On this page
- Define the outcome before comparing engines
- What the evidence should show
- Why raw complaint counts can mislead
- A recall does not measure every failure risk
- Do not use MOT results as an engine-health ranking
- What anecdotes are useful for
- How to assess your own vehicle
- What would justify a future ranking here?
- Sources and further reading
- Related guides
Start hereThis guide does not present a “most and least reliable” ranking. We have not established a sufficiently comparable dataset to publish one responsibly.
Define the outcome before comparing engines
“Reliable” can mean no breakdowns, no internal engine failure, no expensive repair or few warranty claims. Those measures can rank the same vehicles differently.
Decide whether the question concerns the engine alone, the complete powertrain or the whole vehicle. An electrical or emissions-system complaint should not silently become a confirmed internal engine failure in a table.
What the evidence should show
| Required context | Why it matters |
|---|---|
| Exact engine/vehicle population | Broad brand names can combine unrelated designs and generations |
| Defined failure outcome | Keeps different repair types from being counted as one event |
| Number of observed units | Provides a denominator for the reported events |
| Observation period and exposure | Age, mileage and follow-up affect opportunities for a fault to appear |
| Collection method | Warranty, workshop, survey and fleet records capture different groups |
| Missing cases and uncertainty | Prevents an apparently precise result concealing major gaps |
A neat percentage is not enough when the underlying population and method are unclear.
Why raw complaint counts can mislead
Here is an illustration of arithmetic only, not real engine data: 20 confirmed events among 1,000 observed units is 2%; 10 among 100 is 10%. The larger event count has the lower proportion in this example.
Even those proportions would not establish a fair real-world comparison unless the populations, observation periods and event definitions were comparable. Social-media post counts usually provide much less context.
A recall does not measure every failure risk
A recall concerns a defined issue and affected group. It does not provide a complete score for the engine's overall reliability, and absence of a recall does not establish absence of faults.
Use official recall checks for the individual vehicle and our recall guide for the difference between a campaign and a general engine-family claim.
Do not use MOT results as an engine-health ranking
GOV.UK states that the MOT does not cover engine, clutch or gearbox condition. Treat the test and any dataset derived from it within that scope. A general MOT pass/fail statistic is not automatically an engine-failure statistic.
What anecdotes are useful for
A well-described case can identify a possible mechanism or a question worth investigating. It can also help an owner recognise what to discuss with a technician. It cannot, by itself, tell you how frequently the issue occurs across all units or predict the result for your vehicle.
A repair business's experience can be valuable, but its customers are not a random sample of all healthy and faulty engines. Its observations need a clearly described basis before becoming a population ranking.
How to assess your own vehicle
Confirm the exact engine and history, check relevant campaigns, review actual symptoms and obtain an appropriate condition assessment. A family's reputation should not replace evidence about the particular unit.
Our lifespan guide and used-car verification checklist help organise those questions.
What would justify a future ranking here?
A future study would need a disclosed dataset, comparable groups, defined outcomes, a reproducible method, uncertainty reporting and a clear explanation of commercial connections. Until then, a careful explanation of the evidence is more useful than an unsupported league table.