TMC Research structures versioned hypotheses, challenges them through analysis by multiple AI agents and produces reproducible evidence before any human decision is made.
Every idea becomes a traceable formulation with clear validation criteria.
Specialized agents propose, challenge and document their reasoning.
The findings inform decisions without replacing human accountability.
Data, strategies, backtests and metrics are versioned to remain verifiable.
TMC Research is OCTOVEX’s private AI-assisted quantitative research lab.
Its mission is to turn trading ideas into testable hypotheses, structured strategies, controlled datasets and reproducible quantitative experiments.
TMC reduces uncertainty through research and validation. OCTOVEX then provides the architecture, development and production delivery when the evidence supports the transfer.
Decision principle
Specialized agents examine the same hypothesis from multiple angles.
Reproducible data and backtests test the idea against observable results.
The final decision remains explicit, justified and tied to a specific version.
Each area maps to modules, capabilities or protocols that are documented in TMC Research.
Formalize an idea, version its criteria and translate it into a structured strategy without losing its history.
Study regimes, liquidity and market structure to define the conditions surrounding a hypothesis more precisely.
Examine limitations, biases and the robustness of the results before drawing a conclusion.
Validate, normalize and seal versions of market data with a deterministic quality report.
Simulate strategies under a conservative policy and retain versioned trades, metrics and curves.
Cross-check analyses, syntheses and indicator code to detect inconsistencies, repainting and look-ahead bias.
The method separates exploration, experimentation, measurement and decision-making so that every conclusion remains verifiable.
Define the scope of a project, a research question and its known limitations.
Create a versioned hypothesis with explicit criteria.
Have multiple specialized agents analyze and critique the hypothesis.
Combine a strategy, controlled data and a deterministic backtesting protocol.
Synthesize the evidence, quantify the risks and preserve the human decision.
Turn validated knowledge into architecture and an operational product.
The repository describes internal programs, infrastructure and protocols. It publishes no returns, trading signals or promises of performance.
Projects, versioned hypotheses, multi-agent AI runs, summaries and immutable human decisions form a complete, traceable cycle where the human keeps the final call.
A separate Python engine simulates one position at a time under a conservative policy, then persists trades, metrics, equity curves and comparisons atomically and versioned.
An initial program focused on XAUUSD, designed to document hypotheses, limitations and evidence scientifically before drawing any conclusion.
An immutable framework connecting source, validated indicator specification, instrument, economic profile, sealed data and execution conventions, while separating gross results, costs and net results.
The transfer does not promise that an idea will become a product. It establishes a clear path from validated knowledge to an engineering decision.
TMC Research
TMC Research
TMC Research
OCTOVEX
OCTOVEX
OCTOVEX
These principles are applied throughout TMC Research’s architecture, workflows and business rules.
Every hypothesis has a framework, criteria and a documented rationale.
Versions, data and parameters make it possible to rerun an experiment.
Decisions, audits and snapshots preserve the context behind every result.
Analyses actively seek out biases, inconsistencies and fragile assumptions.
AI assists the research process without becoming the final authority.
A lack of evidence or reliable data remains visible and prevents unsupported conclusions.
TMC Research structures the exploration. OCTOVEX provides the architecture, development and production delivery.
