ValthorLabs

Current Validation Status

Status before performance claims.

Performance records require clearly separated data sources, modelling assumptions and forward-observation status.

SystemHistoricalOOSStressForwardLiveLast Update
MMO-XIn progressIn progressActiveActive demo observationNot started14 Jun 2026
XRM5Candidate frozenCompleteCompleteDemo pendingNot started6 Jul 2026
XAU-RLExperimentalN/AN/AN/AN/A12 Jun 2026

Status Legend

Validation terms used across systems.

Each label describes evidence progress, not expected financial outcome.

Not started

No defined test run has started for the selected evidence type.

In progress

Testing is underway, but the planned evaluation is not complete.

Partial

Only part of the planned validation scope has been completed.

Active

The test or observation is currently running or regularly monitored.

Complete

The planned scope is complete for the stated evidence type.

Rejected

The hypothesis did not meet the validation criteria.

Pending publication

The work exists internally, but the public note is not yet prepared.

Experimental

The work is exploratory and not attached to production deployment.

Validation Evidence

What turns a claim into a status.

Evidence is grouped by test type: historical testing, out-of-sample checks, stress review, demo forward and live deployment.

Data

Test Environment

Instrument coverage, real-tick availability, spread and commission assumptions, and broker settings are recorded with each evidence note.

Robustness

Stress Review

Selected candidates are checked against spread stress, execution delay, parameter sensitivity and period dependency.

Decision

Reject Criteria

Fragile parameter behavior, single-period dependency, weak forward behavior or unacceptable execution quality can reject a hypothesis.

How It Works

Research flow before conclusions.

Each system is reviewed as a controlled research track. Methodology is described without exposing internal execution rules.

01

Observe

Market state, session context and basic execution conditions are monitored before any system logic is evaluated.

02

Classify

The research track identifies whether the current environment matches the intended operating profile.

03

Validate

Candidate behavior is tested across historical, stress and forward-observation windows before conclusions are drawn.

04

Constrain

Risk limits, exposure controls and execution quality are treated as design constraints, not afterthoughts.

05

Review

Results are compared against drawdown, stability and regime-specific behavior before status changes.

Execution Pipeline

Decision flow without exposing strategy rules.

The pipeline view documents how a system moves from market observation to execution readiness. It is shown as architecture, not as a trading recipe.

01Snapshot

Collects current market, session and execution context.

02Adapter

Normalizes observations into the system's internal decision format.

03Selector

Chooses the active research regime or strategy family.

04Gate

Blocks candidates that fail context, quality or timing checks.

05Intent

Builds a candidate action only when the prior checks align.

06Guard

Applies risk, exposure and operational constraints.

07Plan

Prepares execution assumptions before any broker interaction.

08Executor

Handles the final execution path and broker response monitoring.

FAQ

Validation questions before contact.

Short answers about validation status, evidence and research boundaries.

What does active validation mean?

It means a system is still being tested across controlled historical, stress or forward-observation conditions. It does not mean a live performance record is available.

How are backtests separated from forward results?

Backtests, out-of-sample checks and forward observation are treated as separate evidence types. Status labels do not combine them into a single performance claim.

What data quality is used?

Data source, modelling assumptions, spread and execution conditions are considered part of validation. Detailed datasets are documented only when evidence notes are prepared.

How is overfitting controlled?

Candidate logic is reviewed across multiple periods, stress windows and forward conditions. A result that only works in one narrow slice is not treated as validated.

Are results independently verified?

Not yet. Performance records remain pending until source, method and limitations are clearly stated.

Why are strategy rules not published?

Methodology, status and limitations can be documented without exposing entry rules, execution parameters or code-level logic.

Are systems running on demo or live accounts?

Current work is research and controlled demo observation. Live capital deployment is not marked as started.

What would cause a system to be rejected?

Instability across regimes, excessive drawdown, fragile parameter behavior, poor execution quality or failure in forward observation can all invalidate a candidate.