EXECUTION POLICY RESEARCH & DECISION INFRASTRUCTURE — VERSION 1.0

Deterministic, auditable execution policy for live market regime intelligence.

Narion Execution Research is the execution-side infrastructure of the Narion ecosystem — a transparent, fail-closed decision pipeline that fuses live signal intelligence, six-dimensional regime classification, and a research-validated policy universe into a single deterministic execution decision, from live market state through to a fully reconstructable audit trail.

Access is intended for research partners, engineering collaborators, and technically qualified reviewers studying the methodology. Full risk and research disclosure is available in the Research & Risk Disclosure section below.

Request Research Access → View Console Preview

REGIME UNIVERSE

650 → 77 Validated

VALIDATION STAGE

Step 23.6 — Pass

LIVE EXECUTION

Controlled / Gated

METHODOLOGY

Paper-First, Fail-Closed

01 THE RESEARCH CONSOLE

Every decision is shown, not hidden.

This is the actual internal operations console. It exposes live signal state, regime classification, and the research-approval reasoning behind every candidate in real time, so the pipeline can be inspected directly rather than taken on faith.

N Operations Dashboard — Narion... ×

CURRENT SESSION

Active trading session

P_LONG

0.010

Long propagation probability

TRADES TODAY

Shadow dataset files

OBSERVATIONS

185,858

Total rows ingested

LIVE SIGNAL DETAILS — BTC/USDT PERP STREAMING

CONFIDENCE BUCKET

NOISE

EMA CONFIDENCE

0.010

VOLATILITY PERCENTILE

9.3%

BTC PRICE

$78,956

BTC/USDT PERP

REGIME INTELLIGENCE LIVE

CURRENT COMPOSITE REGIME

Live regime classification

RESEARCH APPROVAL

NOT APPROVED

IGNORE

Live capture of the internal research console. "IFAE ACTIVE" denotes the signal-generation engine — order execution is handled by a fully isolated downstream layer (see Architecture).

02 RESEARCH METHODOLOGY

What the system studies — and how it is deliberately constrained.

Narion Execution Research separates market intelligence from execution mechanics, and separates both from any claim about outcomes. Every stage exists to make one narrow question answerable and auditable: given the current state, should an execution attempt even be permitted?

WHAT THE SYSTEM STUDIES

Whether live market state maps to a research-validated regime

Whether a candidate meets confidence and freshness requirements

How a deterministic policy gate should respond given that context

Whether the resulting decision is fully reconstructable after the fact

How paper execution compares to the same policy applied live

ENGINEERING SAFEGUARDS

No live exchange orders outside a controlled, explicit configuration

No manufactured signal when the market does not produce one

No treatment of a low long-probability reading as a short signal

No live behavior that has not first passed paper validation

No component approved as production-ready from a single runtime session

STEP 01

Live Intelligence

Consume LiveState from Narion Intelligence

STEP 02

Runtime Context

Join intelligence with local market data

STEP 03

Regime Classification

Six-dimensional composite regime

STEP 04

Research Lookup

Validated regime, or fail-closed IGNORE

STEP 05

Candidate + Policy

Confidence gate, regime gate, decision

STEP 06

Paper / Controlled Live

Journaled, auditable, reversible

03 SYSTEM ARCHITECTURE

Eight layers, each independently testable.

No layer can act outside its responsibility — the regime engine cannot submit orders, the candidate engine cannot manage positions, and the policy engine cannot implement broker logic. This separation is what makes each transition independently researchable, and what lets the system evolve one layer at a time without destabilizing the rest.

Signal Intelligence

Consumes Narion Intelligence and produces LiveState — the authenticated, real-time model output.

Market Intelligence

Constructs local market data from price, EMA structure, volatility, market structure, impulse, VWAP, and momentum features.

Regime Intelligence

Classifies trend, strength, cycle, expansion, VWAP context, and structural location into a composite regime.

Research Policy

Maps the live composite regime against the validated research universe of 77 production regimes.

Candidate Generation

Determines whether an execution candidate exists — not whether it is approved.

Execution Policy

Applies confidence, regime, and candidate gates deterministically via ExecutionPolicyV1.

Execution

Routes approved decisions to PaperBroker, or a controlled live adapter (Delta Exchange, PI42 in development).

Audit & Research

Journals decision, outcome, market and model state into a continuous research feedback loop.

04 REGIME RESEARCH

A deliberately narrow production universe.

The research pipeline observes a far larger universe of market states than the execution layer is permitted to act on. That gap is intentional — production is constrained to regimes with sufficient research evidence behind them, each scored on reliability, expectancy, profit factor, win rate, and risk/reward before it is ever eligible for a live decision.

650

OBSERVED COMPOSITE REGIMES

The full space of six-dimensional market states observed during research qualification.

VALIDATED PRODUCTION REGIMES

The subset carried forward with sufficient research evidence to inform a policy decision.

IGNORE

DEFAULT FOR THE UNKNOWN

A regime without validated evidence produces no approval — fail-closed by design, not by omission.

TREND TREND STRENGTH MARKET CYCLE EXPANSION VWAP CONTEXT STRUCTURAL LOCATION

05 HYBRID EXECUTION & STATE MANAGEMENT

A negotiated fill, not a single order type.

Approval by the policy layer is not the end of the execution problem — it is the start of a separate, equally deliberate one: how to get filled without paying away edge to slippage, and how to guarantee nothing is left open once a position closes. Narion Execution treats order placement and position lifecycle as their own engineered subsystems, isolated from the policy decision that triggers them.

STEP 01

Maker Attempt

Post-only limit priced just inside the live spread

STEP 02

Fill Window

Bounded wait before any fallback is considered

STEP 03

Standard Limit

Rejected maker retried as a normal GTC limit order

STEP 04

Market Fallback

Guarantees entry when neither limit variant fills

STEP 05

Bracket Attachment

Stop-loss and take-profit placed immediately on confirmed entry

HYBRID EXECUTION LADDER

Maker-first routing prioritizes price improvement over speed

Automatic fallback chain removes manual intervention from the fill path

Every fallback branch is logged, so the realized entry path is always reconstructable

Stop-loss and take-profit are structural — no confirmed entry exists without a bracket

A single existing-position check gates the entire sequence before it starts

EXECUTION STATE MANAGER (ESM)

Registers ownership of a position the moment an entry is confirmed

Continuously monitors live position state against the exchange, not an assumption

Detects position closure and automatically sweeps remaining open orders for that instrument

Prevents orphaned stop-loss or take-profit orders from persisting after a position is gone

Resets to a clean idle state only once cleanup is confirmed complete

06 ENGINEERING VALIDATION

Architecture validated, boundary by boundary.

Step 23 validated the full transition from decision architecture to controlled paper execution — signal ingestion, six independent regime engines, the composite regime builder, regime lookup, candidate generation, policy gating, journaling, and execution isolation, each confirmed independently.

Live signal ingestionPASS

Runtime context assemblyPASS

Six regime enginesPASS

Composite regime builderPASS

Regime lookup enginePASS

Unvalidated regime → IGNOREPASS

One-sided p_long semanticsPASS

Candidate enginePASS

Confidence & regime gatesPASS

ExecutionPolicyV1PASS

Decision journalPASS

Paper / live execution isolationPASS

26 architecture boundaries validated through Step 23.6 — full validation ledger, including decision-journal schema and controlled paper-run logs, available on request for research partners.

07 PAPER-FIRST & FAIL-CLOSED CONTROLS

Live execution is a controlled exception, not the default.

The single most important safety boundary in the system is a switch, not a strategy: when disabled, no exchange adapter, order manager, or broker connection is even initialized — real orders are architecturally impossible, not merely discouraged.

Live-Execution Gate

A single explicit switch. When disabled, no live adapter, order manager, or broker connection initializes.

Fail-Closed Regimes

An unvalidated market state produces IGNORE by default, never an assumed approval.

Stale-Signal Rejection

Signals beyond a freshness window are rejected rather than acted on.

Leverage & Margin Checks

Independent validation ahead of any position sizing decision.

Cooldowns & Trade Caps

Explicit limits on execution frequency, independent of signal frequency.

Decision + Trade Journals

Two separate, append-only records — why a decision was made, and what actually happened.

NO ARTIFICIAL SIGNALS

If the underlying market and model do not produce a qualifying signal, there is no trade — paper or live. Validation runs are never permitted to manufacture a signal simply to demonstrate that execution "works."

08 RESEARCH & RISK DISCLOSURE

Positioning, scope, and risk — stated once, plainly.

The engineering sections above describe what the system does and how it is validated. This section states, in one place, what that validation does and does not establish, and the risks involved in studying or using this infrastructure.

Research & Education, Not a Product Guarantee

This is decision-support and execution-policy research infrastructure. It is not a signal service, not a managed trading account, and not offered as a means of generating profit.

Architecture Validation ≠ Profitability

Passing engineering validation confirms the decision pipeline behaves as designed. It says nothing about future market performance, expectancy, drawdown, or trading outcomes.

High-Risk Instruments

Leveraged derivatives carry substantial risk of loss. Anyone studying or building on this research should test extensively in paper mode and validate risk logic independently before considering any live use.

FULL DISCLOSURE

Not financial advice. Nothing in this system, this page, or any related materials constitutes investment, trading, or financial advice, and none of it is a recommendation or solicitation to buy, sell, or hold any instrument.

No profitability claim. Model profitability, long-term paper expectancy, slippage, fees, drawdown, and risk-adjusted performance have not been established and require sustained, independently reviewed observation.

Experimental infrastructure. This is research and engineering work in progress. It is not production-grade brokerage software and carries no uptime, accuracy, or outcome guarantee of any kind.

Paper before anything else. Any evaluation of this architecture should begin — and, absent extensive independent validation, remain — in paper mode with live execution disabled.

Own-risk use. Anyone using, adapting, or studying this infrastructure does so entirely at their own risk and is responsible for independently verifying any exchange integration, risk control, or execution logic before relying on it.

Research access, not retail onboarding. Access is intended for research partners, engineering collaborators, and technically qualified reviewers studying the methodology — not for individuals seeking an automated path to trading profit.

09 WITHIN THE NARION ECOSYSTEM

Research, manual execution, and execution-policy study.

Three distinct products, three distinct responsibilities — no overlap between market research, manual trading, and execution-policy research.

MARKET RESEARCH

Narion Intelligence

Generates probabilistic market intelligence and signals from institutional order flow research.

MANUAL EXECUTION

Narion Trade Panel

Executes a trader's own manual decisions with institution-inspired execution quality.

EXECUTION-POLICY RESEARCH

Narion Execution Research

Studies how intelligence, regime research, and policy combine into an auditable, fail-closed decision pipeline — paper-first, by design.

YOU ARE HERE

10 RESEARCH ROADMAP

Built on the validated junction, extended without redesign.

Step 23 established the validated regime/policy junction. Future work extends measurement and infrastructure on top of it — WebSocket streaming to close the latency gap, a dynamic risk layer, deeper execution analytics, and eventual multi-venue routing — without reopening the fail-closed design.

Infrastructure

WebSocket signal streamingAsync networkingLower-latency architectureVPS deployment

Risk Layer

Dynamic sizingPortfolio exposure controlsAdaptive leverageVolatility-adjusted execution

Analytics

Execution analyticsLatency metricsSlippage monitoringTrade attribution

Multi-Venue

Async executionMulti-venue routingPortfolio managementAdvanced risk engines

Expansion of controlled live execution is gated on independently-reviewed evidence — the roadmap governs measurement and infrastructure, not a fixed timeline.

Study the methodology.

Research briefings and methodology walkthroughs for qualified research partners and engineering collaborators are reviewed within 48 hours.

Request Research Access → All Contact Options