Vault

The comparison studies — paired deep-reads

The four titles the series started as, organized in paired comparisons. The pair structure forces the comparison rather than letting each game stand alone — which is why these are indexed by arc, not by console. All four live in a single monorepo.

Links: series hub → Game Annotation · project SDK → projects/CLAUDE.md · hardware → TIA (2600) · PPU (NES)

Identity — one repo, four titles

Field	Value
Logical name	`game-annotation` (private) · `github.com/chrisaacson69/game-annotation`
Sibling path	`../game-annotation` (resolved via `.claude/local-paths.md`)
Layout	`adventure/` · `mappy/` · `mule/` · `utopia/` · `tools/`
Tools	`asm-relabel.py`, `mesen-labels.py`, `mesen-workspace.py`, `vm-disasm.py`

⚠️ Not cloned on this machine (2026-07-21).

⚠️ Its vm-disasm.py predates the NA1 extraction — verify against na1-decompiler before reusing it. The Mesen tooling here is one of three copies; see nes/tools.

Arc 1 — Hardware capability (6502 family)

How does the platform’s hardware constraint shape the design? Same CPU family, different display/memory architecture.

Game	Platform	ROM	Status
Adventure	Atari 2600	4 KB	complete (8 chapters)
Mappy	Famicom/NES	16 KB + 8 KB CHR	complete (8 chapters)

What the arc reveals: 2600 (no frame buffer, beam racing, 128 bytes RAM) → NES (PPU does rendering, 2 KB RAM, larger code budget) is roughly a 4–8× capability jump that enabled a different kind of game, not just a bigger one. Mappy also carries the NES design vocabulary inherited from Pac-Man.

Arc 2 — Mechanical complexity (early strategy engines)

How does the strategic decision engine evolve once the platform stops being the limiting factor? Different ISAs on purpose — mechanical depth is the comparison axis, not the chip.

Game	Platform	Status
Utopia	Intellivision (CP1610)	complete — 4,075-line source; one of the earliest console turn-based strategy games (1981)
M.U.L.E.	Atari 800 (6502)	complete — 49,981-line 6502 disassembly + Kroah’s algorithm doc (ch 7.5 reserved for Chris’s strategic-frontier analysis)

The arc question: what does player-vs-player auctioning add to a Utopia-style solo economic engine? Utopia (1981) already runs closed-loop adaptive AI via score-driven rebel spawning; M.U.L.E. (1983) keeps the economic spine and adds an auction layer.

Key M.U.L.E. finding — the mechanism-vs-equilibrium gap: the auction enables specialization, but spoilage + zero-sum incentives may pull the equilibrium back toward self-sufficiency + a Crystite lottery. A mechanism existing is not the same as a mechanism mattering — the seed of The Dead-Verb Test.

Nobunaga’s Ambition opened as the third entry in this arc, then broke out of the comparison format entirely and became a full decompiler → nes/.

Why these aren’t filed under a console

Both arcs are cross-system by design. Arc 1 pairs Atari 2600 against NES precisely to isolate the display-architecture jump; Arc 2 spans Intellivision (CP1610 — not even 6502) and Atari 800. Filing them by console would sever the comparison, which is the deliverable. See Author Web, Derive Hierarchy.