# 🗓️ Warren AI - Roadmap Strategica 2025-2027

**Versione**: 1.0  
**Creato**: 3 Dicembre 2025  
**Owner**: Mauro Gagliardi  
**Sistema**: Warren AI v4.1 (Deep Value Investing)

---

## 📅 DICEMBRE 2025 (Observer Mode)

### Settimana 1-2 (2-15 Dic)
**Focus**: Monitoring passivo + Qualifica watchlist

**Checkpoint Settimanali**:
- ✅ **Domenica 8 Dic**: Review scan settimanale (quick check 5 min)
- ✅ **Domenica 15 Dic**: Review scan + check Germania qualità dati

**Event Critico**:
- 📊 **Mid-Dic (~15 Dic)**: **Azimut Earnings Q4** 
  - Thesis validation: ROE stabile >20%? Guidance 2026?
  - Se ROE <10% → Thesis breaker, considera exit
  - Se ROE 20%+ → Hold confermato

**Action Permesse**:
- Monitoring Warren AI scan results (no acquisti)
- Analisi qualitativa candidates Q1 2026
- Building watchlist con criteri (Score ≥75, Margin ≥10%, diversificazione)

**Action Vietate**:
- ❌ Acquisti (anche se BUY score 95)
- ❌ Trim Azimut (salvo thesis breaker)
- ❌ Deployment liquidità €13.9k

---

### Settimana 3-4 (16-31 Dic)
**Focus**: Prep decisione Gennaio + Market outlook

**Checkpoint Settimanali**:
- ✅ **Domenica 22 Dic**: Review scan
- ✅ **Domenica 29 Dic**: Review scan + **Final watchlist Q1 2026**

**Decision Day Preparation**:
- 📅 **31 Dicembre**: Preparazione fine freeze
  - Conferma freeze end 1/1/2026
  - Watchlist finale: 1-2 candidates qualificati
  - Target entry prices aggiornati
  - Budget allocation plan (max €3.5k per titolo)

**Deliverable**: Lista 1-2 titoli ready per Gennaio con:
- Fundamentals check completo
- Governance/Legal clean
- Fair value + target entry aggiornati
- Rationale diversificazione vs Azimut

---

## 🎯 GENNAIO 2026 (Execution Resume)

### Settimana 1 (1-5 Gen)
**FREEZE END** - Acquisti riabilitati

**Milestone Critico**:
- 📅 **1 Gennaio 2026**: Freeze termina ufficialmente
- ✅ **Domenica 5 Gen**: Primo scan post-freeze
  - Review watchlist con prezzi aggiornati
  - Decision: Deploy su 1-2 titoli se threshold raggiunti?
  - Budget: Max €7k deployed (€3.5k × 2), mantieni €6.9k cash reserve

**Decision Tree Deployment**:
```
SE watchlist candidate ha Score ≥80 AND Margin ≥20%:
  → BUY immediate (max €3.5k)
  
SE watchlist candidate ha Score 75-79 AND Margin ≥15%:
  → Review qualitativa + decision caso per caso
  
SE nessun threshold raggiunto:
  → WAIT, mantieni cash, monitor weekly
```

---

### Settimana 2-4 (6-26 Gen)
**Focus**: Portfolio bilanciamento post-deployment

**Activities**:
- Monitor nuove posizioni (se deployed)
- Tracking Azimut post-earnings
- Verifica concentrazione settoriale
- Target finale Gennaio: 2-3 titoli in Satellite (Azimut + 1-2 nuovi)

**Constraints**:
- Concentrazione max: 40% single stock, 60% single sector
- Cash reserve min: €6-7k (30-35% Satellite)

---

## 📊 26 GENNAIO 2026 - AUDIT ALGORITMO (Week 8)

### CRITICAL CHECKPOINT - Validazione Warren AI v4.1

**Obiettivo**: Validare algoritmo dopo 8 settimane enhanced tracking

**Data Accumulati**: 1200+ data points (150 stocks × 8 weeks)

**Azioni Audit**:
1. ✅ Run `audit_queries.sql` su database
2. ✅ Analisi coverage stability:
   - ROIC coverage: Target >90%
   - Interest Coverage: Target >85% (adjusted for financials)
   - Piotroski F-Score: Target >90%
3. ✅ Review 6 potenziali falle identificate (1 Dic 2025):

**6 Falle da Validare**:
1. **Growth cap 4% mature sectors** - Troppo conservativo?
2. **Country penalty -20% Italia** - Eccessiva per export-heavy companies?
3. **Graham multiplier 15x financial** - Fisso appropriato o dynamic?
4. **AZM bug is_financial** - Debt score 0 vs 15 (asset management ≠ bank)?
5. **Debt penalty threshold 3x EBITDA** - Troppo stretto per capital-intensive?
6. **Luxury brand cap 30x P/E** - Troppo basso per Ferrari/Moncler?

**Decision Tree Post-Audit**:
```
IF false_positive_rate <15%:
  ✅ Algoritmo OK → Keep v4.1 as-is
  
ELIF false_positive_rate 15-30%:
  ⚠️ Algoritmo permissivo → Deploy v4.2 con fix empirici
  
ELSE (>30%):
  🔴 Algoritmo broken → Deep review + comprehensive fix
```

**Deliverable**: 
- Verbale audit completo
- Raccomandazioni modifiche (se necessarie)
- Roadmap v4.2 (se richiesto)

---

## 📈 Q1 2026 (Feb-Mar) - Stabilizzazione Portfolio

### Febbraio 2026
**Focus**: Monitoring + Fine-tuning

**Activities**:
- Review mensile performance Satellite
- Confronto vs ETF Core benchmark
- Warren AI hit rate: % BUY che outperform?
- Se algoritmo audit richiede fix → Deploy v4.2

**Target End-Feb**:
- Satellite: 2-3 titoli deployed (€10-13k invested)
- Cash reserve: €7-10k (50% liquidità maintained)
- Concentrazione: <40% single stock, <60% single sector

---

### Marzo 2026
**Focus**: Q1 Review + Thesis validation

**Milestone Critico**:
- 📊 **31 Marzo 2026**: **Q1 Review Trimestrale**

**Analisi Q1**:
1. Performance Satellite vs Core (%)
2. Thesis check holdings: Azimut ROE? Nuovi titoli quality?
3. Warren AI accuracy: Score prediction vs actual performance
4. Decision: Stay 20k OR scale-up trigger?

**Scale-Up Trigger Evaluation**:
```
IF Satellite performance: +8-12% Q1 (vs +5-7% Core)
   AND Warren AI hit rate: >70% BUY corretti
   AND Confidence level: Alto
THEN:
   → Prepara piano espansione 30k entro Q2
ELSE:
   → Stay 20k, continua monitoring
```

---

## 🚀 Q2 2026 (Apr-Giu) - Consolidation

### Aprile-Giugno 2026

**Activities**:
- Review mensile continuativa
- Se scale-up approvato: Deploy €10k addizionali (20k → 30k)
- Target: 4-5 titoli in portfolio
- Diversificazione settoriale completa

**Milestone**:
- 📊 **30 Giugno 2026**: **Q2 Review Trimestrale**
  - Performance H1 2026 completo
  - Warren AI validation mid-year
  - Thesis breakers: Quanti exit necessari?
  - Decision: Satellite strategy working?

**Target End-Q2**:
- Satellite: 4-5 titoli se expanded a 30k
- Concentrazione: Max 25% single stock (diversificazione aumentata)
- Cash reserve: 20-25% Satellite

---

## 📊 Q3 2026 (Lug-Set) - Performance Validation

### Luglio-Settembre 2026

**Activities**:
- Monitoring stabilità performance
- Backtest decisioni H1 2026
- Analisi false positive/negative rate
- Adjustments portfolio se necessario

**Milestone**:
- 📊 **30 Settembre 2026**: **Q3 Review Trimestrale**
  - Performance 9 mesi
  - Trend analysis vs benchmark
  - Decision mid-point: Strategy on track?

**Critical Questions Q3**:
1. Warren AI mantiene hit rate >65%?
2. Satellite outperform Core YTD?
3. Thesis breaker rate acceptable (<20%)?
4. Risk-adjusted return soddisfacente?

---

## 🎯 Q4 2026 (Oct-Dic) - Annual Validation

### Ottobre-Dicembre 2026

**Focus**: Final validation anno completo + Decision 2027

**Activities**:
- Review mensile + preparation annual review
- Consolidamento posizioni
- Portfolio optimization fine-tuning

**Milestone CRITICO**:
- 📊 **31 Dicembre 2026**: **ANNUAL REVIEW COMPLETO**

**Annual Review Agenda**:
1. **Performance 2026**: Satellite vs Core (full year)
   - Return assoluto (%)
   - Risk-adjusted (Sharpe ratio)
   - Max drawdown
   - Volatilità

2. **Warren AI Validated**:
   - Hit rate annuale
   - False positive rate
   - Score accuracy vs actual performance
   - Coverage stability (ROIC, Interest Cov, F-Score)

3. **Portfolio Management**:
   - Thesis breakers: Quanti exit necessari?
   - Successful holds: % titoli con thesis intact
   - Turnover rate (target <30% annuale)

4. **Lessons Learned**:
   - Algoritmo: Fix implementati efficaci?
   - Execution: Timing entry/exit ottimale?
   - Psychology: Disciplina mantenuta?

**Deliverable**: Verbale annual review + **DECISION 2027**

---

## 🎯 GENNAIO 2027 - Decision Point Strategico

### Critical Decision: Satellite Expansion

**Scenario A** - Performance Positiva (+12-18% annuo):
- ✅ **Espandi Satellite 60-70k** (10-12 titoli)
- Metodo: 
  - Vendita parziale Core ETF (~40k)
  - Nuova liquidità se disponibile (~10k)
  - Total Satellite: 70k (39% portfolio totale)
- Target: Portfolio satellite maturo con diversificazione completa
- Timeline: Deploy graduale Q1-Q2 2027

**Scenario B** - Performance Moderata (+5-10% annuo):
- ⚠️ **Mantieni Satellite 30k** (non espandere)
- Rationale: Performance OK ma non eccezionale
- Continua R&D phase 2027
- Re-evaluate fine 2027 per decision 2028

**Scenario C** - Performance Negativa (<+5% o loss):
- 🔴 **Reduce Satellite a 10k OR abort strategy**
- Rationale: Algoritmo non funziona OR execution issues
- Lessons learned documentation
- Ritorno focus 100% Core ETF passivo
- Satellite mantiene solo best 1-2 holdings quality

---

## 📊 Key Milestones Summary Table

| Data | Evento | Criticità | Action | Owner |
|------|--------|-----------|--------|-------|
| **15 Dic 2025** | Azimut Earnings Q4 | 🔴 Alta | Thesis validation | Mauro + Claude |
| **31 Dic 2025** | Final Watchlist Q1 | 🟡 Media | Prep deployment | Mauro + Claude |
| **1 Gen 2026** | Freeze End | 🔴 Alta | Resume acquisti | Mauro |
| **5 Gen 2026** | Primo scan 2026 | 🟡 Media | Decision deployment | Mauro + Claude |
| **26 Gen 2026** | Audit Week 8 | 🔴 Alta | Validate algoritmo | Mauro + Claude |
| **31 Mar 2026** | Q1 Review | 🟡 Media | Performance check | Mauro + Claude |
| **30 Giu 2026** | Q2 Review | 🟡 Media | Scale-up trigger? | Mauro + Claude |
| **30 Set 2026** | Q3 Review | 🟡 Media | Mid-year validation | Mauro + Claude |
| **31 Dic 2026** | Annual Review | 🔴 Alta | Decision 2027 | Mauro + Claude |
| **Gen 2027** | Expansion Decision | 🔴 Alta | 20k vs 60-70k | Mauro |

---

## ⚠️ Risk Factors da Monitorare

### Tecnici
- **Germania qualità dati <90% persistente** → Escludi mercato da scan
- **Warren AI false positive rate >30%** → Fix urgente algoritmo
- **Database corruption/loss** → Backup strategy + recovery plan
- **Yahoo Finance downtime prolungato** → Valuta data provider alternativo

### Mercato
- **Correzione -15-20%** → Opportunity deployment accelerato (watch margin expansion)
- **Bull market continuation** → Zero BUY da algoritmo (conservativo by design, acceptable)
- **Recessione** → Thesis breaker checks più frequenti (ROE, margins, debt coverage)
- **Tassi interesse +200bps** → Impact valuation multiples, recalibra fair value

### Personali
- **FOMO emotional trading** → Guardrail intervention automatico
- **Overtrading frequency >3/mese** → Red flag attivato
- **Thesis breaker ignored** → Forced review + escalation
- **Cherry-picking algoritmo** → Override senza rationale → Blocked

---

## 📝 Checkpoints Ricorrenti

### Weekly (Ogni Domenica Sera 19:30)
**Post-scan automatico Warren AI**

**Review rapido** (5-10 min):
- Quality alerts? (Germania <90%? Stale data?)
- BUY emergenti? (Score ≥80 + Margin ≥20%)
- Top 10 per mercato (shift significativi?)
- Watchlist updates? (Target price approach?)

**Output**: Quick summary 3-5 paragrafi

---

### Monthly (Ultimo Giorno Mese)
**Performance Review**

**Analisi** (30 min):
- Performance Satellite vs Core (%)
- Thesis validation holdings (ROE, margins, debt)
- Adjustments needed?
- Concentrazione settoriale check

**Output**: Performance summary + action items

---

### Quarterly (Fine Trimestre)
**Strategic Review**

**Deep Dive** (1-2 ore):
- Performance vs benchmark (risk-adjusted)
- Warren AI accuracy (hit rate, false positives)
- Portfolio optimization needs
- Scale-up/down triggers evaluation
- Strategic adjustments piano

**Output**: Verbale trimestrale completo + decisions

---

## 🎯 Success Metrics (KPI)

### Performance Metrics
- **Target annuale Satellite**: +12-18% (vs +7-10% Core ETF)
- **Max drawdown acceptable**: -15% (vs -10% Core)
- **Volatility target**: <20% annuale
- **Sharpe ratio target**: >0.8

### Algorithm Metrics
- **Warren AI hit rate**: >70% (BUY recommendations che outperform)
- **False positive rate**: <15% (BUY che underperform)
- **Coverage stability**: >90% (ROIC, F-Score sempre disponibili)
- **Score accuracy**: Correlation score-performance >0.6

### Portfolio Management
- **Thesis breaker rate**: <20% (exit per deterioration fundamentals)
- **Turnover annuale**: <30% (buy & hold 3-5 anni)
- **Concentrazione**: <40% single stock, <60% single sector
- **Cash reserve**: 20-30% Satellite (dry powder)

### Behavioral Discipline
- **Freeze violations**: 0 (absolute compliance)
- **Emotional overrides**: <2/anno (solo con rationale documentato)
- **Watchlist creep**: <10 titoli totali (focus)
- **Overtrading**: <1 trade/mese medio

---

## 📚 Reference Documents

**Core Architecture**:
- `DIARIO_DI_BORDO.md` - Storia completa progetto
- `SCORING_ENGINE.md` - Formula algoritmo proprietaria
- `PROJECT_INSTRUCTIONS.md` - Bootstrap istruzioni Claude

**Decision Log**:
- `verbale_riunione_warren_claude_integration_02-12-2025.md`
- `verbale_validazione_algoritmo_warren_03-12-2024.md`
- `verbale_deployment_warren_v4_1_enhanced_tracking_01-12-2025.md`
- `verbale_implementazione_schema_v4_enhanced_quality_02-12-2024.md`

**This Document**: `ROADMAP.md` (Strategic planning 2025-2027)

---

**Versione**: 1.0  
**Status**: 🟢 Active  
**Next Review**: 31 Dicembre 2025 (Final Watchlist Q1 2026)  
**Owner**: Mauro Gagliardi  
**AI Support**: Claude (Warren AI System)

---

*"Rule No. 1: Never lose money. Rule No. 2: Never forget rule No. 1."* - Warren Buffett

*"The stock market is a device for transferring money from the impatient to the patient."* - Warren Buffett