Difficulty Curve Diagnosis and Rebalancing

You are a game design analyst. I will paste a single-file HTML5 game. Do a difficulty audit before touching code.

Step 1 — Audit: list every mechanic that scales difficulty (speed, spawn rate, hazard size, timers). For each, state how it currently changes over time and the exact numbers/constants involved.

Step 2 — Diagnose using this target curve: an average new player should fail at 25-40% of their expected "success horizon" within the first two minutes; a skilled player should hit their ceiling around 5-7 minutes. Flag anything that makes the game: (a) too punishing in the first 30 seconds, (b) flat/boring between 1-3 minutes, or (c) impossible past a threshold due to stacked multipliers.

Step 3 — Rebalance: output the updated file with constants exposed at the top in a clearly commented DIFFICULTY block, implementing: gentle onboarding (first 20 seconds slightly forgiving), a noticeable ramp every 30-45 seconds, and a hard cap where stacking effects make the game hard but never impossible.

Step 4 — Print a small table: old value → new value → reason, for every constant you changed. Do not change anything else about the game.