How the Three-Tier System Works

Introduction

In certain sector-specific bubbles driven by excessive speculation, a pattern has emerged from historical analysis:

Companies building speculative infrastructure tend to fail first, equipment providers (suppliers) fail second, and narrative leaders (the core story) fail last.

This framework is based on TWO historical examples:

Important Limitations: This pattern did NOT appear in 2022 (macro-driven), 2020 (COVID), 1987 (Black Monday), or other crashes. This is EXPERIMENTAL with limited data (n=2). May not predict future outcomes.

The Three-Tier Framework

🏗️ TIER 1: INFRASTRUCTURE DEMAND

Who: Companies building speculative infrastructure

Why they often fail first: High debt, concentrated customers, no moat. When demand softens, they collapse quickly.

Signal: Composite of infrastructure stocks declining from peaks

Why Composite Approach:
We track FOUR components (CRWV 30%, DLR 25%, EQIX 25%, GLW 20%) instead of just CoreWeave to reduce single-stock risk, smooth IPO volatility, and provide broader confirmation.

⚙️ TIER 2: EQUIPMENT PROVIDERS

Who: Companies selling picks and shovels

Why they often fail second: When Tier 1 stops buying, their revenue disappears.

Signal: Sustained underperformance vs broad market (4-week relative to SPY)

Why 4-week measurement: Balances responsiveness vs noise. Requires 4 consecutive weeks at -10% to confirm sustained weakness.

🚀 TIER 3: NARRATIVE LEADERS

Who: Companies at center of bubble story

Why they often fail last: They ARE the bubble narrative. When they break, the story itself is dead.

Signal: Sustained underperformance vs market (equal-weight composite)

Why equal-weight: Prevents NVDA/AAPL from dominating. All 7 equally important to AI narrative.

Historical Validation

2000-2002: ✓ WORKED

Infrastructure (Feb-Mar 2000) → Equipment/Cisco (Apr 2000) → Narrative/MSFT (Sep 2000)

Result: 6-12 month warning window. Signal triggered at SPY -11%, avoided additional -37% decline.

Drawback: 3-5 false positives during 1998-1999 melt-up.

2007-2009: ✓✓ WORKED EXCEPTIONALLY

Mortgage originators (Early 2007) → Banks (May 2007, 5 months before SPY) → Financial sector

Result: Exceptional early warning. Signal at SPY -7%, avoided -51% decline. ZERO false positives.

2022: ✗ FAILED

No cascade observed. All sectors fell together.

Why: Macro-driven (Fed rate hikes), not sector-specific bubble. System not designed for this.

Lesson: Framework has clear boundaries. Works for sector bubbles, NOT macro corrections.

Key Insight

The tier system has worked for SECTOR-SPECIFIC bubbles (2000 tech, 2008 housing finance).

It has NOT worked for MACRO-DRIVEN corrections (2022, 2020).

The 2025 AI situation appears sector-specific, suggesting framework may apply. However, with only 2 confirmed examples, confidence should be limited.

How to Use This System

  1. CHECK WEEKLY (Sunday evenings)
  2. MONITOR TIER 1 for early warnings
  3. WAIT FOR TIER 2 CONFIRMATION
  4. ACT ONLY WHEN TIER 3 CONFIRMS
  5. ALWAYS CONSULT ADVISOR

This is ONE framework among many. Use multiple sources. Maintain diversification. Understand this is experimental with limited validation.

View Case Studies Full Methodology Back to Dashboard