Introducing Macro Analysis: 10 Conditional Markets, 22 Signals, 2 Live Themes

Today we're launching a new analytical vertical: macro conditional analysis. Where our equity research asks “is this stock what it claims to be?”, macro analysis asks a different question: if a specific policy event happens, what are the measurable downstream consequences? We've built the infrastructure to answer that with the same multi-lens rigor we apply to equities — and we're launching with two live themes covering 10 conditional market pairs, 22 signals, and 180 model predictions.

Why Conditional Analysis

Macro forecasting typically produces a single point estimate: “core PCE will be 2.4% by June.” That number is worse than useless if you can't trace why. Does it assume the Fed cuts? Holds? What if tariffs persist? What if they expire? A probability without a causal model is just a guess with decimal places.

Conditional markets solve this by splitting every question into two branches: IF the condition is true and IF the condition is false. The gap between branches — the causal delta — is the actual analytical signal. It tells you how much a specific event matters for a specific outcome, isolated from everything else.

Multi-Lens Analysis

Each macro theme is analyzed through 5–6 specialized lenses, following the same adversarial process we use for equities: two independent analysts (Opus + Sonnet) produce parallel assessments, a synthesizer reconciles them, and a reporter distills the output into structured signals with evidence grades. The result is not a single narrative but a map of where the models agree, where they disagree, and what the evidence actually supports.

Seven lenses are currently available across both themes, each producing two calibrated signals:

Each lens page documents its core question, analysis stages, required data sources, and signal spectrum in full. Explore the full framework →

Signal Dashboard

Every signal is assessed on a four-level spectrum with severity coloring — green for benign conditions, yellow for watch-worthy, orange for concerning, red for crisis-level. The signal dashboard provides a real-time snapshot of where conditions stand across all analytical dimensions.

For example, in our trade policy analysis, the Tariff Pass-Through signal currently reads ABSORBED (green) — firms are eating the cost through margin compression rather than passing it to consumers. But the same analysis flags that this absorption is running on depleting buffers. The Financial Conditions Stance reads LOOSE (orange) — credit is still flowing easily despite the tariff shock, which complicates the Fed's decision framework.

Two Live Themes

What Makes This Different

Most macro commentary gives you a narrative. We give you a structured decomposition: 7 lenses, 22 signals, 10 conditional market pairs, and 180 individual model predictions that can be scored against reality when the outcomes resolve. Every claim is traceable to specific data, assessed at a specific evidence level (E0–E3), and produced through an adversarial multi-model process that surfaces disagreement rather than hiding it.

The prediction ensemble uses 9 model configurations per market branch — Opus, Sonnet, and Haiku each run 3 times with varied prompts — producing 18 independent estimates per conditional pair. We report the median, the spread, and the model agreement level. When 8 of 9 models agree, that means something different than when they split 5-4.