Pre-computed step-by-step influence sequence for each candidate. Phases are ordered by cascade dependency — complete Phase 1 before Phase 2 unlocks. Probability impacts are anchored on Grade-A historical effect sizes (see methodology card). Status tracks execution progress.
/api/v1/invisible-primary/ops/compute-path, which overrides the editorial sequences when reachable. In offline mode, the editorial sequences are shown. The “+X–Ypts” impact ranges per step cite the historical cascade that anchors them (e.g. Biden 2020 SC-Clyburn = +35–40pt, Obama 2007 Hollywood = +8–12pt Q1 fundraising). These are analogues, not 2028-specific forecasts.
docs/EFFECT_SIZES.md as of …
docs/CAUSAL_GRAPH.json as of …
docs/EFFECT_SIZES.md Tables 0–6, docs/CAUSAL_GRAPH.json. Replaceable with live Dijkstra solver output when backend operational.
Track live influence operations across Planning → In Progress → Executed → Measuring stages. Each card represents a specific actor engagement. Cascade dependencies are shown — some operations cannot start until predecessors complete. Drag cards between columns as operations advance. Click any card to expand the full action brief.
Real-time nomination probability for all candidates, updated as operations are logged. The 'what would move the needle most' section below each bar is the key action signal — it identifies the single highest-expected-value intervention available right now. Use 'Scenario' to model what happens if a specific actor commits to any candidate.
Log and interpret incoming signals — endorsements, public statements, donor meetings, fundraising reports, polling shifts. Each signal is assessed for its downstream cascade effect on nomination probabilities. Use the manual entry form to log new intelligence; the system will recompute affected probabilities automatically.
/api/v1/invisible-primary/ops/intelligence-feed; in snapshot mode (shown here), three editorial signals are pinned to the 2026-04-12 baseline. 2026-05-17 Phase 8.2.2 Pod 3: methodology card surfaced inline (was buried in the data-integrity footer, violating E3 — cards must live next to the numbers they explain).
Source: backend/src/invisible_primary/ops_routes.py::intelligence_feed · effect-size anchors: EFFECT_SIZES.md Tables 0–6 (Battaglini et al. 2024, NBER WP 32649; Cohen-Karol-Noel-Zaller 2008).
Comprehensive strategic intelligence document for each candidate — synthesizes Ferguson model position, actor network status, threat landscape, and 30/90-day action priorities into a single decision-support brief. Update when: a signal donor commits, a rival secures a major endorsement, or Q-filing deadlines pass. Generated by Catalyst Intelligence Engine using Ferguson Model v2.1.
/api/v1/dark-money/funding-by-race).
All cited URLs verified clickable 2026-05-17.
| Race | State | FEC (R) | Dark (R) | Total R | FEC (D) | Dark (D) | Total D | Grand Total | Dark % | R/D Balance |
|---|
| OPERATIVE ↕ | FIRM ↕ | PARTY ↕ | RACES ↕ | TOTAL FUNDS ↕ | WIN RATE ↕ | TOP 3 RACES |
|---|
For each viable candidate, we estimate donor concentration (share of past funding from a single top backer), replaceability (probability of finding a comparable backer within one cycle, baseline 19.4%), and the projected drop in win probability if their top donor departed tomorrow (Battaglini baseline −2.8pp, scaled by concentration and race competitiveness; up to 10× in close races). Risk tier ranges from 🟢 Resilient to 🔴 Critical. Watchlist below is exportable to CSV.
| CANDIDATE ↕ | TOP DONOR ↕ | CONCENTRATION ↕ | REPLACEABILITY ↕ | PROJECTED Δ WIN PROB ↕ | RISK TIER ↕ | RECOMMENDATION |
|---|