I noticed a blind spot in my own engine recently. The algorithm wasn't properly factoring in the massive extraordinary dividend triggered by the TSB subsidiary sale.
Why? Because the sale happened over 90 days ago. The engine's "memory" rolling window was too short and it dropped the context.
To fix this, I just built and deployed an "AI news condensation" layer into DividendQuad. It periodically reads the entire history of a stock and writes a permanent context file for structural events (like the TSB sale or the BBVA takeover attempts) so the math engine never forgets major corporate actions.
After the update, Sabadell's analysis is fully dialed in:
🟢 OPTIMAL
Quality: 65/100 (Solid business)
Opportunity: 75/100 (good entry point)
Yield: 24% TTM (extraordinary + regular)
Forward normalized yield: ~6.9%
Building the tools you actually use to invest is a wild process.

