← Back
Optimization Lab Research Control Center Champion Governance

Optimization Dashboard

Model calibration, search execution, experiment tracking, and validation governance for the BettingOdds projection lab.
Dashboard Experiments Models Validation Leaderboard Market Data
Operator
Manual Runs
Edge Run Operator
Execute deterministic edge runs from persisted market snapshots and feed the Edge Terminal with live opportunity data.
Snapshot → Model Join → Edge Persistence
Active Model
structured-0010-0003
Version ID 45
Model Versions
114
Stored model versions in the optimization lab.
Experiments
18
Total optimization runs created across research history.
Candidates Tested
101
Challenger models evaluated across all experiments.

Lab Control Surface

Launch research runs, review model inventory, and move candidates into governance.
Grid Random Structured Adaptive
Create Experiment
Run Grid Search
Run Random Search
Open Validation
Research Guidance
Use Grid for tightly bounded exhaustive searches, Random for broader bounded exploration, and Structured Adaptive for deterministic champion-centered refinement with smarter candidate allocation.

Lab Posture

Governed
Champion Discipline
Production model changes should pass through experiment ranking, validation evidence, and final governance review.
Optimization Philosophy
The lab is designed for controlled convergence, not ad-hoc tuning. Stronger models should emerge through versioned, reproducible research loops.
Best Practice
Prefer challenger validation before promotion. A strong experiment leaderboard is a research result, not automatic production approval.

Research Workflow

Standard operating flow for the BettingOdds optimization lab.
Lifecycle
1. Generate Challengers
Use bounded grid, random, or structured adaptive search to create candidate model versions.
2. Rank Experiment Results
Identify the strongest challengers under the target metric and inspect secondary signals.
3. Validate Against Champion
Run walk-forward evaluation and holdout checks before governance decisions.
4. Govern Promotion
Apply policy gates, reviewer judgment, and audit-trail decisions before activation.

Navigation Shortcuts

Jump directly into the major research surfaces.
Shortcuts
Experiments
Review search runs, winners, and candidate leaderboards.
Models
Inspect version inventory, active champion, and model lineage.
Validation
Review governance runs, holdout outcomes, and final decisions.
Leaderboard
Compare strongest candidates and research winners across runs.

Recent Experiments

Most recent optimization runs across the BettingOdds research lab.
10 experiment(s)
Dashboard Reading Guide
Use recent experiments to monitor lab activity, identify active search lanes, and jump directly into candidate analysis and validation preparation.
Experiment Status Metric Candidates Created Actions
Optimization 2026-03-31 22:21 Draft MinutesMae 5 2026-03-31 View
Minutes Calibration Sprint 01 | Run 8 | 2026-03-31 22:14 Running MinutesMae 7 2026-03-31 View
Minutes Calibration Sprint 01 | Run 7 | 2026-03-24 21:12 Completed MinutesMae 7 2026-03-24 View
Minutes Calibration Sprint 01 | Run 6 | 2026-03-24 19:45 Completed MinutesMae 6 2026-03-24 View
Minutes Calibration Sprint 01 | Run 5 | 2026-03-24 18:05 Completed MinutesMae 6 2026-03-24 View
Minutes Calibration Sprint 01 | Run 4 | 2026-03-24 17:27 Completed MinutesMae 6 2026-03-24 View
Minutes Calibration Sprint 01 | Run 3 | 2026-03-24 16:27 Completed MinutesMae 6 2026-03-24 View
Minutes Calibration Sprint 01 | Run 2 | 2026-03-24 15:03 Completed MinutesMae 6 2026-03-24 View
Structured Adaptive Minutes 3 Completed MinutesMae 12 2026-03-11 View
Structured Adaptive Minutes 2 Completed MinutesMae 3 2026-03-11 View