Created GAME_STANDARD.md — a universal 4-phase architecture for every assessment test in Baseline (Welcome → Practice → Test → Results). Established per-hand scoring over total scoring as the primary metric for the Bilateral Coordination Test. Learned the hard way that version control is non-negotiable after nearly losing a session of work.
GAME_STANDARD creation: Worked with Claude to articulate what all tests had in common — onboarding, practice, measurement, results. The standard emerged from that conversation, but the key decisions (4 phases, ready screen as UI overlay, single phase variable) were mine.
AI suggested a single total score — simpler to communicate, easier for users to understand immediately. I kept per-hand scoring as the primary metric: Left Hand: 24, Right Hand: 23, Symmetry: 96%. The Bilateral Coordination Test exists to measure bilateral coordination — which requires knowing how each hand performed independently. A total score obscures the very thing the test measures. This is a measurement validity issue, not a UX preference.
I had been treating git commits as optional — pushing work at the end of sessions if I remembered. This session I forgot to commit and nearly lost work when the browser tab crashed. Hard lesson: version control is non-negotiable. Not a tool I'll use when convenient; a minimum condition for every session. This shouldn't have been a lesson I learned through failure, but it was.
The most important decision in the standard: the ready screen and countdown are UI overlays, not phase changes. Phase stays at 'practice' through the countdown. The phase only changes to 'test' when data recording actually starts. This prevents premature data capture and keeps the state model clean.
// Single phase variable — four valid states
let phase = 'welcome'; // welcome | practice | test | results
// Ready screen = overlay, NOT a phase change
// Countdown = overlay, NOT a phase change
// Data recording starts ONLY when phase === 'test'The standard freed up creative energy for the mechanics of each test — instead of redesigning the onboarding flow every time, that structure was already solved.
Standardization enables scale. Creating GAME_STANDARD.md in Session 8 meant Session 12's Rhythm Sync and Session 17's Fine Motor Precision were architecturally coherent from day one. The investment in a standard paid off in every subsequent test.
Also learned: accuracy over convenience when measurement validity is at stake. Per-hand scoring requires more UI real estate and more explanation. That cost is worth paying when the test is specifically about the relationship between two hands.
The symmetry metric — min(L,R) / max(L,R) * 100% — is a direct answer to the quarter question. It measures something real about bilateral coordination that a total score doesn't capture. A person at 40 dots, 20+20 has different coordination than a person at 40 dots, 36+4. The metric makes that difference visible.
Bug fixes on the Bilateral Coordination Test. Screen split (left-side dots stay left, right-side dots stay right) to remove the cross-midline confound. Build a comprehensive results screen with timeline visualization.