Balance: an optimal-play simulation showed the game finishing in
under 3 minutes, mostly because torso_beacon + grand_ritual_prep
alone pushed ritualRate to 2.5 — enough to clear the last 50 ritual
points in 20 seconds. Softened those boosts (0.5→0.15, 2→0.35),
raised deepen_ritual's cost curve so it's a real endgame sink instead
of a one-off top-up, and added two new late-game upgrades (sunken
choir, nine filled chairs) between grand_ritual_prep and the win.
Re-simulated: a greedy optimal-play bot now takes ~5 minutes instead
of ~3 and surfaces ~40 distinct events instead of ~30 out of a much
larger pool — real (non-bot) play should run considerably longer.
Content: added 8 new high-dread late-game standard events (gated up
to minDread 400, since nothing previously escalated past 75) and 7
new kaos events, plus one more choice event.
Kaos events no longer get marked as permanently used — they're
chaos-relief flavor, not narrative progression, and a ~24-entry pool
(now 30) was depleting for good over a long game. Standard/chain
events remain single-use.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- Adds a full-screen intro/title screen (new 'intro' game phase) with
a moody Malmö/Turning Torso background and a "Börja ritualen"
button; the game no longer starts ticking until the player begins.
- Win and loss overlays now show matching generated artwork behind
the existing text, instead of a flat scrim.
- New "Avbryt" header button lets the player abandon a run mid-game
(with a confirm prompt) via a new 'abandoned' phase, reusing the
existing overlay/restart flow.
- Source images converted from PNG (~2.1MB each) to WebP (~120KB
each) and placed under public/img/ so Vite serves them as-is.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds an influence currency that scales with follower count and a new
"exert influence" action to trade it for secrecy, uncaps recruiting
with scaling cost, adds a late-game Andakt sink, and makes hemlighet
drain scale with dread/ritual progress so the endgame carries real
risk. Includes a Vitest suite that simulates play to check balance,
which caught the influence-accrual rate being too slow to be useful
early game.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
