An experiment in what an agentic-AI research framework can do — published in full so you can check its work, not a tip service to follow. Generic research and education, not advice, not a personal recommendation, not FCA-authorised. Hypothetical book, not real money. Capital at risk.
Portfolio Lab · Research Lab
Portfolio Lab — transparent, low-cost portfolios with real out-of-sample performance, net of all costs. View the portfolios →

← All concepts

Confluence — why we stack independent signals

The best ideas are rarely built on a single reason. They're built where several independent reasons all point the same way: a company is cheap against its peers, and it's buying back its own shares, and its own executives are buying with their own money, and management has a track record of doing what they said. Any one of those alone is interesting. All of them together is rare, and far more convincing.

What it is

Confluence is the idea that conviction should come from how many genuinely independent things agree, not from how loudly any single one shouts. Think of it like witnesses to an event: one account is interesting; four unconnected people describing the same thing is compelling.

Why it matters

Every individual signal is noisy — it's right often enough to be useful, but wrong plenty of the time. Stacking independent signals is powerful because the noise tends to cancel while the genuine information adds up. The crucial word is independent. Two signals that are really the same fact dressed up differently are one reason, not two — for example, "the company is doing a buyback" and "its share count is falling" are the same thing, and counting both would just be fooling ourselves into false confidence.

How we use it

We keep a map of which signals are genuinely independent of one another — ownership behaviour, valuation, business quality, how management spends cash, catalysts, and management credibility. An idea's conviction rises with the number of independent strands that line up. A single strong signal can earn a small position; several independent ones aligning is what earns real size.

The honest caveat

More strands is not automatically better. Piling up weak signals — three coin-flips — doesn't make a strong case; it makes a cluttered one. Each strand has to carry its own evidence first (see Evidence over narrative). Confluence multiplies good signals; it can't rescue bad ones.

A plain-English explainer of how we think — part of our evidence-driven framework. Not investment advice.