How we write your report

No AI writes your report

When you check a car, no AI writes anything. Every sentence comes from a library of wording a person has written and checked. It happens in three steps.

1. We read the facts in the MOT record: the fails, the advisories, the mileage readings and the gaps between tests.
2. We pick the sentences that match those facts, from a library of wording a person has written and checked.
3. We put the report together the same way every time.

• • Same car, same day, same words: the same car checked on the same day always gives the same report. Two people who check the same number plate that day see the same words.
• • Words we never use: a check before every release blocks wording that tells you what to do with the car, or that blames its past owners. The decision stays with you.

Why ranges, not single figures

Where the report mentions money, such as running costs or likely repair bills, we always show a low-to-high range, not a single number. A range is honest about what one MOT record can and can't tell us.

• • For planning, not a quote: the figures are our best estimates from public data and our own models. They aren't firm prices.
• • Sourced: every figure says where it comes from and how we worked it out, linked from its card.
• • An honest spread: the gap between low and high allows for what we can't know. That includes who drives the car, how far, and what fuel costs.
• • Repair prices: typical independent-garage prices, parts and labour, from UK price guides and booking sites during 2026. A few rest on owner reports. Many were checked in October 2026 against booking sites' published averages (FixMyCar, ClickMechanic, BookMyGarage). Others come from repair price guides (Bumper, car.co.uk), retailers' fitted prices (Autodoc, Blackcircles), a car-buying firm's guide (webuyanycar) and welding price guides (getcostidea, MyBuilder). We also use the What Car? garage labour-rate survey (April 2023: £76 an hour on average, £47 to £141 by area). The table below shows the named repairs. Other work, such as brake pads, discs, tyres, bushes, CV boots, steering, exhausts, bulbs and wipers, uses our earlier rate card from the first half of 2026.

The decision stays yours. The report's job is to lay out what the data shows, not to make the call for you.

Where the repair prices come from

Each price on the running costs tab is a range for one repair, at an independent garage, with parts and labour. Each range aims to cover the usual repair on a common car. A very cheap fix, or a premium part, can fall outside it.

Many of these prices were checked in October 2026. Others come from our earlier rate card. A few are our own estimates, as marked in the table. Prices vary with the car, the garage and the area, so treat them as a guide, not a quote.

How the risk rating is worked out

The rating you see (Very Low to Very High) comes from a score built from the car's full MOT history. Each factor below adds or takes away points. The total is put on a 0 to 100 scale and mapped to one of five levels.

• Dangerous faults: Anything a tester marked as dangerous. Faults on the last two MOTs count the most.
• Parts that failed more than once: The same part failing on more than one MOT. Parts that keep the car safe count the most.
• Warnings left unfixed: An advisory in the same area coming up on three or more MOTs in a row. An advisory is a warning that did not fail the test.
• Number of advisories: Lots of advisories on one MOT or on the latest MOT, and whether the count is rising.
• Recent MOT fails: How many MOTs ended in a fail. Recent fails count for more than old ones.
• Mileage readings that don’t add up: Readings that go down, a year or more with almost no miles, or sudden big jumps.
• Gaps and MOT status: Stretches with no MOT on record, and whether the car has a current MOT.
• Old warnings not listed at the latest MOT: Advisories noted at two or more MOTs in a row that were not listed at the latest MOT.
• Age and miles a year: Older cars score a little higher. Very high or very low miles a year add a little.
• Critical recalls for this model: Recalls we grade critical for this model and age add a few points.
• A steady record: Passes with few or no fails, no gaps and steady mileage bring the score down.
• Similar cars (cars only): How cars of the same make, model, fuel type and age do in national MOT results. These figures are about the model, not this car, and they move the score up or down. Vans, motorcycles and minibuses are scored without it.

Every report links to its score page, under the rating. That page shows the score and each factor behind it, with its points. So you can see which factors drove that car's rating. How MOT Wise works explains where the data comes from.

How we count fail rates on the car pages

A fail rate on our make, model and generation pages is the share of MOT tests that ended in a fail. A retest counts as a test of its own. So a car that fails, is fixed and passes a week later adds one fail and one pass.

That makes the figure a few points lower than the share of cars that fail at their first attempt. That share is the figure on our MOT failure rates page. Both come from the same DVSA records. They answer different questions, so they are never compared with each other.

When a model page ranks the model against others, it counts only models with at least 500 MOT tests at that age. It shows no rank when fewer than three models have that many.

What this report is not

The report works from public MOT data alone. It doesn't see the car, drive it, or speak to anyone who has. Use the figures to plan your own viewing, not as a judgement on the car's condition today.

• • Not an inspection. A mechanic checking the car in person will catch things MOT data can't.
• • Not a price. Where money figures appear, they're estimates from our models, shown as a range with their source. They're not a quote.
• • Not a final answer. The record is one part of the picture, alongside the test drive, the paperwork and a mechanic's view.