Pokémon TCG — Current State Reconciliation and Provenance Ledger
As-of-commit reconciliation of live code, database evidence, handoff blueprints, tests, transcripts and Wikifita claims.
Pokémon TCG — Current State Reconciliation and Provenance Ledger
Verified snapshot
This page is the current-state anchor for the project. It is not a replacement for the dated timeline or historical experiment pages.
| Surface | Verified value | Role |
|---|---|---|
| Project branch | develop | Live implementation checkout |
| Project HEAD | 20d7d0d | Adds AGENTS.md, the Odyssey Diary and convergence contract |
| Project commit count | 144 reachable commits at HEAD | Direct git rev-list --count HEAD; the older Wiki figure of 147 is not reproducible at this snapshot |
| Wikifita branch | main | Public knowledge-base checkout |
| Previous Wikifita incorporation anchor | 383a0ea | Completed corpus/provenance pass before this coherence pass |
| Source tree | Clean before this documentation pass | No project-code change authorized or made |
The repository and Wikifita are distinct Git histories. A project commit hash proves source evolution; a Wikifita commit hash proves editorial incorporation. Neither replaces the other.
Evidence hierarchy for current claims
For an implementation statement, use this order:
- live source code at a named project commit;
- executable artifacts, logs, database and dataset probes tied to that commit or snapshot;
- project handoff and design documents;
- Antigravity reports and transcript artifacts;
- older Wikifita pages and user interpretations.
The lower layers preserve intent and reasoning. They cannot override a direct absence in the current source tree.
Current implementation boundary
| Area | Current verified state | Provenance |
|---|---|---|
| MLX training | scripts/bc/bc_train_mlx.py is FP32-native: it calls model.set_dtype(mx.float32) and rejects non-FP32 model leaves | Project HEAD 20d7d0d, source lines around the trainer initialization |
| PyTorch inference | rl/policy_infer_torch.py converts and validates floating tensors as strict torch.float32 | Project HEAD 20d7d0d |
| Runtime action selection | agent/main.py recomputes logits for each multi-select substep, masks picked options and carries memory_out into the next decision | Project HEAD 20d7d0d |
| Temporal memory | Scratch registers provide a memory interface; memory_in and memory_out are carried per tracker state, with TBPTT lane isolation | rl/policy_mlx.py, rl/policy_infer_torch.py, agent/main.py |
| Core architecture | Entity/action Transformer, d_model=128, four heads, three-layer default or four-layer session configuration, FFN 512 | Current source/config; exact run config wins for a checkpoint |
| Position encoding | Typed zones, scalar/meta features and explicit source/target gathers; no tracked RoPE-ND implementation | Direct source-tree inventory |
| MoE / vehicle / Apex | No tracked rl/ropend, rl/moe, policy_moe_*, vehicle_draft.py or Apex runtime path | Direct source-tree inventory |
| Database | SQLite Schema 2.0.0, populated but not clean: foreign-key audit reports 2,946,336 violations | Read-only database probe and pokemon_tcg_aug14_data_etl_database_audit |
The FP16 trainer and FP16 PyTorch artifact remain valid historical delivery records. They are not the current source contract after the FP32 correction.
Read-only database revalidation
The native ResultsDB API was used without network synchronization or mutation. The current snapshot returns:
| Table or metric | Value |
|---|---|
matches | 139,783 |
match_steps | 3,065,706 |
agents / decks | 1,117 / 948 |
days / datasets | 30 / 30 |
agent_elo_daily / deck_elo_daily | 28,745 / 10,742 |
tournaments / matchups | 128 / 1,621 |
| direct source split | 138,023 remote / 1,760 local matches |
| Deck #633 local invariant Elo | 412.21 over 420 games; MD10 complete |
| Deck #251 local invariant Elo | 577.39 over 420 games; MD10 complete |
The invariant Elo values are database observations under the current ResultsDB.get_invariant_deck_elo implementation. They do not replace the deck-conditioned tournament win rates, and they should not be interpreted without the source, season and matchup context.
Handoff milestone status
The August 14 handoff commit f988eb0 added PROJECT.md, TEST_INFRA.md, .agents/ survey artifacts, scratch probes and a Tier 1 E2E test file. It did not add the proposed RoPEND, MoE, vehicle-draft or Apex modules.
| Milestone | Status in the project handoff | Direct reconciliation |
|---|---|---|
| M1 RoPEND/MoE | IN_PROGRESS; gate PENDING | Blueprint/explorer material exists; implementation modules are absent |
| M2 elite dataset/DB parity | IN_PROGRESS; survey phase open | No purge/rebuild executed; FK and ZIP/DB deltas remain |
| M3 PageRank/Abelian monograph | IN_PROGRESS; gate not passed | Documents exist; the claimed implementation-level isomorphism is not verified here |
| M4 Wikifita sync | PLANNED in the project plan | Wikifita integration has progressed externally, but this does not rewrite the project plan's historical milestone status |
| M5 E2E/tournament hardening | PLANNED | No verified 500-match acceptance run with the stated >40% Yan target |
| E2E test writers | Initialized or dispatched | Progress files retain unchecked execution and handoff boxes; generated tests are not proof of passing gates |
“In progress” in PROJECT.md or .agents/ means orchestration status. It does not mean that the source modules or acceptance gates exist.
Post-blueprint Git lineage
These commits belong in the project timeline and explain the current handoff state:
| Commit | Change | Evidence interpretation |
|---|---|---|
f988eb0 | Added blueprints, milestone scopes, survey artifacts, scratch probes and E2E test scaffolding | Planning and exploration artifact; not delivery proof |
ea9317a | Ignored SQLite binary sidecars and synchronized GEMINI.md | Repository hygiene plus memory-state edit |
dbf298a | Recorded three-tier idempotency sync, bulk Elo decoupling and KaTeX directive changes | Small source/memory refactor; requires code inspection for behavior claims |
14ec607 | Updated channel/rules files, ResultsDB and card-stat build path; added follow-me rule | Operational rule and ETL/Elo maintenance lineage |
20d7d0d | Added the Odyssey Diary and multi-agent convergence contract in AGENTS.md | Current project instruction/handoff boundary |
The earlier documentation pass stopped its provenance table at 12595d4. This ledger closes that gap without rewriting the historical claims those pages recorded.
Antigravity transcript and brain cross-check
The August 14 handoff was not reconstructed from one summary. The main
Antigravity session 9189fa2e-93c2-4a04-9bf0-6d090880de27 and its linked
brain artifacts were checked against the repository copies and milestone
ledgers. The portable transcript location is
~/.gemini/antigravity-cli/brain/<session-id>/.system_generated/logs/, with
both compact and full JSONL variants. The repository handoff also identifies
the deck analyst 687d75b1-19ff-4999-a56e-c5d40d869f0f, migration engineer
245904ec-c5c7-4dca-a796-e04db0afb510, sentinel
da74ca6c-c155-4353-82d1-5c5695a60da1, orchestrator
cd851a4f-6875-4819-9f25-1b23dd14cc1b, and M1/M2/M3 milestone sessions.
The cross-check resolves the latest handoff into four evidence classes:
| Brain or transcript material | What it establishes | What it cannot establish |
|---|---|---|
fp16_precision_crisis.md and the migration transcript | A reported FP16 export collapse and an FP32 restoration path | That every historical run used the same artifact or that the report replaces live code |
ablation_matrix_results.md | The 420-game Stage 1-4 FP32 matrix and deck-conditioned peaks | That its rows share one database lifecycle or one later tournament cohort |
project_architecture_and_state.md | The Stage 3 loss-scale narrative, Stage 4/5 chronology and scratch-register interpretation | That scratch registers causally absorbed the shock; this remains an interpretation |
master_project_state_and_handoff_plan.md, .agents/ transcripts and milestone ledgers | The intended M1-M5 topology, acceptance criteria and handoff state | Completion of absent modules, gates, parity or the 500-match threshold |
Recent transcript decisions are also preserved in pokemon_tcg_cross_harness_provenance: bulk/idempotent ETL, 28-hour leaderboard TTL, separate Elo recomputation, invariant-Elo preference, docs-only continuation after the quota boundary, and explicit preservation of future architecture. These are decisions and operating constraints; each implementation claim still resolves through the live source and artifact hierarchy above.
Contradictions resolved in Wikifita
The following were genuine stale-reference problems, not legitimate experiment cohorts:
- FP16 versus FP32: FP16 is retained as the July 25 and August 7 delivery history; strict FP32 is the current source contract.
- Scratch state: current inference carries
memory_out; the old statement that scratch outputs are discarded is no longer current. - Multi-select: current
agent/main.pyperforms autoregressive substeps; the oldtopk(count)statement belongs to the earlier defect record. - Commit count: 147 was a previous timeline snapshot; 144 is the direct count at current
developHEAD. - Blueprint completion: M1/M2/M3 “in progress” and M4/M5 “planned” remain handoff statuses; absent modules and pending gates control implementation truth.
- Database cleanliness: the handoff's zero-FK-error and full-parity acceptance target is not current evidence; the direct audit still finds 2,946,336 violations and unresolved archive deltas.
- Historical architecture sanity versus Stage 3/4 failure: the August 6-7 ten-run suite found no broad encoding, conversion, cache or collapse failure within its own cohort and therefore supported a data- or regime-limited interpretation. The later Stage 3/4 audit found a loss-scale and validation contract defect. These are time-scoped observations, not competing claims about one undifferentiated current run.
- Prospective sidecar status: the sidecar and lateral planner are historical
artifacts removed at
a942373; current auxiliary labels and the would-KO contract survive in the primary pipeline. Pages that call the sidecar “current” are stale and must defer to pokemon_tcg_prospective_v2 and pokemon_tcg_research_backlog. - Entity normalization status: the L1/L2 ghost-team work is an analysis
backed by Antigravity scratch probes, not a tracked production resolver or a
persisted identity-mapping table. The prose
<2,000episode heuristic and the saved+60,000probe window are distinct artifacts; see pokemon_tcg_entity_normalization.
Legitimate differences remain separated: FP32 Stage 1–4 tournament cohorts, deck-conditioned results, teacher sweeps, database resets and historical design snapshots are not collapsed into one ranking.
Cross-references
- pokemon_tcg_ai_battle — project hub and reading map.
- pokemon_tcg_repository_timeline — full semantic Git history.
- pokemon_tcg_docs_corpus_provenance — documentation-file provenance.
- pokemon_tcg_stage3_training_failure_postmortem — loss and validation incident.
- pokemon_tcg_empirical_ablation_monograph — source-bound quantitative monograph and cohort separation.
- pokemon_tcg_data_augmentation_vehicle_draft — future synthetic-data boundary and provenance contract.
- pokemon_tcg_aug14_data_etl_database_audit — raw ZIP, Parquet and FK audit.
- pokemon_tcg_aug14_architecture_and_handoff_audit — blueprint and Antigravity boundary.
- pokemon_tcg_orchestration_artifact_ledger —
.agentsorchestration topology and pending gate evidence. - pokemon_tcg_curriculum_stage_lineage — stage labels, commit lineage and separated tournament cohorts.
- elo-as-experimental-protocol — rating semantics across related research projects.
- pokemon_tcg_training_pipeline — reconciled training contract.
- pokemon_tcg_torch_inference — historical FP16 delivery and current FP32 correction.
- pokemon_tcg_implementation_specification — target services and acceptance gates.
- pokemon_tcg_schema_evolution — target relational v2 versus current physical schema.
- pokemon_tcg_arena_future_architecture — future arena boundary.
- pokemon_tcg_manuscript_chapter_lineage — narrative provenance and claim status.