x-algo-engagement
Original：🇺🇸 English
Translated
Reference for X algorithm engagement types and signals. Use when analyzing engagement metrics, action predictions, or understanding what signals the algorithm tracks.
1installs
Sourcecloudai-x/x-algo-skills
Added on
NPX Install
npx skill4agent add cloudai-x/x-algo-skills x-algo-engagementTags
Translated version includes tags in frontmatterSKILL.md Content
View Translation Comparison →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.rsrust
pub 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)
| Action | Proto Name | Description |
|---|---|---|
| Favorite | | User likes the post |
| Reply | | User replies to the post |
| Retweet | | User reposts without comment |
| Quote | | User reposts with their own comment |
| Follow Author | | User follows the post's author |
Sharing Actions
| Action | Proto Name | Description |
|---|---|---|
| Share | | Generic share action |
| Share via DM | | User shares via direct message |
| Share via Copy Link | | User copies link to share externally |
Engagement Metrics
| Action | Proto Name | Description |
|---|---|---|
| Photo Expand | | User expands photo to view |
| Click | | User clicks on the post |
| Profile Click | | User clicks author's profile |
| VQV | | Video Quality View - user watches video for meaningful duration |
| Dwell | | User dwells (pauses) on the post |
| Quoted Click | | User clicks on a quoted post |
Negative Signals
| Action | Proto Name | Description |
|---|---|---|
| Not Interested | | User marks as not interested |
| Block Author | | User blocks the author |
| Mute Author | | User mutes the author |
| Report | | User reports the post |
Continuous Actions
| Action | Proto Name | Description |
|---|---|---|
| Dwell Time | | Continuous value: seconds spent viewing post |
How Scores Are Obtained
The () calls the Phoenix prediction service:
PhoenixScorerhome-mixer/scorers/phoenix_scorer.rs- Input: User history + candidate posts
- Output: Log probabilities for each action type per candidate
- Conversion:
probability = exp(log_prob)
rust
fn 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 of 0.15 means 15% predicted chance of like
favorite_score - Negative signals have negative weights: High reduces overall ranking
report_score - VQV requires minimum video duration: Only applies to videos >
MIN_VIDEO_DURATION_MS
Related Skills
- - How these signals are combined into a weighted score
/x-algo-scoring - - How Phoenix model predicts these probabilities
/x-algo-ml