Mark/mesh-seed

Fork 0

[Evolution] Weekly championship tournament #15

New issue

Closed

opened 2026-03-10 20:50:56 +00:00 by Mark · 2 comments

Mark commented

2026-03-10 20:50:56 +00:00

Owner

Tournament system for capability competition.

Functionality

Run weekly (Friday 2pm)
Defenders: current stable capabilities
Challengers: experimental variants
Sample historic problems as test set
Promote winners, demote losers

Files

evolution/championship.py (enhance)
logs/championship.log

Output

Tournament results
Registry updates
Performance metrics

See examples/task-complex.md for impact

Tournament system for capability competition. ## Functionality - Run weekly (Friday 2pm) - Defenders: current stable capabilities - Challengers: experimental variants - Sample historic problems as test set - Promote winners, demote losers ## Files - evolution/championship.py (enhance) - logs/championship.log ## Output - Tournament results - Registry updates - Performance metrics See examples/task-complex.md for impact

Mark referenced this issue

2026-03-14 18:59:12 +00:00

[Phase 1] Evolution - Self-Improvement #28

Mark commented

2026-03-14 19:11:47 +00:00

Author

Owner

Phase Dependency Mapping

Linked to Phase Tracker: #28 (Phase 1: Evolution)

Prerequisites:

#14 (Daily rumination) - feeds championship data
#10 (Intent classification) - evaluates agent performance
#18 (Model spec analysis) - identifies fine-tuning candidates

Blocks: None directly, but enables Phase 2 specialization

Acceptance Criteria:

Weekly automated tournament runs
Performance metrics collected from all workers
Winner selection algorithm implemented
Model promotion/demotion workflow

Estimated Complexity: Medium (4-6 days)
Priority: P2 (after rumination stable)

## Phase Dependency Mapping **Linked to Phase Tracker:** #28 (Phase 1: Evolution) **Prerequisites:** - #14 (Daily rumination) - feeds championship data - #10 (Intent classification) - evaluates agent performance - #18 (Model spec analysis) - identifies fine-tuning candidates **Blocks:** None directly, but enables Phase 2 specialization **Acceptance Criteria:** - Weekly automated tournament runs - Performance metrics collected from all workers - Winner selection algorithm implemented - Model promotion/demotion workflow **Estimated Complexity:** Medium (4-6 days) **Priority:** P2 (after rumination stable)

Mark referenced this issue

2026-03-14 19:11:55 +00:00

[Integration Test] End-to-end: intent → execution → result #8

Mark referenced this issue

2026-03-14 19:12:04 +00:00

[MVP] Worker capability: writer_v1 #12

Mark referenced this issue

2026-03-14 19:12:11 +00:00

[Model Spec] Analyze for fine-tuning candidates #18

Mark referenced this issue

2026-03-14 19:12:48 +00:00

[Meta] Create project milestones and organize roadmap #26

Mark referenced this issue

2026-03-14 20:01:09 +00:00

[Phase 0] Foundation - MVP Completion #27

Mark closed this issue

2026-03-14 20:17:30 +00:00

Mark commented

2026-03-14 20:17:42 +00:00

Author

Owner

✅ COMPLETED - Phase 1 Evolution

Implementation Summary:

Migrated from JSON to SQLite ledger (MeshLedger)
Contestants, history, results, pathways in ledger tables
Real capability execution via runtime.capability_execution
Fallback to simulation if execution unavailable
Tournament runs: Friday 2pm via cron
10-20 historic problems sampled from ledger
Promotion: >5% improvement over defender
Demotion: <0.7 success rate
Logs to logs/championship.log with match details

Cron Schedule:

0 14 * * 5 cd /path/to/mesh-seed && python evolution/championship.py

Files Changed:

evolution/championship.py (SQLite, real execution, logging)
tests/evolution/test_championship.py (new)

Verification:

pytest tests/evolution/test_championship.py -v
# Tests passing (scoring, sampling, promotion/demotion)

Closed as part of Phase 1 completion.

## ✅ COMPLETED - Phase 1 Evolution **Implementation Summary:** - Migrated from JSON to SQLite ledger (MeshLedger) - Contestants, history, results, pathways in ledger tables - Real capability execution via runtime.capability_execution - Fallback to simulation if execution unavailable - Tournament runs: Friday 2pm via cron - 10-20 historic problems sampled from ledger - Promotion: >5% improvement over defender - Demotion: <0.7 success rate - Logs to logs/championship.log with match details **Cron Schedule:** ```bash 0 14 * * 5 cd /path/to/mesh-seed && python evolution/championship.py ``` **Files Changed:** - evolution/championship.py (SQLite, real execution, logging) - tests/evolution/test_championship.py (new) **Verification:** ```bash pytest tests/evolution/test_championship.py -v # Tests passing (scoring, sampling, promotion/demotion) ``` **Closed as part of Phase 1 completion.**

Mark referenced this issue

2026-03-14 20:17:58 +00:00

[Phase 1] Evolution - Self-Improvement #28

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

Mark/mesh-seed#15

No description provided.

Rows
Columns