Twenty products in each of eight category methods.
Every ranked product has an inspectable evidence profile.
All 160 records expose mapped inputs, score sensitivity, source coverage and open gaps. Profiles make decisions auditable; they do not hide missing evidence or turn rankings into purchase advice.
9 records still need a separate official source.
Independent evidence is the largest source-coverage gap.
Each pool remains source-specific, not internet consensus.
34 also retain return terms; 11 retain manual-derived safety context.
The remaining profiles do not manufacture atomic claims.
Eight complete categories
160 profiles, with missing evidence left visible
Evidence checked 29 Jul 2026, 18:17 BST. These records describe the retained snapshot, not the page render time.
portable-ac-v1.0
Portable air conditioners
20 retained candidates across 2 brands. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead8/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
2 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:87a4a4170b5158239c8d83de357e011bb7c9ad4cae3b2a5a9a9a4a30e73b2065dehumidifier-v1.0
Dehumidifiers
20 retained candidates across 5 brands. All 20 have a readiness profile.
Weight stress testRank #1 remains a leader15/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
1 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:f6433e4485d5d78992c1bed12ab78fd12c79c0323385d743173927966632ccfcpower-bank-v1.0
Power banks
20 retained candidates across 1 brand. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead14/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
2 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:e8f2843a54e2cff5c1753e349a91749ba48a001ecd39e3b09e490846feec4b0cair-fryer-v1.0
Air fryers
20 retained candidates across 2 brands. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead14/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
3 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:71bacc2156d271afce2ac024ecb1e793da0218c7b230f46f0b033001ff59acf5cordless-vacuum-v1.1
Cordless vacuums
20 retained candidates across 2 brands. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead13/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
3 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:fbb796d23e6084371723e2e538d0005688b2b242d7cd73ce653fdd18a609d225robot-vacuum-v1.1
Robot vacuums
20 retained candidates across 2 brands. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead14/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
12 products lead at least one scenario. 2 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:146710998d71835e018f2d60090e3e7df7ac29a31f9b27b2dbb8dab1ff2c9969wireless-earbuds-v1.1
Wireless earbuds
20 retained candidates across 1 brand. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead13/15 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
3 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:c2d5d928b3225dd96907f395c74c2708319c18d8068d511446d501ab6c93c797air-purifier-v1.0
Air purifiers
20 retained candidates across 3 brands. All 20 have a readiness profile.
Weight stress testRank #1 loses the lead5/6 scenarios retain the published rank-one product
Candidate-set boundary
The candidate set is the published 20-product category, not a claim of complete UK market coverage.
All 20 candidates have normalized profiles. Ranks 6 to 20 remain outside the historic leading-five deep-research band, so profile coverage must not be mistaken for complete independent evidence.
2 products lead at least one scenario. 0 scenarios produce at least one leader outside the deep top-five cohort.
candidate-set sha256:1415ec83fa10aa7006f22d4b4993bb50933c1e1c1e737e2cf8be400e6b0a6add“Computed sensitivity makes a ranking easier to challenge. It does not turn published claims into measured performance.”
Controlled measurements and human second-person review remain absent. Second retailers, safety context and return terms are only partially covered. Those gaps prevent purchase advice regardless of the numeric score.