Bike Helmet Sizing, Fit and Safety Ratings
Three questions decide whether a helmet does its job: does it fit your head, what is it actually certified to, and how old is it. None of them are answered on the marketing page.
On this page
A helmet is one of the few pieces of cycling equipment where the specification sheet tells you almost nothing useful. Weight and vent count are easy to print. Whether the shell matches the shape of your skull, what the certification sticker inside actually covers, and how much of the foam is already spent are the things that decide the outcome of a crash, and none of them appear in a product listing.
This guide covers all three. It is written to be the reference the rest of our helmet coverage points back to, so it goes further into the standards than a buying guide would. If you are still choosing a category rather than a size, start with the bike helmet buying guide and come back here once you have a shortlist.
Measure your head, properly, once
Take a soft tailor's tape and wrap it around the widest part of your head. That is roughly one inch, or 2.5 cm, above your eyebrows at the front, and above your ears at the sides, sitting on the bulge at the back of the skull rather than sliding down onto your neck. Keep the tape level all the way round. Snug, not tight.
No soft tape? Wrap a shoelace or a strip of string around the same path, mark where it overlaps, then lay it flat against a ruler. It is accurate enough. Measure twice, because a centimetre of error is enough to put you in the wrong band on some charts.
Write the number down in centimetres. Every helmet brand publishes its ranges in centimetres, including US brands, because the letter sizes are marketing shorthand and the range is the real specification. If you are buying for a child, measure them again before every purchase rather than working from age, since head circumference and age correlate loosely and badly. That point is worth repeating in our kids helmet guide because it is the single most common sizing mistake parents make.
The size chart, and why it is only a starting point
| Size | Head circumference | Inches | Notes |
|---|---|---|---|
| XS / youth | 46 to 51 cm | 18.1 to 20.1 | Toddler and small child shells |
| S | 51 to 55 cm | 20.1 to 21.7 | Also sold as youth large |
| M | 55 to 59 cm | 21.7 to 23.2 | The most common adult size by a distance |
| L | 59 to 63 cm | 23.2 to 24.8 | Where universal-fit shells start to run out |
| XL | 61 to 65 cm | 24.0 to 25.6 | Overlaps L, and not every brand offers it |
| XXL | 63 to 67 cm | 24.8 to 26.4 | Rare, a handful of brands only |
Swipe sideways to see all columns →
Notice that L and XL overlap. That is not a mistake in the table, it is how the industry actually works. Bands are set per model, they shift between a brand's road and mountain lines, and they change when a shell is redesigned. Someone measuring 62 cm may be a large in one brand and an extra large in another, and both fit.
So treat the letters as a filing system and the centimetre range as the specification. If your measurement lands near the boundary between two sizes, the general rule is to take the smaller one if the dial still has adjustment left when it feels snug, because a retention dial can tighten a helmet that is slightly large but cannot make a big shell hug your head evenly. If you wear a thin cap or a winter liner underneath in cold weather, size for the cap.
One more category worth understanding: universal fit. Some budget and mid-range helmets use one shell across a wide range, letting the internal band do all the work. It cuts the number of moulds a brand has to pay for, which is how a $50 helmet exists. It also means the fit is competent rather than excellent for most heads, and poor at the extremes of the range.
The 2-4-1 fit check
Circumference gets the helmet onto your head. This is the sequence that tells you whether it is doing anything once it is there. Run it in the shop, with the helmet on, before you pay.
Two fingers above the eyebrows
The front rim should sit about two finger widths above your brow line, level front to back. The most common error in the world is wearing a helmet tipped back on the crown, which looks casual and leaves the forehead completely exposed. Forehead and temple impacts are extremely common, so this is not a cosmetic point.
Straps in a V under the ears
The front and rear straps should meet at a splitter just below and slightly forward of the earlobe, forming a narrow V with your ear inside it. If the splitter sits over your ear or well behind it, slide it. A correctly placed splitter is what stops the helmet rotating forward over your eyes or backward off your skull.
One finger under the chin strap
Buckle it, then tighten until you can slide one finger between the strap and your chin, and no more. Open your mouth as wide as you can: you should feel the helmet pull down against the top of your head. If nothing happens, the strap is loose.
Then try to move it
Grab the helmet with both hands and try to rotate it forward, backward, and side to side. Your scalp should move with it. If the helmet slides independently, tighten the retention dial first. If it still slides at a tension that is uncomfortable, the shell is the wrong shape for your head and no amount of tightening will fix that. Put it back and try another model.
Large heads, and why shape matters more than size
Anything above roughly 61 cm is into the top few percent of adult head sizes, and it is where the market thins out fast. Three practical points.
Measure before you assume. A surprising share of people who go looking for helmets for big heads measure 59 or 60 cm and are a straightforward large. The perception often comes from having tried on universal-fit helmets that top out around 61 cm and having nothing fit properly.
Skip universal fit above 60 cm. A single shell stretched across 54 to 61 cm is being asked to do too much at the top of its range. At that point the retention band is fully open, the shell perches rather than surrounds, and the coverage at the back of your skull is worse than the same helmet gives a smaller head. Buy from a brand that moulds a genuine XL, and prefer mountain and commuter lines, which more often extend to XL and XXL than aero road lines do.
Shape decides comfort, and it is not on the box. Skulls vary in the ratio of length to width. A longer, narrower head that is forced into a rounder shell gets a hot spot in the middle of the forehead. A rounder head in a longer shell gets gaps at the temples and a helmet that rocks side to side. Motorcycle helmet brands publish head shape openly; bicycle brands almost never do, which leaves you with two options: try helmets on, or buy from retailers with a genuine returns policy and try them at home.
Riders commonly report that the fit character is consistent within a brand, so if one model from a maker presses on your forehead, others from the same maker probably will too. That is a more useful shopping heuristic than any published shape chart, because it lets one bad fitting session rule out a whole catalogue.
Low profile helmets deserve a note here too, since they are often searched for by people who dislike how bulky a helmet looks on them. Low profile describes a shell moulded tight to the head rather than one that has had protection removed. It is achieved with denser foam and a closer shell, which is why low profile models are often slightly heavier for the same coverage. The thing to actually inspect is where the shell ends at the back of your skull.
The certifications that exist, and what each one means
This is the part that gets muddled most often, so it is worth being exact.
- CPSC 16 CFR Part 1203
- Mandatory in the US. Every bicycle helmet legally sold in the United States must pass it. Drop tests onto flat, hemispherical and kerbstone anvils after hot, cold and wet conditioning, with a ceiling on peak head acceleration, plus a roll-off test for positional stability and a retention strength test. Pass or fail. There is no score and no grade.
- EN 1078
- The European bicycle helmet standard, required for CE marking. Slightly lower drop heights than CPSC but a lower acceleration ceiling, so the two are not simply ranked against each other. Also pass or fail.
- NTA 8776
- A Dutch standard written for speed pedelecs. Tests at higher impact energy than EN 1078 and requires more shell coverage at the rear of the skull and around the temples, plus a stronger retention system. Voluntary in the US, and the right thing to look for if you ride at sustained high assisted speeds.
- ASTM F1492
- The skateboard helmet standard. Built around repeated impacts to the same helmet rather than one large one, which drives a different foam and shell choice. Not a substitute for CPSC on a bicycle unless the helmet carries both markings.
- Snell B-90 / B-95
- A voluntary standard from the Snell Memorial Foundation, historically stricter than CPSC. Rare on modern bicycle helmets, and its absence is not a red flag.
- AS/NZS 2063
- The Australian and New Zealand standard, mandatory there. Helmets sold in those markets carry it in place of, or alongside, CPSC.
Three consequences follow from that table, and they are the ones that change how you shop.
CPSC is a floor, not a ranking. Because it is pass or fail, it cannot tell you that one certified helmet is better than another certified helmet. Any marketing that implies a helmet is "more certified" is meaningless. What CPSC does tell you is that the helmet is not a novelty shell.
The standards measure linear impact. They drop a headform straight down and measure the deceleration. Real crashes are usually angled, and angled impacts rotate the head, which is a mechanism these tests were never designed to capture. That gap is the entire commercial reason rotational liners exist.
Speed changes the maths. Kinetic energy rises with the square of speed, so a crash at 28 mph carries roughly twice the energy of the same crash at 20 mph. CPSC 1203 is calibrated to bicycle impact speeds. If you ride a Class 3 ebike that sustains 28 mph, you are asking a standard bicycle helmet to work outside the envelope it was certified in, which is exactly the problem NTA 8776 was written to solve. The argument is laid out properly in our ebike helmet guide, and if the class system is new to you, start with ebike classes explained and how fast ebikes actually go.
MIPS, and the ratings that do give you a score
What MIPS is
MIPS stands for Multi-directional Impact Protection System. It is a Swedish company's product: a low-friction layer fitted between the helmet liner and your head that allows a small amount of relative movement during the first few milliseconds of an angled impact, with the goal of reducing the rotational forces transmitted to the brain. Brands pay to license it, and it typically adds $15 to $25 to the price of an otherwise identical helmet.
The engineering rationale is credible and the cost is small, so it is easy to recommend. What must not be blurred is the category it belongs to. MIPS is not a certification. No standard requires it, no government body verifies a helmet because of it, and its presence on a box says nothing about how that helmet performed in any independent test. A yellow MIPS dot means a licence fee was paid.
The same applies to the competing systems: WaveCel, the collapsing cellular structure used by Trek and Bontrager; Koroyd, the welded tube structure several brands use; Spherical, Giro's two-layer ball-and-socket implementation of the same slip idea; and various in-house rotational liners. All are engineering approaches to the rotational problem. None is a certification.
The Virginia Tech ratings
The Virginia Tech Helmet Lab is the most useful public comparison data available, for one structural reason: it produces a score rather than a verdict. The lab buys helmets at retail, drops them onto an angled anvil at multiple impact locations and at more than one speed, and measures both linear and rotational acceleration. Those results are combined into a single number that estimates relative concussion risk, where lower is better, and that number is converted into a one to five star rating.
Because it scores, it lets you do the thing CPSC cannot: rank one certified helmet against another certified helmet. It is also independent of the manufacturers and it publishes the full table publicly.
Two cautions. First, the ratings are relative, not absolute; a four star helmet is not unsafe, and every helmet in the table already passed CPSC. Second, the table is updated continuously as new helmets are tested and old models are dropped, so any specific star count quoted in an article ages fast. Look up the current table yourself before you buy, and use it to break a tie between two helmets that both fit you well. Do not use it to override fit, because a five star helmet that rocks on your head is worse than a four star one that does not.
Skate helmets and bike helmets are not interchangeable
They look similar enough that people swap them, and the difference is genuinely important because it comes down to how the two are meant to fail.
A bicycle helmet is a single-impact device. Its expanded polystyrene foam absorbs energy by crushing, permanently, converting your deceleration into destroyed foam. That is the whole design. It works once, and then it is finished, whether or not you can see the damage.
A skate helmet certified to ASTM F1492 is built around a different assumption: that you will hit your head repeatedly at lower energy over the life of the helmet, coming off a board on concrete again and again. That drives a hard outer shell and a firmer, more resilient foam that recovers its shape instead of crushing. The trade is that it does not manage a single large impact as well.
| Bike helmet (CPSC 1203) | Skate helmet (ASTM F1492) | |
|---|---|---|
| Designed for | One large impact | Repeated smaller impacts |
| Foam behaviour | Crushes permanently | Largely recovers |
| Shell | Thin, often bonded in-mold | Hard ABS outer |
| Coverage | Varies by category | Low and even, rear and sides |
| Ventilation | Much better | Poor |
| Weight | Lighter | Heavier |
| Legal for cycling in the US | Yes | Only if it also carries CPSC |
Swipe sideways to see all columns →
The practical answer is simple: plenty of skate-styled helmets are dual certified to both CPSC 1203 and ASTM F1492, and those are fine for cycling. Turn the helmet over, look inside the shell, and read the sticker. If it lists CPSC, it is a legal and legitimate bicycle helmet whatever the styling. If it does not, it is not, regardless of how tough the shell feels in your hands. The same logic applies to the extended-coverage helmets increasingly sold for scooters and ebikes, and to anything marketed as a multi-sport helmet.
When to replace it
After any impact, without exception
If you crashed and your head hit anything while wearing the helmet, replace it. Not because it looks damaged, but because it very often will not. The foam crushes under the shell, and a thin in-mold outer shell can spring back over a compressed area and hide it completely. A helmet that has taken one impact has already spent the protection you paid for, and the compressed area will not perform again.
Many manufacturers run crash replacement schemes offering a discounted replacement within a few years of purchase if you send in the damaged helmet, so check before you bin it.
The dropped-helmet question
Knocking an empty helmet off a table or dropping it in a car park is a different case, and the honest answer is that it is usually fine. The energy in a crash comes overwhelmingly from the mass of your head and body behind the helmet. An empty shell falling a metre has very little behind it. Inspect it rather than replacing it reflexively.
Three to five years otherwise
Most manufacturers advise replacing an undamaged helmet every three to five years. That guidance is not about the polystyrene degrading on a shelf, which it does very slowly. It is about everything around the foam:
- UV. Sunlight degrades the shell polymer and, more importantly, the strap webbing. A helmet that lives on a hook by a sunny window ages faster than one in a cupboard.
- Sweat and hair products. Salt and product residue attack webbing, glue lines, and the adhesive holding the shell to the foam. Rinse the straps and pads with plain water occasionally, and never machine wash a helmet or clean it with solvents.
- Compression. The comfort padding packs down over a couple of seasons, which quietly loosens the fit. Most brands sell replacement pad kits for a few dollars, and a new pad set can restore a helmet that has started to feel sloppy.
- Heat. Leaving a helmet on a parcel shelf in summer, repeatedly, is the fastest way to cook the glue and warp the shell.
- Brittle plastic. Buckles, splitters and retention dials are the parts that actually snap on old helmets, and they snap without warning.
How to inspect one
- Pull the pads out and look at the foam underneath in good light. You are hunting for dents, hairline cracks, and any area that looks compressed relative to the rest.
- Flex the shell gently. Stress lines, crazing, or any place the shell has lifted away from the foam are all reasons to retire it.
- Run the webbing through your fingers for fraying, stiffness, or discolouration, and check the buckle clicks firmly and releases cleanly.
- Work the retention dial through its full range. Grinding or slipping means the cradle is done.
- Find the date stamp. It is moulded into the foam or printed on a label inside, and it records manufacture, not purchase. A helmet that sat in a warehouse for two years is already two years old.
Once the helmet fits and you trust it, the remaining decisions are about what goes on it and where you are required to wear it. Mounts, covers, visors and winter liners are covered in helmet accessories worth buying, integrated lights and crash detection in smart bike helmets, and the rules, which vary far more by state and by age than most riders expect, in bike helmet laws by state. If you have not chosen a category yet, go back to the main helmet buying guide or straight to mountain bike helmets.