How We Score Gear

Last updated July 18, 2026. Every product we recommend earns a single 0 to 100 score, and that number is not a gut feeling. It is the weighted sum of a few sub-scores, each one measuring something a rider actually cares about, like whether a helmet is certified for e-bike speeds or whether a lock can slow down an angle grinder. This page shows how those sub-scores are weighted for each gear category, how the 0 to 100 number becomes the star rating on a pick, and when a score changes. Want the bench and street methods behind the numbers? Read how we test.

How the 0 to 100 score works

We split each category into four or five sub-scores, each graded from 0 to 100 on its own. Then we multiply each sub-score by its published weight and add up the results. A helmet that aces protection but vents poorly does not get punished like a helmet that fails protection, because protection carries far more weight. The weights are the honest part. They say out loud what we think matters most for that kind of gear, so you can disagree with us and re-rank the picks yourself.

Take a real example. Say a commuter helmet scores 90 on protection, 70 on ventilation, 80 on weight, 85 on fit, and 75 on value. Using the helmet weights below, that lands at (90 x 0.40) + (70 x 0.15) + (80 x 0.15) + (85 x 0.15) + (75 x 0.15) = 84.0, which maps to 4.5 stars. Change one input and the total moves. That is the point.

From score to stars

The star rating is a plain-language read of the 0 to 100 score. Each half star covers a 10 point band, so a small move in the number does not flip the stars unless the product truly crosses a threshold. We do not publish anything that scores below 50. If it cannot clear that bar, it does not belong on a list of gear we would put on our own bikes.

Score band	Stars	What it means
90 to 100	5.0	Class leader, we would buy it ourselves
80 to 89	4.5	Excellent, a confident recommendation
70 to 79	4.0	Strong pick with minor tradeoffs
60 to 69	3.5	Solid for the right rider or budget
50 to 59	3.0	Has real caveats, read the fine print
Below 50	Not listed	Does not make our guides

Helmet rubric

A helmet has one job that dwarfs the rest, so protection and certification carry nearly half the score. E-bikes move faster than pedal bikes, and a dual CPSC and NTA 8776 rating like the one on the XNITO E-Bike Helmet gets tested against impacts a standard bike helmet never sees. After protection, we weigh what decides whether you actually wear it every ride: ventilation, weight, and fit. Value rounds it out, so a certified $30 pick like the Retrospec Dakota can still beat helmets that cost ten times more.

Sub-score	Weight	What we grade
Protection and certification	40	CPSC, NTA 8776, MIPS or equivalent; coverage and impact standard for the intended speed class
Ventilation	15	Vent count and airflow; how it feels at commuter speed and at a stop
Weight	15	Measured mass and how the load sits on the neck over a long ride
Fit and retention	15	Dial adjuster range, pad options, strap comfort, stability in a shake test
Value	15	Protection and features earned per dollar against the category
Total	100

Lock rubric

In 2026, the way expensive e-bikes get stolen is a cheap cordless angle grinder, so grinder resistance is the heaviest single input in the lock rubric. A lock like the Hiplok D1000 that stalls a cutting disc earns points a hardened U-lock cannot, no matter how good its steel is against hand tools. Independent certification comes next, because a Sold Secure tier is an attack test we did not run and cannot fake. Usability and weight fill out the score, since a lock you leave at home protects nothing.

Sub-score	Weight	What we grade
Grinder resistance	40	How long the lock stalls or resists a cordless angle grinder, from published tests and maker claims we can corroborate
Sold Secure tier	25	Independent rating: Diamond, Gold, Silver, or Bronze, plus maker security levels where they apply
Usability	20	Shackle clearance, keyway quality, mounting or carry options, how easily it locks the frame to a real rack
Weight	15	Measured mass against security delivered; the penalty for carrying it every ride
Total	100

Light rubric

A light is graded on whether it lets you see and be seen at the speed you actually ride. Brightness in lux and lumens leads, because e-bike speeds outrun a be-seen light fast. Beam quality matters almost as much, since raw lumens thrown in a bad pattern blind oncoming traffic without lighting the road. Runtime and mount security round it out, because a light that dies mid-commute or rattles off a pothole is no light at all. Radar tail lights like the Garmin Varia RTL515 get scored against other radar units, not a basic tail light.

Sub-score	Weight	What we grade
Brightness (lux and lumens)	35	Rated output checked against real-world usable light for the speed class, day and night modes
Beam quality	25	Beam shape, throw, spill, and whether it lights the road without dazzling oncoming traffic
Runtime	25	Battery life at the mode a commuter would actually use, plus charge time and port type
Mount	15	Mount security on rough pavement, tool-free swap, and weather sealing rating
Total	100

Scores update on re-test

A score is a snapshot, not a verdict for life. We re-score a product whenever a sub-score input changes: a new certification, a firmware or hardware revision, a fresh round of independent grinder testing, or a price move big enough to shift the value sub-score. When a score changes, the last updated stamp at the top of the guide changes with it, so that date tells you how current the number is. If a product gets worse, its score drops and it can fall off the list entirely. We would rather move a pick than defend a stale one.

One honesty note. Where a sub-score leans on lab-style measurements like grinder cut time or beam lux, we say so and cite the source, whether that is our own bench work or independent testers we trust. We never present a number we did not measure as if we ran the test ourselves. The full bench and street protocol lives on the how we test page.

READ HOW WE TEST

Scoring FAQ

Why does protection outweigh everything else on a helmet?

Because a helmet that vents beautifully and weighs nothing does you no good in the one moment it exists for. E-bikes ride faster than pedal bikes, so we weight certification for the right speed class, like CPSC plus NTA 8776, above comfort features. Comfort still counts, because a helmet you refuse to wear protects nobody, which is why fit, weight, and ventilation together carry as much as protection alone.

Do you run your own angle grinder tests on locks?

Where we have run our own bench work, we say so and show it. Where a grinder resistance sub-score leans on independent testing, we cite that source instead of claiming the measurement as ours. Either way the number is grounded in a real attack test, not a marketing claim, and the how we test page spells out the method.

How does a 0 to 100 score become a star rating?

Each half star covers a 10 point band. A score of 90 or above is 5.0 stars, 80 to 89 is 4.5, 70 to 79 is 4.0, and so on down to 50. Anything below 50 does not make our guides at all. The banding keeps stars stable, so a one or two point change on re-test will not flip the rating unless the product truly crosses a threshold.

Can a cheap product still score well?

Yes. Value is a weighted sub-score in every category, so a certified budget helmet or a strong commuter light earns credit for the protection and performance it delivers per dollar. Price never buys a place on our lists, and a low price never keeps a product off one if it clears the bar.

How often do scores change?

Whenever an input changes. A new certification, a hardware revision, fresh independent testing, or a real price swing can all move a score. When it moves, the last updated date on the guide moves with it, so that stamp is the fastest way to see how current a score is.