Add game-type format profiles (final roadmap item)

[testtest126]

Summary

• GameFormat (6 cases: .mttRegular, .mttTurbo, .mttHyper, .pko, .satellite, .cash) + GameFormatProfile — a pure defaults/config layer carrying per-format starting-point values: default effective stack, whether antes are the sensible assumption, whether the PKO bounty overlay defaults on (and its seed size), whether/how much the ICM-aware UI should lean in, and blind-level speed.
• This is design judgment, not ground-truth math or a transcribed source — unlike ICM (validated against a published worked example) or the range models (transcribed from cited charts), there's no external "correct answer" for what a hyper-turbo's typical starting stack should default to. Every value is disclosed in the doc comments and ai-docs/FORMATS.md as this project's own chosen default, with a one-line rationale — not defended as more "correct" than a reasonable alternative choice.
• Defaults only, never a mutation: no model this feeds (PushFoldRange/OpeningRange/CallingRange/ThreeBetRange/FourBetRange/BountyEquity/ICM/ICMRiskPremium) reads a GameFormat at all — a caller reads a profile once, at format-selection time, to pre-fill an existing parameter. Nothing clamps or overrides a value the user has already set.
• Judgment calls worth flagging explicitly:
  • Turbo/Hyper share Regular MTT's ICM weight (0.5) — blind speed changes stack depth, not payout shape, so this project chose not to invent an unjustified distinction there.
  • PKO's ICM weight (0.4) is set below Regular MTT's — reflecting the commonly-discussed PKO tension that a collectible bounty incentivizes looser play, partially offsetting ICM's tightening pressure (already described in BOUNTY.md's "two competing forces" section).
  • PKO's default bounty size (33% of starting stack) reuses the exact figure already used as a worked example in BountyEquity's own doc comment, rather than inventing a second independent guess.
  • Satellite's ICM weight (0.9) is the highest of any format — every seat pays the same, so min-cash ≈ the entire goal.
• New Game Format picker view: selecting a format resets seeded fields (stack/bounty/ICM) to that format's defaults, freely editable after. Deliberately self-contained rather than retrofitting every existing screen's own independent state — scoping rationale is in the view's doc comment and FORMATS.md's Consumers section.

Test plan

• swift test (PokerKit): 235/235 green (220 baseline + 15 new GameFormat tests — internal-consistency checks: exhaustive/stable case list, Codable/rawValue round-trip, pko-is-only-bounty-format, cash-has-neither-icm-nor-bounty, satellite-leans-hardest-on-icm, pko-leans-less-than-regular, stack-depth-shrinks-with-speed, and structural invariants like bounty-fraction-set-iff-bounty-enabled).
• App build succeeds on simulator (iPhone 17, iOS 27.0, Xcode 27 beta).
• App UI test suite: 21/21 green (excludes BankrollTrackerUITests.testLogSessionUpdatesListAndSummary, previously confirmed pre-existing/flaky on a clean main checkout across the last two PRs — unrelated to this one).
• New GameFormatUITests covering the default (Regular MTT) seed, PKO seeding bounty-on at 33%, and Cash seeding ICM/bounty off with the speed row hidden entirely.

Roadmap status

This is the last item on the original roadmap (push/fold → RFI → calling → bounty → equity → 3-bet/4-bet → ICM → formats). Per standing instruction, stopping here for direction before starting the variants/PLO track.

🤖 Generated with Claude Code