How we judge and rank

A ranking nobody can argue with is a ranking nobody should trust. Everything below is how ours is assembled, written out so that it can be taken apart.

One habit holds the rest together

Every family is built the same way. Read what the maker committed to print. Put it next to the standard that defines the figure. Listen to people who have owned the thing through a winter. Then write down, next to every number, where it came from.

That last habit carries the weight. A number with a source can be argued with; a number without one has to be believed. Where a manual and a sales listing disagree, the manual is what appears here. Where neither says anything, the field stays empty and the page says it is empty — because once a plausible substitute has been set in type, nothing on the page separates it from a real figure.

Six published facts, taken in a fixed order

Anyone buying on this shelf is weighing three worries: whether it will fit — the machine or the body — whether it will hold up, and whether it will still be serviceable in five years. Six published facts settle those, in this sequence:

• First, the figure that decides whether two things meet at all. The pitch and link count a chain kit is cut for. The sprocket bolt pattern. The visor mount a brand uses, which almost never crosses to another. The pad shape a caliper takes. The bar diameter a mount clamps. The valve angle a gauge can reach. It leads because it is why most of these products are sent back, and because no photograph has ever answered it.
• Second, the headline number, carried together with the condition that makes it mean anything. A shell weight, reported with the size it was weighed at when the maker says so, and reported as unstated when they do not. An armour level, which is the outcome of a transmitted-force test and not a grade of quality. An oil viscosity, held apart from the specification that actually decides clutch compatibility. A lock's certification, held apart from the diameter of its shackle. When a maker prints the headline and omits the condition, the omission is what gets recorded.
• Third, the small thing that decides whether it is still in use in five years. Whether the liner comes out and washes. Whether replacement visors are still catalogued for a helmet three seasons old. Whether the chain tool takes replaceable pins. Whether the intercom's firmware is still updated. Every family has one, and that is the one written out.
• Fourth, the parts the box turns out not to contain, and what going without them costs. The anti-fog insert a visor is cut to accept but ships without. The rivet a chain kit does not include. The mounting plate a top box needs before it meets any rack. The thread locker a sprocket bolt asks for. This line carries deliberate weight: anything requiring a second purchase before it works was never the price on the label.
• Fifth, what it costs to keep using. Whether consumables are made by anyone but the original brand. Whether a cheek pad thickness can be changed rather than the whole helmet. Whether a chain is O-ring, X-ring or plain, and what each asks of the lubricant. Products that lose on the headline figure regularly climb above stronger ones on this line alone.
• Sixth, whether you can actually buy it from where you live. Store by store, every reference is confirmed before publication, and each page lists the markets that had an offer. Something excellent that nobody stocks in your country is not a recommendation.

Three kinds of source, weighted unequally on purpose

For anything structural — pitch and link counts, pad references, thread sizes, bar diameters, visor fittings, what an accessory claims to fit — the manufacturer's own documentation comes first. For anything printed as a headline it comes last, because a headline exists to sell.

Second come approvals and markings: an ECE version, an armour level, a viscosity grade and its published specification, a lock's certification, a tyre's load and speed index. These are treated as things to confirm against the current listing rather than as permanent properties, since a maker can revise a liner, a vent or a supplied insert without changing a model name — and an approval version is revised on its own schedule.

Third, and irreplaceable for one question, come owners a year or two in. No specification sheet says whether a visor mechanism still seals after a winter of salt, whether a zip has given way, whether a lock's barrel freezes, or whether anyone still answers the support address. When a judgement rests mostly on one of the three, the page names it. When none of them answers, the page says so rather than reasoning its way to a figure.

Places in a family, not marks out of ten

Nothing here is scored. A single figure would have to trade a touring helmet against a can of chain lubricant, and there is no honest exchange rate between them. What appears instead is a place in a family, with the sentence that earned it printed next to it.

Families run upward from the modest rather than downward from the expensive. Almost everyone arrives with a sum in mind and a specific gap in their kit; a page that opens on the most expensive item on the shelf is answering a question almost nobody asked.

What the income here pulls towards

Affiliate income leans towards whatever is expensive, newly launched and easy to write about. What resists it is arrangement rather than good intentions: no family has ever been reordered for what something earns, several pages end by recommending a cleaning routine or a replacement visor instead of a new helmet, and no position has ever been exchanged for anything at all.

There is one distortion peculiar to this subject, and naming it works better than quietly resisting it. Protective equipment is sold on fear. Build a page around that and it finishes by recommending the most expensive item in every family, since no amount of spending can ever be argued as too much. Nothing here is built that way. Where the honest answer is that an upgrade changes nothing measurable — a second armour level in a jacket whose pockets do not hold the armour in place, a chain kit heavier than the sprocket carrier it bolts to, a lock defeated by the rail it is fastened to — that is the answer the page gives.

One bias cannot be arranged away: what is covered here is what can be bought through the Amazon stores of the United States, the United Kingdom, Germany, France, Italy and Spain. Equipment sold only through dealers, or fitted under warranty, is under-represented as a result, and where that shapes a recommendation the page says so.

Figures are quoted, never derived here

This subject is sold on numbers, and a number is the easiest thing in a listing to repeat without thinking. One rule covers every page carrying one: every weight, shell size, armour level, approval version, viscosity grade, chain pitch and link count, pad reference, disc diameter, torque figure, load index, capacity in litres and quoted temperature on this site is copied from the manufacturer, in the unit it was published in, with its source named — and none of them is recalculated, converted or inferred by us. Where a maker publishes nothing, the page says nothing has been published rather than offering a substitute.

Chain slack is where that rule earns its keep, because the figure is meaningless on its own. The same chain reads differently with the machine on its side stand, on a paddock stand, and with a rider seated, and manufacturers specify which one they mean. So this site reports the figure together with the position it belongs to, and never the figure alone.

What an approval actually undertakes

One idea sits underneath all of the above, and it is why these pages read more flatly than the ones beside them in a search result: an approval records a test that was passed; it does not undertake an outcome. ECE 22.06 describes impacts at defined points and defined speeds on a sample of a model. EN 1621 describes a force transmitted through a piece of armour on a rig. A lock's certification describes an attack, with named tools, for a named duration. Each is a real, repeatable measurement, and none of them is a statement about a particular crash, a particular theft or a particular rider. Treat any of them as a promise and you sell more equipment, at the cost of misleading somebody about what they bought.

Two sentences that never appear here

Two limits are fixed, and both are better stated than left to be inferred. Nothing on this site — no helmet, no armour, no airbag, no lock — is described as making anything safe. A page reports what the maker specifies for its own product: the approval a listing cites, its version and date, the construction, the fitting, the inspection asked for before use, the replacement interval. The rest belongs to those instructions, because whether equipment protects depends on how it fits, how it is fastened and what it meets — and no listing knows any of that. And a stated size is never presented as a fit: a circumference is a published measurement, an internal shape is not, and a ranking that blurred the two would be recommending a fit it has no way of knowing.

Where a page stops

None of this stands in for the manufacturer's instructions. For a helmet's replacement interval, for an airbag's service and cartridge, for a torque figure on a safety-critical fastener, for the bedding-in of new pads and for the specification an engine's clutch requires, the page points at those instructions instead of rewording them — even though a page that stops there earns nothing.

A third limit is repeated on every page it touches, rather than stated once and assumed: whether two things fit is settled by a figure or a list the maker published, and never by how they look. The pitch and link count decide which chain a machine takes. The bolt pattern decides which sprocket mounts. The brand's own fitting decides which visor closes on a helmet. The plate decides which top box meets which rack. The bar diameter decides which mount clamps. Where a maker publishes none of that, the page reports the answer as unpublished rather than deducing it from a photograph or a product name.