How we score products

A score belongs to a matchup

Each available factor contributes its assigned weight. Headphones award that weight to the better specification, splitting it for ties or differences below the model’s meaningful threshold. The other categories split numeric-factor weight proportionally between the values, favoring higher or lower values as appropriate. Boolean features split equally when equal and favor the product with the feature otherwise.

Points are divided by the available scoring weight and scaled to 100. The two scores add to 100, subject to rounding. A score of 60 is a relative share of the matchup, not “60% product quality.” Scores from different categories should not be compared.

Missing evidence is not a loss

A factor needs verified, applicable values for both products before it contributes. Unknown facts are excluded rather than assumed to be “no” or zero. Publication also requires the category’s evidence-coverage checks and an approved verdict tied to the current model and product data. A comparison can become ineligible when that evidence changes.

How rankings are calculated

The Rankings page averages each product’s Balanced scores against its eligible published opponents in the same category. With a complete ten-product catalog, that means nine opponents. Results are rounded to one decimal place and sorted highest first. Equal averages are ordered by number of eligible opponents, then product name. Products without an eligible comparison are not ranked.

Rankings can change when the catalog, evidence, or model changes. Incomplete comparison coverage can give products different sets of opponents. Other priorities are available on comparison pages and in Find my match; the Rankings page always uses Balanced.

Votes and money stay separate

Community votes express preferences and do not rewrite specifications or the scoring weights. Affiliate commission and sponsorship do not enter the score or ranking.