How Bike Reliability works

Bike Reliability is a score from 0 to 100 that shows how well a motorcycle holds up compared with other bikes of the same age and mileage. It comes from more than 18 million official roadworthiness tests on real bikes.

Where the data comes from

In Great Britain, every motorcycle over three years old must pass a roadworthiness test, called the MOT, every year. The government publishes the result of every test since 2005, with the make, model, engine size, mileage and first registration date of each bike. That is roughly a million motorcycle tests a year.

No other country publishes anything like it. In the US there is no national inspection, and the few states that inspect bikes either do not publish results or record almost no failures. Germany and Japan inspect bikes but do not release the results.

Why it matters wherever you ride

Most motorcycles are built the same way for every market. A Honda CBR600RR sold in Ohio, Sydney or Leeds comes from the same factory with the same brakes, lights, wheels and suspension. The test checks exactly those parts, so how a model fares in Britain is a good guide to how it will wear anywhere.

Large samples. Popular models have thousands of tests, so one owner's bad luck does not move the score.
Real bikes, not new ones. Every tested bike is at least three years old, many over twenty, so the score shows how a model ages.
Independent. The tests are carried out by thousands of separate garages, not by manufacturers, magazines or owners.

Some things are different in Britain. Its wet weather and salted winter roads are hard on bikes, so corrosion may show up more than in a dry climate. A few models were only sold in Britain or Europe in a slightly different form. Because every bike is judged against others tested in the same conditions, these effects mostly cancel out in the score.

What the test checks, and what it does not

The test checks brakes, lights, steering, suspension, tyres, wheels, frame, exhaust and noise, and the registration plate and frame number. It does not take the engine or gearbox apart, so a bike can pass with a worn clutch or a tired engine. The score measures how well a bike stays roadworthy, which is a large part of reliability but not all of it.

How a bike is looked after matters too. Learner 50 cc and 125 cc bikes, for example, tend to score lower partly because of how they are used. Compare a bike with others of the same type and engine size, as the rankings and each bike's rivals table do.

How the score is worked out

Passed first time. We count a bike as passing only if it passed as presented. A bike repaired during the test counts as a fail.
Expected pass rate. Older and higher-mileage bikes fail more often. For every test we look up the pass rate of all bikes of that age and mileage, then average those for the model. This is the rate an average bike would have scored on the same tests.
Points above or below expected. We subtract the expected rate from the actual rate.
Score out of 100. We rank every model with at least 100 tests by that difference. The score is the share of models it beats: 70 means it does better than 70% of models.

Model years

The test records give the date a bike was first registered, not its model year. Bikes are sometimes registered a year after they were built, so a few "2010" bikes may be 2009 models. A year page shows its own score when that year has at least 100 tests. Otherwise it shows the score for all years of the model, and says so.

Matching tests to our bikes

Test records name bikes in many different ways, such as "CBR600RR", "CBR 600 RR" or just "CBR". We match each one to our models using the make, model name, engine size and year. Where the records do not say which version a bike is, for example a standard or a sport version, the score covers those versions together and the page says so. Records we cannot match with confidence are left out rather than guessed.

Safety recalls

Recalls come from the US National Highway Traffic Safety Administration and are matched by exact model year. Recalls usually apply in every country where a bike was sold, but check with your dealer, because not every market is always included.

Updates

New test data is published once a year, usually between March and June. We update every score when it arrives. The current scores use tests from 2005 to the end of 2025.