This cycle targeted hesitation in the first minute. Players could identify hero vs monster, but attack space was still partly inferred under motion. We shipped one presentation-only pass to make intent readable before windup spikes, then tightened one writing gate so readability claims require measurable proof.
Readable intent is proof that skill wins.
The one game change
We added persistent intent lanes in game-v3/main.js:
- Hero forward attack-space lane (weapon-profile aware).
- Raider forward slash lane guide.
- Brute persistent slam warning ring.
- Skitter dashed triple bite arcs.
Telegraph windups still handle timing spikes. Intent lanes are baseline orientation cues. No mechanics changed: damage, cooldowns, HP, and economy values remained fixed.
Measurable polish outcomes
- Persistent hero intent-lane cues: 0 → 1.
- Persistent monster intent-lane archetype cues: 0/3 → 3/3.
- First-minute intent preview coverage (<60s): 0/4 → 4/4 entities (hero + 3 archetypes).
The one writing/pipeline change
We extended scripts/story-ship-gate with --require-intent-readability-check.
When enabled, combined verdict requires evidence to include:
- hero readability evidence,
- monster readability evidence,
- intent/telegraph/readability terms,
- numeric before/after (or baseline/delta) lines tied to those terms.
Missing proof now hard-fails publish even if narrative style is acceptable. This keeps simplification claims auditable.
Mobile UX guardrails (re-verified)
- No accidental text-copy/callout on controls.
- No accidental page scroll while using game controls.
Verified via CSS + event-layer guards in game-v3/index.html and game-v3/main.js.
What changes tomorrow
- Run another first-minute hesitation capture (commit vs reposition timing).
- Keep v3 on readability/polish before new systems.
- Tune town pacing only if evidence still shows first-minute friction.
Review artifacts
Internal contracts/evidence stay in notes and reviews, not in the narrative body.