraint moved, but whether the market has mapped that constraint to the wrong outlet while the correct outlet remains underpriced.

Section 01

What Conventional Approaches Miss

Three failure modes. Each one producing the same result — a stress thermometer rather than an alpha engine.

Standard stress-factor models test whether a price-based constraint proxy moved above a threshold. They do not test whether the market has recognized the correct downstream outlet. The result is strategies that detect headlines rather than the transmission lag that produces alpha.

Failure Mode 1 — Pressure-Only Automation

Reducing the signal to generic movement detection. The instrument that represents stress in the aggregate — a credit spread, a volatility index, a commodity price — is not the instrument that carries the unpriced residual. Collapsing the architecture to "stress rose, buy defensive exposure" produces a stress thermometer. A thermometer can be accurate and still not make money.

Failure Mode 2 — Instrument Contamination

Many instruments that have genuine exposure to a constraint change also carry unrelated beta — sector correlation, currency exposure, duration, or credit effects that dominate the deliverability signal. A model that does not filter for instrument cleanliness is testing the wrong relationship even when the underlying thesis is correct.

Failure Mode 3 — Late Entry on Constraint Movement

Alpha does not exist when the market has already repriced the correct outlet. It exists in the window between when constraint pressure changes and when consensus recognizes which instrument carries the repricing. Standard models enter on constraint movement. The edge is in entering before the expression moves.

Section 02

The CDLA Architecture

Four gates. Sequential. A signal that does not pass all four does not generate a position.

CDLA — Constraint-Deliverability Lag Alpha — maps four layers sequentially. Each layer is a gate. The system is sparse by design. Sparse signals that meet all four gates outperform dense signals that meet only one.

Gate 01

Source Event

Did the requirement change?

A measurable event changes required demand, access, supply quality, processing capacity, settlement architecture, or permission rails. The source event is not the price move. It is the structural change that precedes the price move.

Source events span: AI infrastructure capex cycles, geopolitical energy disruptions, regulatory reclassification, credit compression regimes, dollar funding stress episodes, and deliverability architecture shifts across commodity and financial markets.

Architecture Relevance

A signal that begins with price movement is entering after the source event has already partially transmitted. CDLA begins upstream — with the structural change that generates the downstream pressure chain.

Gate 02

Deliverability Constraint

Who can actually deliver the required output under current quality, accessibility, capacity, and settlement conditions simultaneously?

This is where most systematic frameworks stop at the wrong answer. The largest resource is not the best deliverable if it lacks the processing step. The geographically optimal route is not the best path if the insurance architecture is broken. The asset with the highest nominal capacity is not the best expression if the settlement channel requires the wrong currency.

CDLA evaluates deliverability across: quality and specification match, geographic and logistical accessibility, processing and refining capacity, export permission and political alignment, settlement and currency fit, and capacity scaling within the relevant lag window.

Architecture Relevance

The winner is the expression with the most direct deliverability to the changed requirement — not the one most correlated with the constraint category historically.

Gate 03

Underresponse Residual

Has the market repriced the correct outlet?

The system looks for the lag between changed requirement and market recognition. When the correct outlet has already repriced, there is no signal. The edge is in the window when the constraint has moved and the outlet has not.

This is what separates CDLA from a factor model. A factor model asks: did this variable predict this return historically? CDLA asks: has this specific outlet been identified and repriced by the market relative to the constraint change that just occurred? Those are different questions with different answers.

Architecture Relevance

The signal exists only in the recognition window. Systematic strategies that enter on constraint movement alone enter late. CDLA enters on the gap between constraint and recognition.

Gate 04

Expression Cleanliness

Can the trade express the deliverability residual without contamination?

Two instruments may both carry genuine exposure to the same deliverability gap. One captures it cleanly. The other imports unrelated beta — currency effects, sector correlation, duration, credit spread — that dominates the deliverability signal in the historical data and in live execution.

The CDLA architecture ranks candidate instruments by residual size relative to contamination. Instruments that fail the cleanliness filter do not generate positions regardless of signal strength in the upstream layers.

Architecture Relevance

The highest-Sharpe expressions are those where the deliverability residual is large and the contaminating beta is minimal. That intersection is rare, which is why the system fires sparsely. Sparse signals that meet all four gates outperform dense signals that meet only one.

Section 03

Carrier Cartography

Some regimes transmit through higher-order carriers before direct deliverability expressions reprice. The map identifies them without contaminating what works.

In credit compression environments, duration-quality spreads carry the pressure before specific deliverability instruments move. In dollar-weakness regimes with sparse signal density, the denominator relief carries the expression before constraint selection can operate cleanly.

The M22 carrier cartography layer maps these higher-order pressure paths across three identified regime types. The M23 surgical overlay applies the carrier map only to the three identified scar years — 2015, 2017, and 2018 — leaving all other years byte-for-byte unchanged.

YearMissing MechanicCarrier Map CorrectionNon-Scar Contamination
2015Balance-sheet defensiveness precedes deliverability repricingDURATION_QUALITY_CARRIER — draw quality/duration carrier before CDLA expressionZERO
2017Denominator relief precedes constraint selection when CDLA density is sparseDOLLAR_DENOMINATOR_CARRIER — draw denominator object before sleeve mapZERO
2018Credit compression distorts weekly CDLA while recognition lag remains validCREDIT_COMPRESSION_RECOGNITION_CARRIER — smoothed recognition-lag distributionZERO

Zero non-scar changed weeks across all overlay variants. The carrier map repairs identified failures without retrofitting working years. That is the non-contamination proof.

Section 04

The Audit Result

358 active weeks. 14 calendar years. Five structurally different macro regimes. Zero hard look-ahead violations.

3.83

BASE_K Sharpe

Return 771.0% · Max Drawdown -9.5% · Full active sample

4.29

M23 Family Conservative

Return 897.6% · 1.59x internal hurdle

4.39

M23 Top High-Resolution

Return 870.0% · 1.63x internal hurdle

97.4%

Bootstrap P(Sharpe > 2.7)

5,000 iterations · 29 active clusters · Internal hurdle 2.7

358

Active Weeks · 2013–2026

29 active clusters · Scar active weeks 57

0

Non-Scar Changed Weeks

Byte-for-byte invariance across all M23 surgical variants

Year-by-Year Consistency

YearWeeksBASE_KM23 FamilyM23 TopReturnNote
20131612.0612.0612.0615.4%Unchanged
2014263.883.883.8812.0%Unchanged
201533-0.232.052.088–11%Carrier Repair
2016174.844.844.847.6%Unchanged
20178-0.282.964.503–5%Carrier Repair
2018161.101.743.902–5%Carrier Repair
20191810.8310.8310.834.2%Unchanged
20203711.5511.5511.5552.2%Unchanged
2021427.217.217.2139.8%Unchanged
2022390.850.850.856.6%Unchanged
2023338.078.078.0757.1%Unchanged
2024473.103.103.1029.8%Unchanged
2025184.844.844.8416.4%Unchanged
2026817.4117.4117.417.8%Unchanged

M28 / M28B Temporal & Instrument Provenance Audit

Pre-Inception Exposure Flags

0

No position entered before instrument inception date across 768 ledger rows

Feature Windows After Signal Date

0

No feature construction window extends beyond the declared signal date

Hard Look-Ahead Violations

0

Zero unresolved ledger-level look-ahead violations across 1,536 price and provenance checks

Selector Training Cutoffs

0

All selector-training cutoffs precede the evaluated position week

Carrier-Label Feature Dates

0

Carrier-label feature dates do not extend beyond evaluated position weeks

Combined Audit Status

PASS

M28 PASS_WITH_WARNINGS (18 warn, 0 fail) · M28B PASS (0 warn, 0 fail)

Cost & Slippage Stress

0 BPS

4.29

10 BPS

3.63

25 BPS

2.64

50 BPS

0.99

75 BPS

-0.66

Implementation Warning

Severe active-week drag is the primary implementation constraint. At 25 basis points per active week the conservative family holds above the 2.7 internal hurdle at 2.64. At 50 basis points it drops below. Instrument liquidity at institutional sizing is the remaining open variable. This is not a reason to dismiss the result — it is the correct framing of what forward paper-trading must validate.

Section 05

The Research Product

Three deliverables. One diligence package. One live signal layer running in parallel.

Deliverable 01

Frozen Architecture for Independent Review

The CDLA architecture is complete and frozen pending forward paper-trading validation. The institutional report, year-by-year matrix, non-scar invariance proof, carrier cartography, cost stress grid, bootstrap distribution, and M28/M28B audit receipt are available as a complete diligence package for institutional review.

Deliverable 02

Forward Validation

Forward paper-trading is the remaining gate before autonomous live deployment. The forward-specifiable carrier labels — identifying which higher-order carrier dominates in each observable market configuration — are documented and testable without requiring access to the executable signal logic. Each new signal will be classified at entry with frozen fields: carrier label, expected outlet, entry date, exit window, recognition trigger, and invalidation condition.

Deliverable 03

Live Signal Layer — CGAM Archive

The CGAM human-led constraint intelligence layer produces published, timestamped market notes across all seven constraint domains on an ongoing basis. The full archive demonstrates real-time alpha identification in public, with mechanism and invalidation conditions stated at publication. The validated note event book returned 143.90% compounded across nine events against 38.64% for matched SPY windows and -2.81% for a generic stress proxy across the same periods.

That archive functions as a live forward-validation record running in parallel with the quantitative architecture. It is the human layer that CDLA is designed to systematize.

Institutional Context

The full research archive is available at hscai.org. Every call is published with its mechanism and invalidation conditions. The misses are visible alongside the hits. That is the architecture of a research track record rather than a marketing document.

Section 06

The Claim Boundary

Three claims stated. One explicitly not stated. The distinction matters.

✓ Claim 01 — Architecturally Complete and Historically Audited

The system passed architecture freeze, invariance proof, cost stress testing, active-cluster bootstrap, and forward-label specification at the ledger level. M28/M28B found zero hard look-ahead violations across 768 ledger rows and 1,536 price and provenance checks.

✓ Claim 02 — Scar Years Explained by Carrier Cartography

2015, 2017, and 2018 improve through higher-order pressure carriers without contaminating working years. Zero non-scar changed weeks across all surgical overlay variants. The correction is surgical and non-additive.

✓ Claim 03 — Full-Sample Result Across All Active Weeks

The headline Sharpe of 3.83 to 4.39 is calculated across all active weeks including scar years at their BASE_K values before carrier repair. This is not a curated subset. The full sample result is the result.

✗ Not Claimed — Live Fund Performance

Autonomous deployment requires forward paper-trading validation. Production implementation at institutional sizing carries transaction cost and slippage risk that the stress grid quantifies but live data must confirm. This is a research architecture. Not a live fund. Not an offer.

The correct institutional statement: architecturally complete, historically audited, forward validation required.

Standard systematic strategies find stress. CDLA finds what stress has not yet priced.

The architecture exists. The audit exists. The track record exists across both the quantitative ledger and five months of published, timestamped constraint intelligence in public markets.

What does not yet exist is the institutional context in which this architecture is applied to a specific mandate, a specific expression universe, and a specific allocator's deployment constraints.

That is what this brief is an invitation to discuss.

Hampson Strategies — hscai.org

Andrew C. Hampson II · @drampson11 · Lafayette, Louisiana

May 2026

Not a live fund · Architecturally complete · Forward validation required · hscai.org

Not investment advice. Research and strategic intelligence for institutional decision-makers. Backtested and model results are hypothetical and do not reflect actual investor returns. Results do not account for fees, slippage, borrow costs, taxes, or execution constraints. Temporal and instrument-provenance audit receipt available for qualified diligence conversations on request. © 2026 Andrew C. Hampson II / Hampson Strategies. All rights reserved.

Institutional Engagement

Constraint-deliverability lag alpha — architecturally complete and audited.

Talk with Us