A historical baseline
Before deep models, an early and famous feed ranking idea scored each post, called an edge, using three factors multiplied together. It is a clean mental model for how ranking works.
The three factors
- Affinity measures the relationship between the viewer and the author. Frequent interaction raises it.
- Weight measures the post type. A comment or a video might carry more weight than a simple status.
- Time decay lowers the score as the post ages, so fresh content stays competitive.
The score is roughly affinity times weight times decay, summed or compared across candidate edges to order the feed.
Why this shape endures
- It is interpretable. You can explain why a post ranked high.
- It captures the three forces every feed cares about, relationship, content value, and freshness.
- Modern models replace the fixed multiply with learned weights, but they still consume the same families of signals.
The lesson is the structure, not the exact formula. Affinity, content weight, and decay remain the backbone of feed scoring.
Key idea
An edge style score multiplies viewer author affinity, post type weight, and time decay into one number, capturing relationship, content value, and freshness.