---

name: x-algo-engagement description: Reference for X algorithm engagement types and signals. Use when analyzing engagement metrics, action predictions, or understanding what signals the algorithm tracks.

X Algorithm Engagement Signals

The X recommendation algorithm tracks 18 engagement action types plus 1 continuous metric. These are predicted by the Phoenix ML model and used to calculate weighted scores.

PhoenixScores Struct

Defined in home-mixer/candidate_pipeline/candidate.rs:

Copypub struct PhoenixScores {
    // Positive engagement signals
    pub favorite_score: Option<f64>,
    pub reply_score: Option<f64>,
    pub retweet_score: Option<f64>,
    pub quote_score: Option<f64>,
    pub share_score: Option<f64>,
    pub share_via_dm_score: Option<f64>,
    pub share_via_copy_link_score: Option<f64>,
    pub follow_author_score: Option<f64>,

    // Engagement metrics
    pub photo_expand_score: Option<f64>,
    pub click_score: Option<f64>,
    pub profile_click_score: Option<f64>,
    pub vqv_score: Option<f64>,              // Video Quality View
    pub dwell_score: Option<f64>,
    pub quoted_click_score: Option<f64>,

    // Negative signals
    pub not_interested_score: Option<f64>,
    pub block_author_score: Option<f64>,
    pub mute_author_score: Option<f64>,
    pub report_score: Option<f64>,

    // Continuous actions
    pub dwell_time: Option<f64>,
}

Action Types by Category

Positive Engagement (High Value)

Sharing Actions

Engagement Metrics

Negative Signals

Continuous Actions

How Scores Are Obtained

The PhoenixScorer (home-mixer/scorers/phoenix_scorer.rs) calls the Phoenix prediction service:

1. Input: User history + candidate posts
2. Output: Log probabilities for each action type per candidate
3. Conversion: probability = exp(log_prob)

Copyfn extract_phoenix_scores(&self, p: &ActionPredictions) -> PhoenixScores {
    PhoenixScores {
        favorite_score: p.get(ActionName::ServerTweetFav),
        reply_score: p.get(ActionName::ServerTweetReply),
        retweet_score: p.get(ActionName::ServerTweetRetweet),
        // ... maps each action to its probability
    }
}

Signal Interpretation

• Scores are probabilities (0.0 to 1.0): P(user takes action | user sees post)
• Higher = more likely: A favorite_score of 0.15 means 15% predicted chance of like
• Negative signals have negative weights: High report_score reduces overall ranking
• VQV requires minimum video duration: Only applies to videos > MIN_VIDEO_DURATION_MS

Related Skills

• /x-algo-scoring - How these signals are combined into a weighted score
• /x-algo-ml - How Phoenix model predicts these probabilities