AssetFlow IntelligenceResearch

Scorecard

Research · Scorecard · Coverage & data quality · Names

Every published interpretation names a claim about a later issue and carries the base rate for that claim, computed at the time it was logged from the same tables. When the later data arrive the claim resolves hit, miss or void, and the score is lift over the base rate — the hit rate minus the mean base rate of the claims made, in percentage points, with a 95% interval that resamples issue dates (claims logged the same day share the market day). Hit rate alone is not reported as a score. Brier skill compares the stated confidence (default 0.75) with the base rate as a forecast of the same outcomes; above zero means the view added information. Trading days are weekdays. Fewer than 5 issue dates gives no interval.

Overall
All modelsLoggedScoreableScoredVoidPendingWithheldMean base rateLift over base rate (95% interval)Brier skillDates
all1290090no scored claims yet0
By model
ModelLoggedScoreableScoredVoidPendingWithheldMean base rateLift over base rate (95% interval)Brier skillDates
anthropic/claude-fable-5.1430030no scored claims yet0
anthropic/claude-opus-5430030no scored claims yet0
google/gemini-3.8-flash430030no scored claims yet0
By publication and model
Publication · modelLoggedScoreableScoredVoidPendingWithheldMean base rateLift over base rate (95% interval)Brier skillDates
Editorial Attention · anthropic/claude-fable-5.1100000no scored claims yet0
Editorial Attention · anthropic/claude-opus-5100000no scored claims yet0
Editorial Attention · google/gemini-3.8-flash100000no scored claims yet0
Fed Language Tracker · anthropic/claude-fable-5.1110010no scored claims yet0
Fed Language Tracker · anthropic/claude-opus-5110010no scored claims yet0
Fed Language Tracker · google/gemini-3.8-flash110010no scored claims yet0
Regime Watch · anthropic/claude-fable-5.1110010no scored claims yet0
Regime Watch · anthropic/claude-opus-5110010no scored claims yet0
Regime Watch · google/gemini-3.8-flash110010no scored claims yet0
Sector Attention Breadth · anthropic/claude-fable-5.1110010no scored claims yet0
Sector Attention Breadth · anthropic/claude-opus-5110010no scored claims yet0
Sector Attention Breadth · google/gemini-3.8-flash110010no scored claims yet0
Claims
DatePublicationModelClaimBase rate at log timeOutcomeResolved
2026-09-02Sector Attention Breadthgoogle/gemini-3.8-flashCommunication Services will record a Top-5 share of gains of at least 31.9% in the next issue.
claim 0f857f72 · confidence 0.8
Population: share of a sector's names whose attention value rose versus the previous session (levels, day over day; flat names excluded from the denominator), sampled weekly as the last session on or before each 7-day step back from the as-of date; rank 1 = the highest share that week, rel = the share minus the median sector's share that week; sectors with at least 20 names, crypto, unclassified excluded (the published row rule). Sector is each name's current label applied to its whole history; 3 years to 2026-08-31. Given that the claim holds now, it held in the following issue too few to rate (n=0) of the time (n=0 transitions); unconditionally too few to rate (n=0) of 0. Subject's own record: 0 of 0 periods at or above 31.9. Window 2023-09-01 to 2026-08-31. No base rate (n=0); this claim is logged but cannot be scored as lift.pendinguntil 2026-09-09
2026-09-02Sector Attention Breadthanthropic/claude-opus-5We expect at most two sectors to print a Top-5 share of gains above 30% in the next issue.
claim b7370a9b · confidence 0.6
Population: rows of table 'breadth' in each earlier issue of 'sector_breadth_weekly' (published issues first, point-in-time snapshots fill gaps), up to 2026-08-31. Given that the claim holds now, it held in the following issue too few to rate (n=0) of the time (n=0 transitions); unconditionally too few to rate (n=0) of 0. Window no history. No base rate (n=0); this claim is logged but cannot be scored as lift.pendinguntil 2026-09-09
2026-09-02Sector Attention Breadthanthropic/claude-fable-5.1Communication Services' top-5 share of gains falls below 50% in the next issue.
claim 407724fb · confidence 0.65
Population: share of a sector's names whose attention value rose versus the previous session (levels, day over day; flat names excluded from the denominator), sampled weekly as the last session on or before each 7-day step back from the as-of date; rank 1 = the highest share that week, rel = the share minus the median sector's share that week; sectors with at least 20 names, crypto, unclassified excluded (the published row rule). Sector is each name's current label applied to its whole history; 3 years to 2026-08-31. Given that the claim does not hold now, it held in the following issue too few to rate (n=0) of the time (n=0 transitions); unconditionally too few to rate (n=0) of 0. Subject's own record: 0 of 0 periods below 50. Window 2023-09-01 to 2026-08-31. No base rate (n=0); this claim is logged but cannot be scored as lift.pendinguntil 2026-09-09
2026-09-02Regime Watchgoogle/gemini-3.8-flashGOOG will remain in attention state 3 through the next two weeks.
claim d617fe63 · confidence 0.6
57% (n=249) · c1/h3pendinguntil 2026-09-16
2026-09-02Regime Watchanthropic/claude-opus-5We expect both AMZN and GOOG to still carry state 3 a week from now.
claim a05bf50b · confidence 0.62
53% (n=249) · c1/h3pendinguntil 2026-09-09
2026-09-02Regime Watchanthropic/claude-fable-5.1AMZN remains in state 3 through the next two weeks.
claim 310ffb94 · confidence 0.6
57% (n=249) · c1/h3pendinguntil 2026-09-16
2026-09-01Fed Language Trackergoogle/gemini-3.8-flashIn his next press conference, Warsh will record a rate of mentions for 'balance sheet' of at least 11.5 per 100k characters.
claim 362de0e8 · confidence 0.6
16% (n=322) · c1/h1pendinguntil 2026-10-31
2026-09-01Fed Language Trackeranthropic/claude-opus-5At the next press conference Warsh's 'inflation expectations' rate stays below 6.0 per 100k characters of his own words.
claim cb334122 · confidence 0.79
76% (n=246) · c1/h1pendinguntil 2026-10-31
2026-09-01Fed Language Trackeranthropic/claude-fable-5.1At the next press conference, Warsh's 'balance sheet' rate stays at or above 5.0 per 100k characters.
claim a68d5768 · confidence 0.7
80% (n=303) · c1/h1pendinguntil 2026-10-31
Views without a scoreable claim
These interpretations were published but carry no scoreable claim — most often because every claim the writer could make resolves the same way nearly all the time, so no lift could be measured. They count in the logged column, never in lift.
DatePublicationModelWhy
2026-09-01Editorial Attentiongoogle/gemini-3.8-flashno claim menu
2026-09-01Editorial Attentionanthropic/claude-opus-5no claim menu
2026-09-01Editorial Attentionanthropic/claude-fable-5.1no claim menu
Withheld interpretations
None.
Earlier interpretations
13 interpretation(s) were logged before every view carried a matched control. They are listed here unscored and enter no aggregate.
DatePublicationClaim as written
2026-09-01Editorial Attention
2026-09-01Fed Language Tracker
2026-09-01Regime Watch
2026-09-01Fed Language Tracker
2026-09-01Sector Attention Breadth
2026-09-01Regime Watch
2026-09-01Editorial Attention
2026-09-01Fed Language TrackerWe expect Warsh to maintain balance sheet mentions above 10.0 per 100,000 characters and zero mentions of inflation expectations at the next FOMC press conference.
2026-09-01Sector Attention BreadthTechnology attention will fail to achieve a 50.0% advancing majority in the next weekly update as public interest stays clustered around leading mega-caps.
2026-09-01Editorial AttentionEditorial restructuring on Apple's entry will keep its revert share below 10.00% in the following monthly snapshot.
2026-09-01Fed Language TrackerAt the next press conference, Warsh will continue to register zero mentions of 'inflation expectations' as he anchors policy communication around balance-sheet dynamics.
2026-09-01Sector Attention BreadthTechnology sector attention advancing share will remain below 50.0% in the next session as crowd attention remains anchored to headline mega-caps.
2026-09-01Regime WatchBoth AMZN and GOOG will maintain state 3 web attention through the next two weekly reporting cycles.
17 superseded row(s) hidden: a re-run of the same issue by the same model replaces its earlier view; only the last view logged before the data arrived is live.
Scorer s1 · built 2026-09-02T22:33:26
Published by AssetFlow.ai LLC. Descriptive research on attention and public-record data; not investment advice and not an offer or solicitation. No return claim attaches to anything above. Issue date 2026-09-02.