Software directs the model.
Predict, look ahead, check, play.
Explore the current agentAct Two / paused indefinitely / updated 13 November 2026
We trained a model to interact with a chess VM, measured where it works and where it fails, and outlined a possible reinforcement-learning continuation. That continuation has not been run.
Act Two is indefinitely paused. No GRPO training has been run and no reinforcement-learning improvement is claimed. The Act One CM-14 agent remains the live Lichess player.
In Act One, the harness asks the model for predictions and decides what to do next. In Act Two, the model emits moves and simple tool-call tokens. A virtual machine (VM) checks them and returns observations to the model. The board stays authoritative in the VM; the harness still manages the clock and the connection to the game.
Predict, look ahead, check, play.
Explore the current agentAsk for the board, propose a move, learn from feedback.
PVM-01 recovery training started from the original CM-14 weights, followed by an eight-hour supervised recovery run and an additional eight-hour continuation on the same data. The evaluated final checkpoint is step 471,623. These were supervised runs, not self-play training. The earlier random-initialization experiment is not the origin of these final results.
The VM returns observations and attributed rejections. There is no legality mask in the following two evaluations. After three rejected move proposals on a turn, the harness inserts a seeded random legal move into history. That safeguard prevents getting stuck; it is not a model-generated move.
84 wins / 16 draws / 0 losses, out of 100 games. All reached chess endings: 84 checkmates and 16 stalemates. No scored VM failures or incomplete games.
Random fallback: 512 / 3,818 PVM turns (13.4%). This measures the model-plus-fallback system, not fully autonomous legal proposals.
32 White wins / 36 Black wins / 32 draws, out of 100 games. Independent seeds and private VM histories on each side; no teacher and no learning during games.
All reached chess endings: 68 checkmates, 24 stalemates, four fivefold repetitions and four insufficient-material draws. Zero scored incompletes.
A near-even White/Black split is not evidence of improved strength or proof that chess is balanced. Both sides use the same checkpoint. The useful finding is that the system can produce complete games with varied outcomes, while fallback dependence remains substantial.
These are offline VM-harness results, not results for the full live CM-14 search harness. Weak-engine and teacher-intervention explorations did not establish a training improvement. Teacher-assisted victories would measure the assisted system, not learned behavior.
A turn, seen from both sides
A model proposes. The VM checks. Its response joins the context the model reads next.
Illustrative opening sequence, not a training result or a recorded game. The recovery path is shown deliberately; a legal first proposal skips it.
Queries and rejected proposals do not move a piece.
e2e4The VM supplies the snapshot. The model does not have to invent the board.
Three separate exchanges. Each VM response contains two tokens: the attempted move, then REJECT. Append that returned pair once; do not append the proposal separately. The board stays unchanged.
The VM lists legal options, not the best move. The model still chooses.
The accepted move appears once in history. The opponent’s real reply, for example e7e5, is appended next.
The proposal is a request, not a second history entry. The harness appends the VM's returned tokens once, then builds the next model input from that updated history.
H = [earlier events, BOARD_QUERY, snapshot]e2e5Send to the VM. Do not append separately.[e2e5, REJECT]No board change.H + [e2e5, REJECT]One attempted-move token, one rejection token.After all three rejected attempts: H + [e2e5, REJECT, e2e6, REJECT, e2e7, REJECT]. White is still to move in the starting position. Query responses and accepted moves extend the same history; only accepted moves advance the board.
Readable labels for the protocol, not a literal tokenizer dump. BOARD_QUERY stands for the board-query command; snapshots and legal lists are VM-supplied observations. “Move accepted” above explains the state change, not an extra ACCEPT token.
BOARD_QUERY[BOARD_QUERY, snapshot] appended once. Human-readable FEN equivalent:rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1[e2e5, REJECT] [e2e6, REJECT] [e2e7, REJECT]LEGAL_QUERY[LEGAL_QUERY, LIST_BEGIN, all 20 legal moves, LIST_END] appended once.a2a3 a2a4 b1a3 b1c3 b2b3 b2b4 c2c3 c2c4 d2d3 d2d4 e2e3 e2e4 f2f3 f2f4 g1f3 g1h3 g2g3 g2g4 h2h3 h2h4e2e4[e2e4] · legal, committed oncee7e5 · illustrative opponent reply, not generated by this modelWhat must be learned: asking for the board, choosing when to request legal options, and recovering from feedback. This illustration does not establish that PVM-01 already does these reliably.
REJECT leaves the board unchanged and the model retries.The earlier design used a 1,024-token context; the evaluated recovery and continuation checkpoint used 512 tokens with a full snapshot plus recent moves and feedback. The supervised data included refresh and partial-history examples; evaluation must measure actual autonomous refresh rather than assume it works. The model must learn to request refresh, rather than relying on an evaluator to force it. Snapshot tokens are VM observations, not text the model must invent.
Start/end markers delimit games. Move, query and outcome tokens share one next-token prediction interface. The outer harness can insert a win/draw/loss query in an isolated diagnostic context; the agent does not call it during normal play. There is no separate WDL head.
Human games and engine-analyzed positions teach next moves. Derived interactions teach board queries, partial-history play, refresh and recovery from illegal proposals. Outcome examples teach the harness-only diagnostic query.
Examples are prepared and checked offline, then randomly mixed with controlled proportions. The optimizer reads cached sequences; it does not call the VM during training. Related examples share family-level exposure limits and held-out boundaries so repetition is not mistaken for new data.
They provide analyzed positions and next-move supervision. Multiple alternatives are not automatically negative examples, and no search trace is trained in this round. Using those alternatives to learn exploration is a later experiment.
REJECT.Each measure needs attempt counts, baselines and a distinction between model behavior and forced safeguards.
The longer-term idea is to let the model generate short what-if lines, with legality checked on isolated VM boards, before committing a real move. A six-ply target is a starting point, not an achieved capability. Board queries and diagnostic WDL calls would stay outside hypothetical continuations.
What-if traces remain in context in the proposed design to preserve useful history and compatible cache prefixes. A host-injected DEADLINE would request a final move, with a separate hard clock limit enforced by the harness. These are unrun design questions, not demonstrated capabilities. The evaluated recovery recipe excludes what-if search and resignation.
Proposed only / not run
The idea was to continue from the supervised PVM checkpoint using Group Relative Policy Optimization (GRPO), without adding an imitation-learning loss. The environment supplies rule-based outcomes; no learned reward judge is required. This is a roadmap for a future experiment, not a scheduled training run.
Engine losses may expose weaknesses that self-play misses and reduce the risk of co-adaptation, but an all-loss group can still have zero outcome variance. Opponent diversity is a hypothesis to test, not a guarantee against training collapse. Requested engine nodes can overshoot, so actual work must be recorded.
Teacher intervention and imitation were explored as alternatives, but the preferred next experiment was pure GRPO. No teacher moves should be silently relabeled as samples from PVM. No GRPO gains, Elo improvement or successful training curriculum have been demonstrated here.