Skip to content

How we verify what we publish

Source labels describe provenance, market and test standard where recorded. A page update date does not certify a personal check of every figure. This page explains what those labels mean, how we grade Chinese-market data, and what we refuse to print.

How we label figures

Every figure carries one of five labels, and the label says where the number came from.

UK tested. Measured in the UK by an outlet we name and link: What Car?, Autocar, Carwow, Electrifying and others.

Official figure. The manufacturer's homologated UK figure or its published on-the-road price: WLTP range from the UK specification sheet, the UK price list, Euro NCAP results, DVSA notices.

Manufacturer figure. A manufacturer claim that has not been homologated for the UK, typically a Chinese CLTC range or a charging time from the China-market brochure. We print it as a claim, never as a UK number.

Estimate. Our own reading, worked out from the fixed factors below and shown with the method. Never a substitute for a road test.

No data. Nothing credible has been published yet. We leave the row out, or say plainly that there is no figure.

Check software on the UK car

A feature shown in China may depend on different hardware, software, accounts or connected services. Check the UK release notes and the exact car rather than assuming a later update will add it.

Can You Leave the Key at Home? UK EV Software Reality | Deepal S05 & XPENG G6 · 5:19

Watch the updated film on YouTube · Original edition · 9 Oct 2026 · Explore the UK XPENG G6 · Read the driver-assistance guide

Range: CLTC, WLTP and what you will actually get

Chinese specification sheets quote CLTC; UK listings quote WLTP. We never compare the two directly and never mix them on one page. Our real-world estimates apply one set of factors site-wide, derived from owner-reported seasonal driving: mild weather WLTP × 0.85, mixed driving WLTP × 0.78, cold weather with heating WLTP × 0.70. The factors are reviewed each autumn against the latest owner data and the change is recorded on the corrections page.

How a page is put together

AI-assisted research gathers Chinese owner reviews and pricing. New pages pass automated publication checks; proposed material changes to published figures are referred for a human decision.

Owner reviews and complaint records from 汽车之家 (Autohome) and 车质网 (12365auto) are crowd-sourced, so we always state the sample size, and a fault pattern must appear in two sources before we call it a pattern rather than a signal. They shape what a page says about a car; they never carry a figure label of their own.

Where those owner reports and complaint records are counted rather than quoted, they become a dataset of their own: the owner range ledger and the fault and recall archive. How both are built sets out the sample rules, what is thrown out and what they cannot tell you.

Local-language news briefings

Short, market-specific explainers separate company announcements, media reporting and our own interpretation. A briefing does not confirm local specification, pricing or availability.

What we will not publish

A byline or source label presented as proof that a person has checked every figure. A Chinese owner figure without its sample size. A fault claim that rests on one anonymous post. A verdict we would not defend to the manufacturer.

Where we have no verified figure, the row is left out rather than filled with a guess.

Independence

Cathay EV does not accept payment, cars, travel or hospitality from manufacturers or dealers. Any commercial link on a page is labelled as such and sits apart from the editorial text. Recorded material corrections are listed on the corrections page, whose historical coverage is incomplete.