Table 2Saved FSM shootout results for model and teacher heads

Empirical result Supported by saved experimental runs, not by a proof.

1 Paper text

Method	Heads	Rollouts	Accuracy	Recall
Single head (model)	1	1.00	0.768	0.536
Multi-head + FSM (model)	4	2.87	1.000	1.000
Single head (teacher)	1	1.00	0.768	0.536
Multi-head + FSM (teacher)	4	2.87	1.000	1.000

Table 2: Single-frame

P/N

evaluation by FSM-controlled shootouts (

N=20

k=4

), reporting overall

P/N

accuracy and

N

-position recall.

Generated from paper_jmlr.tex, lines 1197–1212.

2 Explain this

The table reports, for single-head and multi-head (FSM) evaluation, with the learned model and with the exact teacher heads: the number of heads, the mean number of rollouts, overall $P/N$ accuracy and $N$ -position recall. Model and teacher rows coincide.

3 How this is formalised

How this is covered. Empirical (saved outputs), apart from the theoretical values discussed in Section 7.3, paragraph 3 (arithmetic).

Coverage record
Paper anchor	`tab:fsm-shootout` (manuscript label)
Displayed as	Table 2
Claim	Saved FSM shootout results for model and teacher heads
Coverage category	Empirical result
Statement source	paper_jmlr.tex lines 1197–1212; SHA-256 of the normalised block source `cca551d958baf864…`
Explanations	`explain/tab:fsm-shootout`, `formal/tab:fsm-shootout`
Build evidence	fresh isolated rebuild, audit report

4 Lean statement and source

No Lean statement corresponds to this claim. This is an empirical statement about saved experimental runs. Lean proofs do not certify training runs, curves or checkpoints.

Build evidence: fresh isolated rebuild of all 38 bundled modules with Lean (version 4.26.0, started 2026-10-10T15:21:56Z, finished 2026-10-10T15:27:00Z: 38/38 modules compiled, 1588 axiom reports, no errors. See Build for the commands and the independent audit of 2026-10-10 (dated evidence).