GadgetGauge
Back to Methods

Rule-set changelog

A reviewed history of changes that can affect GadgetGauge's decision guidance. This is a record of behavior, not release marketing.

Current rule set 1.3.0

Only changes supported by reviewed repository evidence appear here. Earlier history is not reconstructed from memory, and an omitted history item should be read as unknown rather than unchanged.

Rule set 1.3.0

val-storage-1024 · version 1.0.0

Added
Previous behavior
No storage valuation rule existed for 1TB (1024GB) configurations. A device configuration referencing 1024GB storage would fail comparison with a DataIntegrityError ("No storage valuation rule for 1024GB").
New behavior
1024GB (1TB) configurations now receive a +30% storage adjustment over the 128GB reference baseline, matching the existing 64/256/512GB tier pattern. This enables the newly added iPhone 16 Pro and iPhone 16 Pro Max 1TB device configurations to price correctly.
Why it changed
Solo todo #32: the device catalog stopped at iPhone 15, a first-contact credibility failure in the Aug-2026 used market (docs/audits/2026-08-06-full-product-audit.md problem #2). Adding the iPhone 16 family (16, 16 Plus, 16 Pro, 16 Pro Max, 16e) required a 1TB storage tier for 16 Pro and 16 Pro Max, which had no prior storage rule to price against. No shipped device configuration used 1024GB storage before this change, so this addition changes results for zero existing requests — it only makes a previously-impossible configuration price correctly.

Affected outputs

  • offer
  • priceContext

Reviewed source: DECISIONS.md#iphone-16-family-catalog-expansion-2026-08-06-rulesetversion-130, docs/audits/2026-08-06-full-product-audit.md problem #2

Rule set 1.2.0

val-reserve-cosmetic-unknown · version 1.0.0

Added
Previous behavior
Cosmetic condition had no "unknown" answer available to buyers. The intake told users unknown was a first-class answer and offered a bulk "mark unanswered as Unknown" shortcut, then handed a validation error to anyone who used it on this question because the enum had no unknown member.
New behavior
Buyers can now answer "unknown" for cosmetic condition. An unknown grade is never deducted like confirmed damage — no cosmetic ledger line is emitted — and instead holds back a 5% reserve until the buyer sees the phone. It also blocks high confidence the same way a confirmed "poor" grade does, and adds daylight, screen-on and screen-off photos to the missing-evidence list.
Why it changed
"I couldn't tell" is not evidence of wear and must never be priced as either the best or the worst case. Treating an unknown cosmetic grade like every other unknown (hold money back, lower confidence) closes a contradiction where the app claimed unknown was always a safe answer but silently rejected it for one question.

Affected outputs

  • confidence
  • offer
  • priceContext
  • missingEvidence

Reviewed source: DECISIONS.md#iteration-7-2026-07-28-unknown-as-a-cosmetic-answer-ruleset-version-120, commit 528d8b5

Rule set 1.1.0

risk-gather-payment-unknown · version 1.0.0

Added
Previous behavior
An unknown payment-before-inspection answer did not independently gate the decision.
New behavior
An unknown payment-before-inspection answer now produces Gather evidence before price-forward guidance.
Why it changed
Whether payment is required before inspection is critical transaction-safety evidence because a confirmed prepayment demand is a hard walk-away condition.

Affected outputs

  • verdict
  • reasons
  • missingEvidence
  • sellerMessage

Reviewed source: DECISIONS.md#engine-gap-closed-paymentbeforeinspection-unknown-now-gates-like-the-other-deal-breaker-unknowns-2026-07-28