Listening for the next measured pulse
The view is fetching the current MarketSchema Attention snapshot. Layout stays stable while the data resolves.
Attention per market-cap dollar. Proof first.
Organic score summarizes coordination tells. Independent voices discounts repeated co-movement among hashed authors.
Awaiting blinded operator labels.
A 15-year, $9.8 billion lease with an investment-grade tenant sent $HUT surging 14% and pulled $IREN and other compute-focused miners sharply higher. With the unnamed tenant and still-unbuilt 352 MW phase, can peers replicate the deal? https://t.co/xm3rs15mKD
Official X API · official x ready · no handle shownNo independently adjudicated manufactured-crowd case is available yet. This public-post observation is evidence, not a labeled outcome or verdict.55 names · deterministic stale-first refresh · quota-gated expansion
MarketSchema Attention measures a bounded 55-name watchlist, not the whole market. The universe widens quota-gated, not on a fixed calendar: each cycle refreshes a deterministic sub-universe ordered by staleness, and a disclosed co-mention rule promotes a capped set of repeatedly co-mentioned names. Auto-broadening runs only inside the operator X-fetch budget — unset, coverage holds at the fixed watchlist. Every cycle publishes a coverage breadth receipt, so dark names are explicit evidence, never a silent gap.
Each point carries an as-of timestamp. Public: tape · /v1/snapshot/leaderboard/history · deeper history: developers. The room is locked to Table · Dig · Scorecard · API. Every other surface is footer or detail.
| KPI | Formula | Unit | Direction | Inputs | Bot use |
|---|---|---|---|---|---|
| Attention densityattention_density | WSA / (market_cap / 1e9) | WSA per $1B market cap | higher means more attention per market-cap dollar | weighted_social_activity, market_cap | Rank abnormal attention per dollar of size. |
| Abnormalityseasonal_z | shrunk seasonal z (WSA vs ticker×dow×hour cell, shrunk toward the ticker session-phase parent, then ticker parent) + basis (cell/phase_shrunk/shrunk), shrinkage weights, borrowed_from_parent flag, structural-break pre_break_excluded receipt, baseline depth status, maturity ratios, observations remaining, parent tier, and served-abnormality-framing receipt with absolute_level/seasonal_z/dominant_basis | seasonal z-score object | farther from zero means farther from that ticker's baseline | weighted_social_activity, seasonal cells (archive) | Detect departures from a name's own baseline. |
| Evidence coverageattention_integrity | weight-renormalized mean of organic, dispersion (1−gini), corroboration (eff_sources/3), persistence (clamp autocorr), and ready independent-voices independence_ratio, plus measured author-graph co-occurrence independence (1−suspicious_pair_share); degraded/blocked independent-voices receipts are ignored; blocked/unmeasured author-graph receipts are ignored; None < 2 components | 0 to 1 score | higher means more measured corroboration, dispersion, and persistence | organic_score, attention_concentration, cross_source_corroboration, attention_persistence, independent_voices, author_graph_evidence | Gate how far to trust a reading. |
| Anti-gaming flagspump_risk_score | 0.25·spam_cluster+0.20·low_float+0.20·low_liquidity+0.15·low_source_quality+0.10·promotional_language+0.10·price_extended; missing float/liquidity/source inputs on small or illiquid names use a conservative floor and are flagged under-measured | 0 to 1 score | higher means more observable anti-gaming flags are present | spam_score, market_cap, float_market_cap, avg_dollar_volume, source_quality_score, recent_dilution | Computed diagnostic only; the public real/manufactured claim is withheld pending validation. |
| Liquidity pressureliquidity_pressure | WSA / (avg_dollar_volume / 1e9) (the crowding ratio) | WSA per $1B average dollar volume | higher means more attention relative to trading capacity | weighted_social_activity, avg_dollar_volume | Size attention against tradable capacity. |
attention_densityseasonal_zattention_integritypump_risk_scoreliquidity_pressure| Metric | Formula | Inputs | Where shown |
|---|---|---|---|
| Velocity (acceleration z-score)attention_density/scoring.py (zscore/velocity) | z-score of current weighted social activity vs its own trailing baseline: (WSA − mean(baseline)) / std(baseline), sample std (n−1); null under 2 baseline points. Label bands: heating ≥ +1σ, cooling ≤ −1σ, else steady. | weighted_social_activity (current), trailing WSA baseline | home velocity chip ("+2.4σ heating"), Markets "Velocity" column |
| Raw density (Posts/$B)attention_density/scoring.py (raw_density) | raw_post_count / (market_cap / 1e9) — raw source posts per $1B market cap; null unless raw_post_count ≥ 0 and market_cap > 0. | raw_post_count, market_cap | Markets "Posts/$B" column |
| Engagement-weighted density (Eng/$B)attention_density/scoring.py (engagement_weighted_density) | engagement_sum / (market_cap / 1e9) — observed public interactions per $1B market cap; null unless engagement_sum ≥ 0 and market_cap > 0. | engagement_sum (public interactions), market_cap | Markets "Eng/$B" column |
| Silent-to-loud triggerattention_density/pipeline/events.py + ignition.py | Fires when velocity_zscore ≥ 3.0 AND source_quality_score ≥ 0.60 (both measured). The page's continuous ignition ratio = current WSA / median(baseline); bands ignition ≥ 10×, surge ≥ 3×, elevated ≥ 1.5×, else normal (null under 3 baseline points). | velocity_zscore, source_quality_score, pump_risk_score (confidence discount) | engine event feed (the /silent-to-loud page is retired; the measurement lives in /v1/signals) |
| Attention-price divergence eventattention_density/pipeline/divergence.py | With rising = velocity_z ≥ 2.0: REVERSAL_RISK when rising and price_return < −2%; ATTENTION_LEADS when rising and price_return ≤ +2%; NOT_SOCIAL when velocity_z ≤ 0.5 and price_return ≥ +2%. Only these three divergent labels emit an event (CHASE_RISK and ALIGNED do not). Strength = clamp(0.5 + 0.25·attention_excess + 0.25·price_component). | velocity_zscore, recent price_return, pump_risk_score (confidence) | engine event feed (the /attention-price-divergence page is retired; the measurement lives in /v1/signals) |
| Wall grade (Moves)attention_density/share_queue.py (WALL_GRADE_PUMP_RISK_THRESHOLD) | The archived wall-grade rule is pump_risk_score ≥ 0.35. It is not currently a public verdict: labeled-outcome predictive validity is not yet established. | pump_risk_score | Moves at /caught-pumps ("wall-grade flags") |
attention_density/scoring.py (zscore/velocity)attention_density/scoring.py (raw_density)attention_density/scoring.py (engagement_weighted_density)attention_density/pipeline/events.py + ignition.pyattention_density/pipeline/divergence.pyattention_density/share_queue.py (WALL_GRADE_PUMP_RISK_THRESHOLD)