Buying Guides
Jacket Temperature Ratings Are (Mostly) Made Up — Here's How to Judge Warmth
Sooner or later every winter-jacket listing shows a number: rated to −20 °C, a "thermal index," a five-snowflake scale. Here's the uncomfortable truth: there is no industry standard behind any of it. Unlike sleeping bags — which have a real lab test (EN/ISO 23537) with defined comfort and limit temperatures — jacket temperature ratings are marketing artifacts each brand invents for itself.
Why a single number can't work
Whether you are warm at −20 °C (−4 °F) depends on variables no jacket label can know:
- Activity. Walking briskly, you produce several hundred watts; standing at a bus stop, a fraction of that. The same parka is sweltering for one and inadequate for the other.
- Wind. −5 °C with a 40 km/h wind strips heat faster than −15 °C in still air.
- What's underneath. A parka over a t-shirt and the same parka over a merino base plus fleece are two different systems.
- You. Metabolism, body size, and cold tolerance vary enormously between people.
A brand printing "−20 °C" on a hang tag is guessing at all four on your behalf. Some brands' guesses are honest; the point is you can't tell from the number.
Where the numbers actually come from
It's worth knowing what a rating can be based on, because a few brands do real work behind the marketing. Clothing insulation can be measured on a heated thermal manikin, which yields a clo value — the standard unit of clothing insulation, where 1 clo is roughly the insulation of a business suit, enough to keep a resting person comfortable at 21 °C (70 °F). A serious expedition brand may test on a manikin, pick an assumed activity level and wind speed, and convert the clo value into a temperature.
Two problems remain even then. First, the assumptions are almost never published — "comfort to −25 °C" might assume moderate walking with a fleece underneath, or standing still over a base layer, and those two scenarios are 15–20 degrees apart in practice. Second, most brands don't test at all; they benchmark against a competitor's claim or let the marketing department pick a number that sounds right for the price point. From the outside, a manikin-tested rating and an invented one look identical on the hang tag.
The specs that actually predict warmth
Ignore the temperature claim, read these instead:
| Spec | What it tells you | What to look for |
|---|---|---|
| Fill weight (grams of down) | The single best warmth predictor for down | Everyday winter jacket: 100–200 g. Serious parka: 200–300 g+. If it's not listed, ask — or assume it's low |
| Fill power (550–900) | Quality/efficiency of that down | Warmth ≈ fill power × fill weight. Both numbers or no conclusion |
| Synthetic insulation weight (g/m²) | Warmth of synthetic layers | 60 g/m² = cool-weather layer; 100–133 = winter; 200+ = deep cold, low activity |
| Length | Coverage matters as much as loft | A knee-length parka at 150 g of down outperforms a waist-length jacket at 200 g for standing around |
| Hood, storm cuffs, hem drawcord | Air-sealing features | Warm air escaping at the openings is the biggest real-world loss |
| Shell fabric | Wind and wet protection | Windproof shell can add the equivalent of a whole midlayer in exposed conditions |
Fill power alone is the most misread spec on the table. A 900-fill jacket with 90 g of down is a superb light jacket — lofty, compressible, expensive — and noticeably colder than a 650-fill parka carrying 250 g. High fill power means the down does its work with less weight and bulk, not that there's more warmth in the garment. Brands lead with fill power precisely because it's the impressive-sounding number; fill weight is the one they bury.
The baffle check for down jackets
Look at the construction between the puffy chambers:
- Sewn-through baffles (stitching goes straight through both fabric layers) create cold lines at every seam. Fine for town, cheaper to make.
- Box-wall baffles (a fabric wall separates inner and outer, keeping full loft over the seam) are what real expedition parkas use — warmer at the same fill weight, and a reliable tell that a brand is building for warmth rather than looks.
- Hybrid construction — box-wall through the torso, sewn-through in the sleeves — is a sensible middle ground and common on good parkas. The core is where the construction matters most.
On a cold day you can feel the difference: run a hand along the seams of a sewn-through jacket and each stitch line is a cold stripe.
Translate your actual use case into specs
Three common scenarios, converted into the numbers that matter:
- School run, commute, errands in a −5 to −15 °C (23 to 5 °F) winter. 120–180 g of 600+ down (or ~100–133 g/m² synthetic), hip length or longer, hood, cinchable hem. Sewn-through construction is fine at this activity level.
- Standing around in real cold — touchline parenting, dog walks at −20 °C, outdoor work with long still periods. 200 g+ of fill, box-wall or hybrid construction, mid-thigh length, insulated hood, storm cuffs. This is the use case brands' optimistic ratings fail hardest, because standing still is the lowest heat output you'll ever have outdoors.
- High-output cold — winter cycling, snow clearing, brisk hiking. Deliberately less insulation (60–100 g/m² synthetic or a light down midlayer under a shell), because you're producing the heat yourself and sweat-soaked insulation is worse than none. A "−30 °C rated" parka is the wrong tool here regardless of how honest the rating is.
How to use brand ratings anyway
Within one brand, ratings rank the range usefully: the brand's "−30 °C" parka is warmer than its "−10 °C" jacket. Across brands, the numbers are incomparable — one brand's −20 assumes a wool base layer and walking pace; another assumes standing still in a t-shirt (almost none publish their assumptions). Use ratings to sort a single brand's catalogue, then verify the shortlist with fill weight, construction and length.
A realistic buying checklist
For a "stand around in real winter" parka, the listing should show:
- Fill power and fill weight (or g/m² for synthetic) — a brand hiding one of them has a reason
- Box-wall or hybrid construction if you're paying real money
- Hip coverage or longer, insulated hood, cinchable hem and cuffs
- Windproof (ideally water-resistant) shell — see our down vs. synthetic guide for the wet-climate caveats
- Room inside for a midlayer without compressing the fill
A jacket that ticks those five will handle deep cold regardless of what temperature is printed on the tag. A jacket that ticks none of them isn't rescued by the boldest claim.
FAQ
Why do sleeping bags get a real standard and jackets don't? A sleeping bag is a closed, predictable system: one person, lying still, on a defined mat, in defined sleepwear — conditions a lab can reproduce. A jacket is worn over unknown layers at unknown activity in unknown wind, so no test protocol has ever stuck.
Is a rating from a serious expedition brand more trustworthy? Somewhat — brands that outfit polar expeditions tend to rate conservatively because their reputation rides on it. But "more trustworthy" still means "calibrated to unstated assumptions." Check the fill weight anyway; a trustworthy brand will happily publish it.
My jacket is rated −25 °C and I'm cold at −10. Is it defective? Almost certainly not. You're likely standing still (scenario the rating didn't assume), losing heat at the hem and cuffs, or wearing it over a shirt. Add a midlayer, cinch the hem, close the hood — sealing the openings is worth more than another 50 g of down.
Do the snowflake / 1-to-5 warmth scales mean anything? Only as a ranking within that retailer's own range, same as temperature numbers. A "5/5 warmth" own-brand jacket with no published fill weight tells you nothing a −30 °C claim wouldn't.
The one-line answer
Temperature ratings on jackets are unregulated marketing. Judge warmth by fill weight × fill power (or g/m² synthetic), construction, length and seals — and by whether you'll be moving or standing still while wearing it.
