feat: complete the beginner quantum learning journey

[GwriPennar]

Beginner comprehension problem

The Guided Experiment already established exact truth before showing quantum evidence, but it did not fully connect binary modelling, repeated measurement, distributions, and hardware noise for a first-time learner.

What changed

• reorders the post-Reveal journey to exact answer → model translation → quantum measurement → registered comparison;
• defines why eight binary choices produce 256 groupings, what energy means, and why lower is better;
• decodes 00001111 and its complement-equivalent 11110000;
• explains shots and why repeated quantum measurements form a distribution rather than one authoritative answer;
• visualises the ten most frequent outcomes from the committed 4,096-shot registered simulation using stable, non-mutating sorting;
• labels both exact-optimum states directly in the chart and keeps exact enumeration authoritative;
• defines hardware noise before the first recorded IBM evidence;
• adds concise glossary entries for energy, noise, superposition, Spearman rho, PUB, hardware backend, and Hadamard.

Evidence and safety

• registered scientific evidence and numerical values are unchanged;
• nothing under experiments/** changed;
• no governed Build Week model or loader changed;
• IBM hardware was not contacted and no workload was submitted;
• no credential, network access, Qiskit installation, or OpenAI key is required for the core journey;
• GPT-5.6 remains a bounded explanation layer after validated evidence;
• no quantum-advantage, speedup, scalability, superiority, or musical-truth claim was added.

Validation

• focused Foundations and Guided Experiment tests: passed (18 tests);
• full non-quantum suite: passed;
• Ruff format check: passed (152 files);
• Ruff lint: passed;
• Build Week release verification: passed;
• public-safety scan: passed;
• diff check and frozen evidence/dependency scope check: passed;
• desktop visual review: passed;
• 700px responsive review: passed with no horizontal overflow;
• pre-Reveal result gating: preserved;
• clean GitHub full mypy and CI are expected on this draft PR; the available local Qiskit environment lacks the development stubs needed for a valid full mypy run.

Visual review

The revised chart was reviewed at desktop and 700px widths. Full bitstrings remain readable, exact optima are labelled textually inside the bars, registered counts are visible, and simulation/hardware/evidence boundaries remain distinct. Temporary review screenshots were not committed.

Reviewer walkthrough

1. Open Experiments → Start here · Guided experiment.
2. Confirm no result appears before Reveal all 256 answers.
3. Reveal and follow: all 256 scores → exact answer → bitstring decoding → model translation.
4. Read the shot/distribution bridge and inspect the registered top-ten outcome chart.
5. Confirm the registered comparison, optional local Qiskit path, GPT-5.6 boundary, downloads, and limitations retain their existing authority and gating.
6. Open Real folk data & IBM results and confirm noise is defined before the first IBM hardware evidence.
7. Search the Glossary for energy, noise, rho, and PUB.

Deferred

Tier 3 Foundations “Go deeper” disclosure bodies were intentionally deferred. They require typed lesson-model changes and were not necessary to complete the bounded beginner journey.

Foundations follow-up

• replaces the awkward “Do not tell beginners…” admonition with learner-facing language about amplitudes, probability, phase, and measurement;
• renders the repository’s restricted flowchart LR / flowchart TB lesson diagrams as visible, dependency-free concept flows;
• escapes every diagram label and fails closed for unsupported syntax;
• retains Mermaid source in a collapsed View diagram source expander;
• verified on the Bits & Qubits page at desktop and narrow widths with no horizontal overflow.