Radar Perene / concept
Vértice — the methodological experiment
Vértice is the methodological experiment of Radar Perene. It does not replace the five lenses; it attempts a different kind of reading — cross-domain hypotheses between regulatory, market, crypto, sentiment, and fragility — under Bayesian discipline.
What it is
A system that formulates hypotheses about relations between domains that normally do not converse (for example, global crypto liquidity and Brazil's short rates; US credit stress and Brazilian country risk; tech VC flow and Brazilian small caps). Each hypothesis is structured with evidence, contradictions, what strengthens it, what weakens it, next signals to watch, and categorical contextual confidence.
Vértice runs at a faster cadence than the principal lenses (the quantitative engine recomputes every fifteen minutes), but the dispatch of hypothesis generation is surgical — only on regime change, relevant crossing, or detected anomaly.
Why it exists
Because there are relevant wealth-side readings that do not fit classical macro. Sophisticated Brazilian capital operates in a globalized environment, with exposure to crypto, EM fragility, tech cycles, US liquidity. Treating those dimensions as noise loses reading; treating them as signal requires a methodology different from classical domestic macro.
Vértice is the attempt to offer that reading without betraying the Radar's principles — every hypothesis is structured, falsifiable, with contradictions listed, and no hypothesis becomes a recommendation.
How to read it
Each Vértice hypothesis appears as a structured object:
- Title — short sentence stating the thesis.
- Type — lead-lag cross-asset, cross-asset divergence, regime change, historical analog.
- Evidence — signals that sustain it.
- Contradictions — signals that push against it.
- What strengthens — events or readings that would increase confidence.
- What weakens — events or readings that would reduce confidence.
- Next signals — what to watch in the coming weeks.
- Contextual confidence — categorical (low, moderate, moderate-high, high).
- Adversarial critique — why this hypothesis might be wrong (mandatory second reading).
Without evidence and contradictions, the hypothesis is rejected before it appears.
What it does not mean
It does not mean a forecast. A hypothesis is an updatable reading; a forecast is a deterministic statement. Vértice never says "X will happen" — it says "this hypothesis gains strength if A, B, C persist; it loses strength if D, E appear".
It does not mean a trade. Recomputing faster (every fifteen minutes) does not become an operational signal. It is a fresher regime reading, not a market signal.
Limits
Vértice requires accumulated history to calibrate its categorical confidences — in early waves, many hypotheses remain in calibration. Formal backtest of hypotheses is performed at t+10, t+20, t+60, and t+120 days, and confidence labels are recalibrated against real historical hit rates. Cross-domain analysis with a short sample is declared as such.
Example
Active hypothesis: US credit stress precedes Brazil risk. Evidence: HY OAS up 80 bps over 60 days; Brazil country risk stable; finance/IBOV ratio at the 25th percentile. Contradictions: parallel deterioration in domestic regulatory regime; positive foreign flow. Contextual confidence: moderate. Historically similar windows: 2018-11, 2020-02. Mandatory microcopy: contextual hypothesis, not forecast.
Vértice formulates hypotheses; history calibrates confidence; the decision remains human.
Related concepts: Historical Analogs · Intermarket BR · Global Regime · see the full method