AI analysis

How to read a stock narrative

Each day (Madrid early hours), we generate a narrative for every watchlist item based on:

• Price and recent technical metrics (last close, 1D/1W/1M/YTD changes, 30d avg volume)
• The stock's sensitivity profile: positive catalysts, negative catalysts, macro sensitivities
• Up to 5 relevant news items from the last 72h (Marketaux with per-entity sentiment)
• The day's market pulse

The model is Claude Sonnet 4.5 with prompt caching enabled: narratives are generated consistently and at controlled cost.

The opportunity score: what it is and what it is NOT

The score is an integer from 1 to 10 that reflects the model's read of how much opportunity it sees in the current context.

What it IS:

• A qualitative synthesis of favorable catalysts vs visible risks
• Useful for prioritizing which stocks deserve more attention this week
• A change signal: if a stock moves from 4 to 8 between Monday and Friday, something happened worth your attention

What it is NOT:

• A buy recommendation. It is not. Say it out loud.
• A price target calculation, nor a certified quantitative model
• A guarantee of anything. Today's 8 can be tomorrow's 3 with no warning

Catalysts and risks

Every narrative lists:

• 2-3 active catalysts — upcoming earnings release, product launch, pending regulation, rumored M&A, etc.
• 2-3 risks — analyst downgrade, sector exposure, macro headwind, geographic dependence, etc.

Read them as testable hypotheses, not certainties. The purpose is to give you the angles to watch, not to tell you what's going to happen.

Short / medium / long-term theses

Each stock has three independent theses:

• Short (1-3 days): expected immediate behavior. Useful if you'll trade this week.
• Medium (1-4 weeks): developing dynamics. Useful for swing trading or tactical adjustment.
• Long (3-12 months): structural direction. Useful for strategic positioning.

The three theses can contradict each other — that's normal. A stock can have a bearish short-term thesis (profit taking) and a bullish long-term thesis (fundamental case intact). None tells you what to do; all are model interpretations of public data.