Under The Hood X Report Analyzer
Engagement Signal Guide

How Quotes Affect X Reach

Quote posts are separately modeled by X's recommendation system. The current Phoenix action set includes quote as an engagement prediction, distinct from an ordinary repost. That means the system can distinguish between a repost and a quote rather than treating both as exactly the same behavior.

(GitHub)

Why Quote Posts Are Different

A quote post adds the user's own content around the original post (Original post + User's commentary). That creates a different type of interaction from simply reposting the original. The current model therefore has a separate predicted action for quotes.

How Quote Prediction Works

Original / Candidate Post
+ Viewer Context
Phoenix
P(Quote)
Weighted scoring
Overall ranking

The important part is P(Quote). The system is predicting how likely the viewer is to perform the action. It isn't simply adding the number of existing quotes to the post's score.

Quote Count vs Quote Prediction

These are different:

Quote count How many people have already quoted the post?
Quote prediction How likely is this viewer to quote the post?

The second is what fits into the recommendation scoring architecture.

Can Quotes Help a Post Reach More People?

A quote can be part of the positive engagement prediction used in ranking. But that doesn't mean: “Get 50 quotes and X will automatically boost the post.” The overall ranking score combines multiple predicted actions. A quote can also create a new post containing commentary about the original, meaning quote behavior can be connected to broader conversation and content discovery.

Under The Hood Takeaway

Quotes are independently modeled from reposts, making them a distinct engagement signal in the current recommendation architecture.

Official Open-Source References