How we work out our MOT figures
Data to December 2025. Last updated 16 November 2026. Responsible for the data: the Checks team.
Our make and model pages show what MOT records say about each model and year of car. This page explains where the figures come from, how we count them and what they can't tell you, so you can judge them for yourself.
The data
Every figure comes from DVSA's anonymised MOT tests and results: a record of every MOT test in Great Britain, published each year by the Driver and Vehicle Standards Agency. It includes the make, model, fuel, engine size, date first used, mileage, result and the reasons for any fail or advisory. It doesn't include registrations or anything about owners.
We use tests from 2024 to 2025. Northern Ireland tests aren't included, because they aren't in this data.
Contains public sector information licensed under the Open Government Licence v3.0. DVSA hasn't reviewed or endorsed our analysis.
What we count
- Cars and light vans (MOT test class 4, up to 3,000 kg).
- "Year" means the year the car was first used, not its model year. A car built in late 2015 and first registered in 2016 counts as 2016.
- First attempt only. Pass rates use normal tests, not retests. A car that failed and was fixed at the test station during the same visit is counted as a fail.
- Failure and advisory areas are DVSA's groups (brakes, tyres, suspension and so on). We count tests with at least one item in an area, so a test with three brake items counts once.
- Like for like. Every figure is shown next to the same figure for all cars first used in the same year, so age doesn't distort the comparison.
Model names
DVSA records the model as it was typed at the test, so one model can appear under many spellings and trim names. We map these to one model name using a list we maintain and check by hand. We never combine genuinely different models to make the numbers bigger. Each page lists which model names it includes, where that isn't obvious.
Minimum numbers
Small samples give misleading percentages, so we only show figures with enough tests behind them.
- We don't make a page for a model and year with fewer than 300 tests in 2025.
- We don't ask search engines to list a page with fewer than 1,000 tests.
- A section is hidden, with "not enough data", if its own count is below 300.
- With 1,000 tests, a pass rate near 50% could be out by about 3.1 percentage points either way. Bigger samples are more precise.
"What stands out"
Each page lists findings only where the model differs from all cars first used the same year by more than twice the combined margin of error of the two figures, and by at least 3 percentage points. If nothing qualifies, the page says so. We don't hunt for negatives, and we don't use words like "reliable", "unreliable", "best" or "worst".
Mileage
- Typical mileage is the median recorded at MOTs in 2025: half of cars were higher, half lower.
- Miles a year is worked out from each car's own readings between one MOT and the next, then the median across cars.
- Readings that are zero, missing or unreadable are left out. Readings taken in kilometres are already converted to miles in DVSA’s data (it does this for 2022 onwards).
Readings that went down
We count cars where a later MOT reading was lower than an earlier one by more than 1,000 miles. This is not a count of clocked cars. A drop can be a typing slip at the test, a replaced instrument cluster or a reading in kilometres. It's a reason to check the individual car, not a conclusion about it.
What these figures can't tell you
- Anything about a particular car. For that, check its own MOT history.
- Repair costs, breakdowns or reliability between MOTs.
- Servicing, accidents, outstanding finance or write-offs.
- How well individual cars were looked after. A model's MOT record reflects its owners as well as its design.
Updates
We update the figures once a year, when DVSA publishes the next year's data. Dates on each page change only when the figures or wording change.
The figures on these pages
- Pass rates, failures, advisories, dangerous defects, mileage and fuel use normal tests in 2025 with a result of pass, fail or pass after rectification. Abandoned and aborted tests are left out.
- “Advisories that became fails”, “miles a year” and “readings that went down” follow each car from one normal test to its next, across 2024–2025.
- Miles a year uses pairs of readings at least 180 days apart where the reading didn’t go down.
- A difference in miles a year is listed under “What stands out” when it’s at least 15% of the figure for all cars first used the same year, from at least 300 cars.
- Cars with a first-use date of 1971 are left out: DVSA uses that date when the real one isn’t known.
- Advisories that DVSA doesn’t tie to one area of the car aren’t ranked.