[chase-manning]

Fix importance weight scaling to use geometric spacing

Problem

The quiz's per-question importance rating (1-5) was used as a raw linear multiplier on each trait's contribution to the match score. This created wildly inconsistent proportional impact across the scale:

Step	Ratio increase	% increase
1→2	2/1	+100%
2→3	3/2	+50%
3→4	4/3	+33%
4→5	5/4	+25%

Choosing between 1 and 2 doubled a trait's weight, while choosing between 4 and 5 only increased it by 25%. Users perceive each step as equally meaningful, but the math didn't reflect that.

Solution

Replaced the raw linear multiplier with a geometrically-spaced weight lookup table ([1.0, 1.5, 2.24, 3.34, 5.0]), where each step is a consistent ~50% proportional increase. This is derived from r^(importance - 1) with r^4 = 5 (preserving the existing 5:1 max/min dynamic range).

This approach is grounded in the Weber-Fechner law from psychophysics, which shows that humans perceive proportional changes — not absolute ones — as equal. It's also the recommended practice in multi-criteria decision analysis (MCDA) literature for transforming ordinal importance ratings into ratio-scale weights.

Changes

• Added IMPORTANCE_WEIGHTS lookup table and getImportanceWeight() helper to dog-rating.ts
• Applied the geometric weight in all 3 places importance is used: rating questions, string checkbox matches, and array checkbox matches
• No quiz cache version bump needed — importance values are still stored as raw 1-5 and transformed at scoring time

[github-actions]

Visit the preview URL for this PR (updated for commit 990cfd3):

https://dogmatch-33--pr1-algorithm-update-funb46yc.web.app

_{(expires Mon, 20 Apr 2026 10:10:19 GMT)}

_{🔥 via Firebase Hosting GitHub Action 🌎}

_{Sign: 3be598a5fc061b587f1c015bd81cd7e7b2c794a8}

[Copilot]

Pull request overview

Updates quiz scoring so per-question importance (1–5) is mapped to geometrically spaced weights, making each step produce a consistent proportional impact on match scoring. Also adjusts GitHub Actions workflow permissions for Firebase Hosting deploys.

Changes:

• Introduced a geometric IMPORTANCE_WEIGHTS table and helper to transform importance ratings into ratio-scale weights.
• Applied the transformed weight in rating, string checkbox, and array checkbox scoring paths.
• Added explicit permissions blocks to Firebase Hosting PR/merge workflows.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
src/app/dog-rating.ts	Replaces linear importance multipliers with geometric weights via a shared helper.
.github/workflows/firebase-hosting-pull-request.yml	Adds explicit workflow token permissions for PR previews.
.github/workflows/firebase-hosting-merge.yml	Adds explicit workflow token permissions for merges to main.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.