100 points. Nothing hidden.
Every e-bike we rate is measured against the same published specifications, grouped into six categories that add up to 100. The score is a sum, not a formula. You can check our arithmetic on any bike, and we would like you to.
16 + 21 + 14 + 32 + 8 + 9 = 100
On this page
Five rulesBefore we score: eligibilityWhy these six categoriesEvery metric, every rungWhat we deliberately do not scorePerformance vs valueHow categories workHow stable is this?Rule changelogQuestions we getFive rules we hold ourselves to
1. Half points, and no finer
Every metric is worth a whole or half point, and every score is a multiple of 0.5 out of 100. Half points exist because without them two rungs on the smaller ladders collapse into the same number and stop distinguishing anything. Nothing needs a second decimal place, and we do not publish one.
2. The parts add up to the whole
6 category subtotals, printed on every product page, that sum exactly to the score. No weighting step after the fact, no normalisation, no rounding you cannot see.
3. Adding a bike never moves your bike
Scores are absolute, not a curve. A bike that scored 62 last year scores 62 today unless its specification changed or we corrected an error — and we publish every correction.
4. Safety is a gate, not a bonus
No bike earns points for being certified or for having brakes that work. Those are the price of appearing here at all. One stated exception: a motor cut-off at the brake lever is scored rather than gated, because no US law, no CPSC rule and not UL 2849 requires one, and at least one manufacturer removes it deliberately. We would rather score that difference than delist a bike over a rule we cannot source.
5. We score what we can check
Every scored metric comes from a published manufacturer specification. Where a manufacturer does not publish one, the bike scores zero for that metric and we say so on the page. We do not estimate.
And one thing we do not do
No metric anywhere in this system takes the brand name as an input. You can read the whole rulebook below and check.
Before anything is scored: 6 eligibility gates
These are pass or fail and worth zero points. A bike that fails any one of them is not ranked, at any price.
Why these 6 categories, and these weights
The weights are a judgement, and we would rather argue for them in public than bury them. Here is the case for each.
| Category | Points | Why it is worth that |
|---|---|---|
| Power & Drive | 16 | How the bike feels under you and whether it gets up the hill. Peak power output alone is 9.5 points, because it is the specification a rider notices within thirty seconds of the first ride. Cadence-only sensing now scores zero: it is the cheapest possible way to detect that you are pedalling, and it feels like it. |
| Battery & Range | 21 | Range anxiety is the most common reason people give for not buying an e-bike, and charge time is the most common complaint after buying one. Usable capacity is the single largest metric in the whole system at 12 points, and it counts both packs on a dual-battery bike. |
| Safety & Braking | 14 | A 70 lb vehicle at 28 mph. Mechanical disc brakes now score zero — they are the bare minimum and nearly every bike has them — while a four-piston hydraulic calliper earns the full 6.5, because it is more than twice the stopping power and more than twice the cost. |
| Ride Quality & Components | 32 | The largest category by some distance, because it decides whether the bike is still good in year three. A rigid steel fork costs about $6 to make; a name-brand suspension fork is $80 or more, and the spread now reflects that. Forks, drivetrains and tyres are what wear out. |
| Utility & Practicality | 8 | Whether the bike does a job on the day it arrives, without a second order. Racks, fenders and a stand that holds a loaded bike upright. No rule changed here — its share fell only because the scale grew around it. |
| Display & Connectivity | 9 | Real, and worth showing — a colour TFT is a genuinely better ride than a monochrome LCD, and it costs more. But a screen is still not a bicycle, so it stays below every category that determines how the bike rides. |
Every metric, and every rung on it
This is the entire rulebook — 27 metrics. Each metric is a ladder: a bike sits on exactly one rung, and that rung’s points are the points it earns. Nothing else happens.
Power & Drive
16 of 100 pointsSystem voltage x controller amperage - the one number that decides acceleration and hill climbing. Scored once, not three times.
A torque sensor measures how hard you pedal and matches it. Every bike has at least a cadence sensor, so cadence alone earns nothing.
Lets you pull away from a stop without pedalling - useful with a load, on a hill, in traffic, or with joint pain.
Class 3 bikes (28 mph) keep pace with city traffic. Class 1 and 2 stop assisting at 20 mph.
Battery & Range
21 of 100 pointsTotal watt-hours you can carry, counting a second battery if one is included. Capacity is scored once whether it arrives in one pack or two.
The charger's own output in watts. Scored on the charger alone so that a large battery is not penalised, and so a 36 V high-amp charger cannot pass for a 48 V one.
Cells from LG, Samsung or Panasonic publish cycle-life and thermal data. Anything else earns nothing.
The published IP rating of the battery pack - not the bike. IPX5 is a water-jet test only; IP65 adds a dust-tight claim; IPX7 adds immersion.
Safety & Braking
14 of 100 pointsMechanical disc is the bare minimum and earns nothing. Hydraulic discs stop harder with less hand effort; a four-piston calliper is more than twice the braking power of a two-piston and over double the price.
Whether pads, rotors and callipers can be bought and serviced anywhere. Tektro and Star Union are both serviceable; Magura is better.
On a throttle bike, squeezing the brake should cut motor power. It is a backstop for a stuck throttle or a controller fault. A bike with no throttle cannot have that failure, so it scores full marks here.
A rear light that brightens under braking is the cheapest way to be seen slowing down.
Signal without taking a hand off the bars.
Ride Quality & Components
32 of 100 pointsA rigid steel fork costs about $6 to make; a name-brand suspension fork is $80 or more. The spread reflects that.
Gear range, shifting quality and maintenance burden on one ladder. An internally-geared hub with a belt is several levels above a single-speed chain and costs roughly eighty times as much.
The best available proxy for grip, rolling resistance and puncture resistance. Deliberately kept small - a name-brand tyre is only about $5 more.
Weld finish is a visible proxy for frame manufacturing control.
A thru-axle holds the wheel more securely than a quick-release; a threaded thru-axle more securely still.
An adjustable stem lets one bike fit several riders. A quick-release adjustable stem costs roughly triple a tooled one.
The part most owners replace first. Quick-release aluminium pedals fold the bike into a smaller space and do not snap.
Lock-on grips do not twist loose or slide off in the wet.
Utility & Practicality
8 of 100 pointsWhether the bike carries anything out of the box, and whether accessories clip to a published standard instead of needing bungee cords.
Full-coverage fenders are what make an e-bike usable in the rain. Composite fenders do not rattle, rust or cut.
An adjustable or double-leg stand holds a loaded bike upright.
Display & Connectivity
9 of 100 pointsHow much you can read at a glance, in sunlight, without stopping. A touchscreen TFT is $40-50 of cost over a colour LCD.
Whether you can change how the bike delivers power, or only how much.
Useful for firmware updates and diagnostics. Nice to have, but most riders do not use it.
What we deliberately do not score
A rating system is defined as much by what it refuses to count. These were all scored under our previous model and are not scored now.
| Was scored as | Old points | Why it no longer affects the score |
|---|---|---|
| Motor brand | 2.1 | Most bikes read simply “OEM”. A supplier label with almost no variance and no verified link to output. Shown as a specification. |
| UL certification | 2.1 | Every bike scored the full 2.1. A metric with zero variance cannot rank anything — and no bike should earn points for meeting a legal minimum. Now Gates 1 and 2, worth nothing and required of everyone. |
| Motor cut-off switch | 1.3 | Kept, rebuilt and moved into Safety & Braking at 2 points. It was briefly an eligibility gate, but no US law, no CPSC rule and not UL 2849 requires one, and at least one manufacturer removes it deliberately and says so — so it is scored, not gated. |
| Quick-release pedals | 1.3 | A yes/no convenience, not a measurement. Now a feature badge. |
| Plastic-free packaging | 1.3 | Recorded for a handful of brands and blank for everyone else. It rewarded research coverage rather than the product, and it does nothing for the rider. Removed outright. |
| Tyre sealant | 1.3 | A genuine benefit, but recorded for only three brands. Until every bike is audited for it, scoring it would reward our own research coverage rather than the product. Shown as a badge. |
| Second battery (as its own metric) | 5.6 | Merged into total capacity. A 5.6-point binary carried 15% of the whole model’s raw-points variance, and it paid twice for watt-hours already counted. |
| Computed motor wattage (as its own metric) | 4.3 | Verified to be exactly volts × amps in every row with data. Merged into one power metric. |
Performance and value are two different questions
The Performance Score is absolute
It describes the bike, and only the bike. A $3,999 bike and a $999 bike are measured on the same 27 ladders, and price is not one of them. This is the number we rank by.
The Value Index is a ratio, not a rating
Points per $1,000, printed to one decimal, never blended into the score. We publish it on its own board so that “what is the best bike” and “what is the most bike for my money” stay separate questions with separate answers.
How categories work — one score, one lens
A cargo bike and a folding bike are not the same product, but nine different scoring systems would be nine different things to trust. So we run one universal score and one secondary lens.
Performance Score — universal
The same 27 metrics and the same weights for all 87 bikes. This is the headline number and the thing we rank by, everywhere on the site.
Category Fit — the same points, re-read
The identical 6 subtotals, re-weighted with a fixed profile for that use case: cargo bikes weight utility double, off-road bikes weight ride quality and power, moped-style bikes discount racks and fenders. The profiles are published on each category page and they never change from bike to bike.
Category Fit is a badge, not a ranking. A bike cannot beat another bike on the leaderboard because it fits its own category well.
How much do the weights actually matter?
We tested this before publishing the current rulebook (September 2026). If moving a weight by a fifth reorders the board, the board was measuring our opinion, not the bikes.
±20% on every weight, 4,000 times
Rank correlation with the published order averaged 0.9968 and never fell below 0.987. On average, 8.5 of the top 10 stayed in the top 10, and the worst draw kept 7.
Shock one category at a time
Move any single category weight by a fifth, in either direction, and the correlation stays at or above 0.996. The most any one bike moves is seven places.
Remove a whole category
Delete any one of the six entirely and the order still correlates above 0.92 — lowest for Ride Quality & Components, which is the largest category and should be the most load-bearing. No single group is carrying the ranking on its own.
Where we are least certain
One bike is meaningfully weight-sensitive: the Priority Skyline Smart.Shift moves between #24 and #41 across defensible alternative weights. Everything in the top four does not move at all. We would rather tell you that than pretend to a precision we do not have.
Today’s top eight, with the arithmetic
| # | Bike | Price | Score | Power /16 | Battery /21 | Safety /14 | Ride quality /32 | Utility /8 | Display /9 | pts/$1k |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Tracer Monarc Bikes | $1,899 | 84.5 | 14.5 | 20 | 12.5 | 21 | 7.5 | 9 | 44.5 |
| 2 | Maker Monarc Bikes | $1,999 | 84.5 | 14.5 | 20 | 12.5 | 21 | 7.5 | 9 | 42.3 |
| 3 | Scrambler Full-Suspension Juiced Bikes | $1,899 | 66.5 | 16 | 13 | 12.5 | 15.5 | 3.5 | 6 | 35.0 |
| 4 | Summit 2 Velotric | $2,099 | 62.5 | 12 | 12.5 | 9 | 20.5 | 4 | 4.5 | 29.8 |
| 5 | Scrambler HardTail Juiced Bikes | $1,699 | 58.5 | 12.5 | 13 | 12.5 | 11 | 3.5 | 6 | 34.4 |
| 6 | Nomad 2 Velotric | $1,899 | 58.5 | 9 | 12.5 | 9 | 17.5 | 5.5 | 5 | 30.8 |
| 7 | XPedition2 Dual Long Range Lectric eBikes | $1,999 | 57.5 | 12 | 12.5 | 9 | 14 | 6 | 4 | 28.8 |
| 8 | RadWagon 5 Rad Power Bikes | $2,399 | 57.5 | 12 | 8.5 | 9 | 18 | 6 | 4 | 24.0 |
Rule changelog
Every change to the rulebook is versioned, dated, and applied to every bike at once. Live rulebook: Scoring Rules v3 (100-point) — Six consumer-facing categories, 27 scored metrics, half-point increments, no normalisation. Built from the CEO meeting of 3 Sep 2026, amended 17 Sep 2026 (motor cut-off demoted from an eligibility gate to a scored metric).
Scoring Rules v3 (100-point) adopted. Six categories, 27 metrics, half-point increments. Motor cut-off at the brake lever moved from an eligibility gate to a 2-point metric in Safety & Braking.
Questions we get
One of three reasons, and the product page will say which: the manufacturer changed the specification, we corrected a data error, or we changed a scoring rule. Rule changes are versioned, dated and applied to every bike at once — see the changelog above.
Probably not. A 48 is close to the middle of our board: the average bike we rate scores 45, and 50 of 87 are below 50. The scale deliberately reserves headroom at the top, so a score in the forties is an ordinary, well-equipped bike rather than a bad one. Scores describe equipment, not enjoyment, and equipment is not the same as fit, comfort or whether you want to ride it. Use the category subtotals to see whether the gap is in something you care about.
Because price is not an input. Across the bikes we rate, price explains only a small share of the variation in score; doubling the price buys roughly eight points on average, out of 100. That is a finding, not a bug.
The bike scores zero for that metric and the product page says “not published”. We do not estimate on a manufacturer’s behalf, because an estimate would be indistinguishable from a measurement to you.
Yes, on some outbound links, and we display advertising. Neither has any path into the scoring function, which takes 27 specifications and nothing else.