Metals Market Intelligence

AI Finance / Macro AnalysisProduction

Regime detection for metals markets across 34 macro, structural, and behavioral signals.

Overview

An analysis system for metals markets. It combines 34 macroeconomic, structural, and behavioral signals to detect which market regime is in force and to catch pattern shifts that price-based analysis misses.

The output is structured for trading decisions rather than for charts: a stated regime and the signals behind it, not a dashboard.

What it does

Multi-factor analysis across 34 signals
Market regime detection
Behavioral pattern recognition
Signals produced for trading decisions

Architecture

Signal ingestion layer
Pattern recognition models
Regime detection module
Decision output interface

Impact

Structured macro signals for trading decisions
Behavioral patterns that traditional analysis misses
Output a decision can be based on, not more charts
Foundation for a broader macro intelligence platform

Macro regime detection for metals markets

Want a system like this?

Book a strategy call

Why 34 signals rather than a model that predicts price

Direct price prediction on liquid markets is where quantitative projects go to die: at short horizons the price is close to a random walk, and any edge that exists is arbitraged away by participants with better data and faster execution. Regime detection asks a different, more tractable question: not where the price goes, but which set of relationships currently drives it.

That is why the input is a wide, mixed panel rather than a long price history. Macro conditions, market structure, and behavioural indicators fail in different circumstances, and a regime shift shows up as several of them changing character at once, well before it is readable in the price alone.

Output built for a decision, not a dashboard

A dashboard hands the work back to the human: it shows thirty-four charts and leaves the synthesis to whoever is looking. That synthesis is the step the system exists to do, and a tool that skips it gets used eagerly for a month and then not at all.

Structured output — a stated regime, the signals that moved, and in which direction — can be checked, logged, and argued with after the fact. It also makes the system's record auditable: what it said, when, and whether it was right. That record is the only good reason to keep using it.