Closed-loop intelligence: settled bets feed back into grading weights.
- Grade accuracy tracking per stat type (A/B/C/D hit rates)
- Signal accuracy tracking (which deltas predict outcomes)
- Kill condition effectiveness (hit_rate_with vs without)
- Conservative weight adjustment (20% cap, 50-pick minimum)
- 4 new DB tables: grade_accuracy, signal_accuracy,
kill_condition_accuracy, weight_history
- Desk-tier endpoints: /api/model/accuracy, /api/model/insights
Spec complete, ready to build when Phase 3 deployment is stable.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Next.js 14+ web app in web/ directory:
- Landing page: Hero, How It Works, Features, 3-tier Pricing with
founder badges, Footer with email capture
- Blog system: MDX-powered, /blog index + /blog/[slug] pages,
reading time, Open Graph tags, JSON-LD structured data
- Auth pages: /login + /signup (Supabase Auth ready)
- Design system: dark theme, grade colors (A/B/C/D), BetonBLK voice
- 1 seed blog post: "How to Read Line Movement Like a Sharp"
- Specs for 3.2 (Scan UI), 3.3 (Bet Tracker), 3.4 (Stripe)
Build passes clean: 7 static pages generated.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three submission methods:
- POST /api/bets/quickslip — structured bet entry
- POST /api/bets/screenshot — stub OCR with confirm flow
- POST /api/bets/sync — coming soon stub
Full bet lifecycle:
- PATCH /api/bets/:id/settle — settle with outcome, recalculates performance
- GET /api/bets — list with status/book/pagination filters
- GET /api/bets/performance — ROI, win rate, profit (weekly/monthly/all_time)
Payout calculator handles straight bets (American odds) and parlays
(multiplied leg payouts). Performance service recalculates on each
settlement and upserts into performance table.
33 new tests, 221 total (194 Node.js + 27 Python), all passing.
All backend features for Phase 1 + Phase 2 now complete.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Line movement system:
- Baseline capture on first odds fetch of the day
- Movement detection >= 0.5 points with direction (up/down)
- Sharp money heuristic (sharp_action/public_action/unknown)
- GET /api/movements with player, stat_type, min_movement filters
- Movements included in GET /api/odds/nba live responses
Cascade detection system:
- Scratch detection: player props disappear from 2+ books
- Affected user lookup via scan_sessions + picks
- Parlay re-grade without scratched legs
- cascade_alerts created for affected users
- GET /api/alerts (Analyst/Desk only), PATCH /api/alerts/:id/read
Zero extra Odds API credits — all detection piggybacks on existing fetches.
Migration 002: line_baselines, line_movements, cascade_alerts tables.
30 new tests, 188 total (161 Node.js + 27 Python), all passing.
Phase 2 Core Product COMPLETE.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Core intelligence for BetonBLK prop analysis:
- POST /api/analyze/prop — single prop analysis
- POST /api/analyze/batch — multi-prop analysis for parlay scanner
- 6-step pipeline: season avg → recent form → situational splits →
cross-book lines → kill conditions → grade (A/B/C/D)
- 6 kill conditions: low_minutes, small_sample, b2b_high_usage,
blowout_risk, split_conflict, no_opponent_data
- Composite scoring with confidence (30-95), bonuses, penalties
- Added spreads market to Odds API fetch (zero extra credits)
- Full reasoning output with step-by-step breakdown
36 new tests (unit + integration), 128 total across all features
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
builtbykev