Research Buying guide 18 August 2026 · 9 min read

Two of these five cars are hiding an accident

We ran 4,939 cars through our records while their buyers were still deciding. Two in every five came back with damage on file, the average one carries 8,040 euros of it, and most had been hit more than once.

Five used cars in a row with the MrVIN bear inspecting them. Two of the five have crash damage.

Two in five is not a scare number. It is the rate we measured on cars people were about to pay for.

These are real used cars, caught in the middle of a purchase, checked by someone who wanted to be certain before they handed over the money. What came back is worth knowing before you are the one standing in that car park.

The short version

  • 40.2% of the cars we checked had at least one accident or damage record on file.
  • The average damaged car carries 8,040 EUR of assessed damage. One in five carries more than 10,000 EUR.
  • 53.7% of damaged cars had been damaged more than once. Repeat damage is the normal case, not the rare one.
  • Among the cars we could test, the damaged ones had their odometer wound back at twice the rate of the clean ones.
40.2%of the cars checked had damage on file
8,040 EURof assessed damage on the average damaged car
53.7%of damaged cars were damaged more than once
4,939cars checked, all of them mid purchase

Skip ahead: check a car now·jump to the brand table

Two in every five had something on file

These were not random cars pulled out of a register. Every one of the 4,939 was a car somebody was about to buy, and asked us to check first.

40.2% came back with at least one recorded accident or damage claim. Two in every five. Not two in fifty.

Look at the five cars at the top of this page. The two with the crumpled panels are not the real risk, because nobody knowingly buys those. The risk is the same five cars photographed the way a seller actually photographs them, after the repairs are done and the paint is matched.

Nothing in a line up like that tells you which two they are. Not the paint, the panel gaps, the photographs or the seller. The only thing that separates them is what is written on file.

The MrVIN bear holding a report that marks two of the five cars as damaged

Make sure you are not buying one of the two

See a car's real mileage, damage, finance and country history across 30+ countries, in under a minute.

Example: WBA3A5G59DNP26082Free preview included

  • 30+ countries
  • Result in under a minute
  • Full PDF report

It is not scratches

The natural assumption is that a damage record means something small. A car park scuff that someone claimed for, had painted, and forgot about.

It is not. The average damaged car in this data carries 8,040 EUR of assessed damage. That is the repair or settlement value an assessor put on the vehicle, not a number we estimated.

The spread matters more than the average, because the spread is where the risk actually sits.

Total recorded damage on the carShare of damaged cars
Under 500 EUR8.8%
500 to 1,000 EUR11.0%
1,000 to 2,500 EUR23.6%
2,500 to 5,000 EUR19.4%
5,000 to 10,000 EUR16.8%
10,000 to 20,000 EUR11.7%
Over 20,000 EUR8.7%

Based on 1,882 damaged cars where we hold a usable amount, adding together every damage record on the vehicle.

One in five damaged cars carries more than 10,000 EUR of assessed damage. One in twelve carries more than 20,000 EUR, which on a typical used car is more than the car itself is worth.

The tail runs further than that table shows. The most expensive one percent of individual damage records sit above 36,352 EUR, and the largest single record in this data is 112,377 EUR. One claim, on one car, now somewhere on the used market.

The middle of the pack sits at 3,177 EUR, well below the 8,040 EUR average. That gap is not noise, it is the shape of the problem. A small number of near total write offs pull the average up, and those cars are back on sale looking completely ordinary. The median tells you what a damaged car usually is. The average tells you what you are trying to avoid.

And it happened more than once

The next assumption to drop is that this was one bad afternoon years ago, properly repaired and behind the car for good.

Among cars with a damage record, 53.7% had been damaged more than once. Nearly one in five had four or more separate incidents on file. The average damaged car carries 2.4 records.

The bills rise with them:

Damage records on the carMedian total damage
1 record1,527 EUR
2 records3,039 EUR
3 records5,596 EUR
4 records7,336 EUR
5 or more records10,590 EUR

At the far end, roughly one in seventy damaged cars carried more than ten separate damage records.

A car with a single repaired incident is the minority case.

Which badges carry the most

Sorted by how often a brand came back with something on file, the order is not the one most buyers would guess. Premium and sporting badges sit at the top next to value brands, while the names people buy for reliability sit lower down.

BrandShare damagedMedian total damage
Lexus55.6%6,626 EUR
Škoda53.2%3,429 EUR
BMW49.6%3,040 EUR
Mini48.0%4,281 EUR
SEAT46.9%3,642 EUR
Porsche45.8%5,776 EUR
Audi44.5%4,346 EUR
Volkswagen44.1%2,666 EUR

Every brand shown was checked at least 40 times. Sorted by the share that came back with a damage record.

Two things are worth taking from this. The frequency: on several of these badges, close to half the cars checked had something on file. And the cost: Porsche and Lexus do not only come back damaged often, they come back expensive, at roughly twice the median bill of a Volkswagen.

Read it as a picture of what our customers bring us, which is mostly used imports, rather than a verdict on how well a car is built. A badge that gets bought in volume also gets checked in volume.

Where the car was hit, and what that tells you

The front of the car takes the largest share of impacts. It is also one of the cheaper places to be hit.

FRONT 1,958 EUR REAR 1,535 EUR CABIN AND FLOOR 8,062 to 12,006 EUR SIDES, CORNERS AND ROOF 2,141 to 4,107 EUR
Damage cost by where the car was hit. The cheap areas are the ones you can see. The expensive ones sit in the middle, where a photograph cannot reach.
  • Serious: structural or mechanical
  • Moderate
  • Usually cosmetic
Area of the carShare of recordsMedian cost
Front44.7%1,958 EUR
Rear30.9%1,535 EUR
Front right27.8%3,139 EUR
Front left26.3%3,082 EUR
Rear right21.7%2,331 EUR
Right side21.0%2,400 EUR
Rear left20.6%2,141 EUR
Left side19.5%2,447 EUR
Windscreen14.4%895 EUR
Interior6.0%12,006 EUR
Underbody5.4%9,844 EUR
Mechanical parts4.9%8,062 EUR
Roof4.1%4,107 EUR

One record can list several areas, so the shares add up to more than 100 percent.

Look at the bottom of that table, not the top. Interior, underbody and mechanical damage are uncommon, but they cost four to six times what a knock to the front does.

That is not because a seat costs more than a bumper. Those areas only get listed when the crash was bad enough to reach the cabin, the floor or the drivetrain, and such a claim names several areas at once, so the figure shown is the whole claim. A windscreen sits at the bottom for the opposite reason: it is usually the only thing on the claim.

Takeaway A history that mentions the interior or the underbody means the car was in a serious accident, whatever the panels look like now.

And where it happened

Most of these cars are imports, so the country a claim was filed in is usually the country the car lived in before it was sold on. That makes this a rough map of where a used import spent its working life.

A car crossing a border, damaged on the side it came from and clean on the side it is sold in
Nothing is repaired at the border. The record stays in the country the car came from, and the next buyer is shown only the clean half.
Country the damage was recorded inShare of recordsMedian cost
Germany32.8%1,322 EUR
Poland17.3%1,316 EUR
Austria13.2%1,685 EUR
Romania8.0%790 EUR
Belgium5.0%1,551 EUR
France4.4%1,151 EUR
Switzerland2.5%2,706 EUR
Bulgaria2.3%298 EUR

The eight countries with the most records. Smaller countries are left out because a handful of records cannot support a number.

Nearly two thirds of all recorded damage happened in Germany, Poland or Austria, which are also the three countries most of these cars were imported from. If you are buying a car that arrived from any of them, its history is more likely to exist than to be clean.

One caution on reading this. Our coverage is not even across Europe, so the table shows where the records are strongest as much as where accidents happen. It is a map of visibility, not an accident league table, and a country low on this list is not a safer place to buy a car from.

The odometer usually lies too

A rollback can only be spotted where we hold at least two odometer readings to compare, which is true of 2,673 of these cars. For the rest the honest answer is that we cannot tell, so they are left out completely rather than counted as clean.

Within the group we can test, the difference is stark. Cars with a damage record had been wound back 12.4% of the time. Cars with no damage record, 6.3%. Twice the rate.

The obvious objection is a fair one. Damaged cars tend to carry more odometer readings, because a claim usually records the mileage, and more readings mean more chances to catch a rollback. So is this simply an effect of seeing damaged cars more clearly?

No. Splitting both groups by how much history we hold answers that directly:

Cars with damage on file Cars with no damage on file 10% 20% 30% 40% 3.2% 1.2% 2 to 3 5.5% 2.7% 4 to 6 9.6% 2.9% 7 to 10 18.3% 10.2% 11 to 20 41.4% 30.3% 21 or more ODOMETER READINGS WE HOLD FOR THE CAR
Rollback rate within each group. Both series climb together, because the more of a car's history we can see the more rollbacks we find. The gap between them at every level is the finding.

The gap holds at every level, from the thinnest histories to the richest. The mechanism is not complicated. Someone who has decided to hide a crash has already made the hard decision, and the odometer is that same decision applied to a second number.

Put both problems on the same car and 6.5% of everything we could test carried a damage record and a wound back odometer together. About one car in fifteen was hiding two things at once.

It is not simply that the car was damaged. It is that someone hid the damage, and was willing to hide the mileage as well.

New does not mean clean

Damage builds up over time, so older cars carry more of it. That much is obvious. What is less obvious is how early it starts, and how quickly a car stops being the safe bet its age suggests.

Model yearShare with damage on fileMedian total damage
2022 and newer25.9%3,609 EUR
2018 to 202143.1%3,829 EUR
2014 to 201751.3%3,346 EUR
2010 to 201350.8%2,675 EUR
2005 to 200952.6%2,735 EUR

Cars built before 2005 are deliberately left out. Claims records thin out the further back you go, so the apparent drop among the oldest cars is our visibility, not their history.

A car from the last three or four years is still roughly a one in four proposition. By the middle of the last decade it is a coin flip, and it stays there. The steep climb happens early, between a nearly new car and one that is merely a few years old, rather than somewhere late in its life.

Now read the last column. The two newest groups are damaged least often, yet they carry the highest median repair bills of any group, around a third more than cars built between 2005 and 2009. Newer cars cost more to put right, so a nearly new car with one accident on file can easily carry more damage value than an older car with two.

The second way to look at this is not how old the car is now, but how old it was when the damage happened:

Age of the car at the time of the accidentShare of damage records
Under 3 years old33.7%
3 to 5 years old31.7%
6 to 10 years old24.9%
11 to 15 years old7.2%
Over 15 years old2.6%

Based on damage records where both the accident date and the car's first registration are known. The median damaged car was four years old when it happened.

Two thirds of all the damage in this data happened before the car turned six. A third of it happened before the car turned three, while it was still under warranty and still being described as nearly new.

That is the part worth taking to a viewing. The damage is not waiting at the far end of a car's life for somebody else to deal with. It happens early, to cars that still look and sell like young ones, and it is on file long before the paint gives anything away. A recent registration is not a clean history. It is a shorter one.

A clean report means no damage was recorded. Damage that was never claimed leaves nothing behind to find, so no check can promise a perfect car. What it can do is tell you which of the five you are looking at.

Before you sign anything

Know which one you are buying

Mileage, damage, finance and import records across 30+ countries, in one report, in under a minute.

  • 30+ countries
  • Under a minute
  • Full PDF report
Check a car now

Not sure yet? See what a full report looks like first

Frequently asked questions

What share of used cars have accident damage on file?

Among the 4,939 cars our customers asked us to check, 40.2 percent had at least one recorded accident or damage claim. That measures cars in the middle of being bought and sold, which are checked precisely because somebody wanted certainty, so it is not a figure for every car on the road.

How much damage does a damaged used car normally carry?

The median damaged car carries 3,177 euros of assessed damage and the average is 8,040 euros. The gap between the two is a small number of near total write offs pulling the average up. One in five damaged cars carries more than 10,000 euros, and one in twelve more than 20,000.

Does a clean history report mean the car was never damaged?

No. It means no damage was recorded. Damage that was never claimed, never insured or never reported leaves nothing behind to find. A clean report is a strong signal, not a guarantee, and any service that tells you otherwise is overselling.

Is a repaired car always a bad buy?

No. A repaired front bumper on an otherwise good car is a bargaining point, not a reason to walk away. What matters is where the car was hit and how much the claim was worth. A history that names the interior, the underbody or the mechanical parts points to a crash that reached the structure of the car, and those are the ones to be careful with.

Are damaged cars more likely to have a wound back odometer?

Yes, about twice as likely. Among the cars where we hold enough odometer history to test, 12.4 percent of the cars with a damage record had been wound back, against 6.3 percent of those without one. The gap holds at every level of history we hold, so it is not simply an effect of seeing damaged cars more clearly.

How can I check whether a used car has been in an accident?

Run its VIN, or its number plate where that is supported, through a history check. A report reads recorded damage and claims, mileage readings and rollbacks, finance and theft flags, and the countries the car has been registered in, across more than 30 countries. It takes under a minute, and it is the only part of a viewing that a fresh coat of paint cannot change.

Where do these numbers come from?

From our own checks. We took 4,939 vehicles that MrVIN customers submitted for a history check and looked at what the damage and claim records held against them actually said. Every figure names the group it is based on, and cars we could not assess are left out rather than counted as clean.

About this data

  • What was analysed: 4,939 vehicles submitted for a history check by MrVIN customers, run against the damage and insurance claim records we hold. These cars are self selected, meaning somebody chose to check them, which is why the figures here run higher than a whole market average would.
  • Damage figures: 1,985 cars with at least one damage record. Cost figures use the assessed repair or settlement value recorded against each incident, across 4,418 individual records with a usable amount on 1,882 cars. Records that could not be placed in euros are left out of every money figure.
  • Rollback figures: restricted to the 2,673 cars where we hold at least two odometer readings, since a rollback can only be found by comparing readings. The remaining cars are excluded rather than counted as clean.
  • Damaged area figures: 2,871 records that identify an area. One record can name several areas, so the shares total more than 100 percent, and the cost shown against an area is the value of the whole claim rather than that one part.
  • Brand figures: limited to brands checked at least 40 times. The mix reflects what our customers bring us, largely used imports, and is not a sample of the wider car park.
  • Averages and medians: where a distribution has a long tail we give the median as well, because a handful of write offs move an average a long way. Both are stated wherever both matter.

Read next

Two in five had damage on file