Volkswagen Golf — MOT statistics 2024

The Volkswagen Golf failed 24.94% of 1,323,641 first MOT attempts in 2024 ("first-time", excluding retests). Adjusted for the age of the fleet, its initial-failure ratio is 85.1 (100 = expected for its age profile). This car fails its MOT about 15% less often than expected for its age.

Age-adjusted peer: Honda Jazz (index 85.0) — the two published results are 0.1 index points apart.

Volkswagen Golf common problems: what actually fails the MOT

Body, chassis, structure stand out at 14 per 100 tests, against 8.4 per 100 tests nationally.

higher bar = fails more oftenrate per 100 tests

family vs national — top component groups, per 100 tests. Legend: This family · National avg.

Named defects ranked by how much more often they appear than the national average — not by raw rate.

Volkswagen Golf MOT advisories: what gets flagged

Suspension is highest at 80.5 per 100 tests; Body & structure are lowest at 8.6 per 100 tests — the gap is 71.9 per 100 tests.

higher bar = flagged more oftenadvisory items per 100 initial tests

Advisories are wear the tester noted but didn't fail — what will need money soon.

How does the Volkswagen Golf perform as it gets older?

At age 10 stands out at 18.4%, against 27.7% for all cars of the same age.

higher = larger sharecompare shares in percentage points

Chart axes: car age, years · failure-rate scale 0–47%.

Takeaway: a 10-year-old Golf has roughly a 18% chance of failing its MOT; from about age 18 it stays above the national age curve.

Volkswagen Golf petrol vs diesel: which ages best

Diesel is highest at 28.2%; Electric is lowest at 11.6% — the gap is 16.6 percentage points.

higher = larger sharecompare shares in percentage points

Chart sample sizes: n=657k · n=649k · n=16k · n=1,274.

How dangerous are Volkswagen Golf MOT failures?

Minor defects stand out at 30.8 per 100 tests, against 24.4 per 100 tests nationally.

higher bar = fails more oftenrate per 100 tests

Chart values: Minor 30.8 · Major 64.8 · Dangerous 13.4.

Per 100 tests on this family. National average: Minor 24.4 · Major 63.3 · Dangerous 12.9 per 100.

Does the month of a Volkswagen Golf MOT matter?

Nov is highest at 26.4%; Mar is lowest at 23.6% — the gap is 2.8 percentage points.

higher = larger sharecompare shares in percentage points

Chart scale: 22%–29%.

bars: Volkswagen Golf · dashed: all class-4 cars

Takeaway: the spread between best and worst month is 2.8 percentage points — booking month barely matters for this family. Seasonality partly reflects registration-date clustering, not weather.

Before reading the shape: most of it is not about cars. Britain changes its registration plates twice a year, and 19.5% of these cars were first registered in March with another17% in September — 37% in two months against the 17% an even spread would give. An MOT falls due three years after first registration, so the buying calendar becomes the testing calendar. Peaks here mark when cars are tested, not when they are worse.

Volkswagen Golf average mileage by age

At age 8, the middle half of Golfs runs from 48,984 miles to 82,837 miles, against 65,087 miles as the median.

further right = more milesmedian is the typical car; band is p25–p75

Chart axes: car age, years · odometer scale 0–170k mi.

Does high mileage mean more Volkswagen Golf MOT failures?

Same age band (6–14 years), split by odometer — this separates wear from age: 9% of the lowest-mileage cars failed versus 34.2% of the highest.

line: median · band: middle half of the fleet (p25–p75)

A typical 8-year-old Golf has ~65,087 miles on the clock. Much more — ask why; much less — check for gaps in the MOT history.

Where in Britain the Volkswagen Golf fails its MOT most

Exeter area is highest at 35.9%; Durham area is lowest at 17.4% — the gap is 18.5 percentage points.

higher = larger sharecompare shares in percentage points

Toughest areas	% failing	vs model overall
Exeter (EX)	35.89%	+10.9 pp
Torquay (TQ)	35.88%	+10.9 pp
Truro (TR)	35.67%	+10.7 pp
Brighton (BN)	33.76%	+8.8 pp
Bournemouth (BH)	33.21%	+8.3 pp

Easiest areas	% failing	vs model overall
Durham (DH)	17.38%	-7.6 pp
Uxbridge (UB)	17.8%	-7.1 pp
Twickenham (TW)	19.51%	-5.4 pp
Romford (RM)	19.58%	-5.4 pp
Sheffield (S)	19.61%	-5.3 pp

Why do areas differ?

For each external factor we correlate the area fail rate with the factor across postcode areas that have ≥800 tests of this family, then compare the bottom vs top fifth (quintile) of the factor. Correlation, not causation — factors travel together (the north is colder and often poorer; urban areas are denser and less salted). Road surface condition is joined now, for the 91 English areas the DfT publishes, and it explains nothing: the correlation with suspension failures is −0.01 and with body and structure failures 0.07. Potholes wreck wheels and tyres between tests, but by the time a car reaches the bay the road it drives on leaves no measurable trace.

In areas with the highest urban share this model fails 4.7 pp less often than in the lowest fifth (r=-0.35, n=116 areas with ≥800 tests).

In the most coastal fifth of areas this model fails 3.4 pp more often than in the most inland fifth (r=-0.28, n=116 areas with ≥800 tests).

In areas with the highest income deprivation this model fails 3.1 pp less often than in the lowest fifth (r=-0.28, n=97 areas with ≥800 tests).

Values shown in the quintile chart

Factor	Low Q	High Q	Gap	Correlation
Urban share	26.8%	22.1%	-4.7 pp	r=-0.35
Distance to coast	26.9%	23.5%	-3.4 pp	r=-0.28
Income deprivation	25.8%	22.6%	-3.2 pp	r=-0.28

Factors that only mean something combined

Some conditions are useless on their own and informative multiplied together. Frost days alone say little — Britain barely freezes. Salt tonnage alone says little — it tracks frost. Distance to the sea alone says little. Multiply the first two and divide by the third and you get an index of how hard an area attacks a car's underside, and it lines up with the corrosion defects we actually record. We built five such combinations and tested each against our own data. Three survived; two did not, and the failures are listed here on purpose — a proxy that was never checked is decoration. One survivor only survived on the second attempt: built from a crude stand-in for street parking it failed outright, and rebuilt from the census — terraced streets carrying more cars than they have driveways — it works, and carries information the salt index does not.

Composite area indicators, each tested against a defect rate we measure

Composite	Built from	Tested against	r	Verdict
Underbody corrosion	frost days × salt load ÷ (distance to coast + 1)	share of tests with a corrosion defect	+0.42	holds (116 areas)
Tyre wear	rain days × annual miles ÷ age at first tyre advisory	share of tests failing on tyres	+0.45	holds (116 areas)
Overnight corrosion	terraced housing × cars per household × frost days	share with a corrosion defect	+0.41	holds (102 areas) — and adds signal beyond the first index
Deferred maintenance	deprivation × advisory-to-failure conversion ÷ labour rate	cars repeating the same advisory	−0.26	fails (118 areas)
Powertrain stress	engine capacity ÷ kerb weight	suspension, brake and emissions failures	−0.16	fails (22 models)

The deprivation result is the one worth dwelling on. The intuition that poorer areas defer repairs is strong enough that most people treat it as established. We tested it twice — once on raw failure rates (r = 0.04) and again, more fairly, on whether an advisory turns into a failure — and it did not hold any of the three times we tried it — on raw failure rates, on advisory-to-failure conversion, and again with deprivation measured properly inside each nation rather than England alone. We are closing it. What does predict local failure rates is the density of test centres, which is a statement about testing, not about cars.

All external-factor correlations (machine-readable)

Factor	r	n areas	Low-Q fail %	High-Q fail %	Gap pp	Low-Q factor	High-Q factor
Urban share (0–1)	-0.349	116	26.8	22.1	-4.7	0.4	1
Distance to coast (km)	-0.280	116	26.9	23.5	-3.4	6	78.7
Income deprivation (score)	-0.278	97	25.8	22.6	-3.1	0.1	0.2
MOT centre density (per 10k vehicles)	0.261	116	22.0	26.1	+4.1	5	8.9
Terrain gradient (% grade)	0.212	69	24.2	26.2	+2.0	0.5	4.2
Road-salt proxy (g/m²-yr)	-0.173	116	26.0	23.9	-2.1	223	674.9
Frost days (days/year)	-0.171	116	26.7	23.8	-2.9	11.8	38

Deprivation uses England IMD 2019 income score only (Scotland SIMD is a different scale). Salt proxy is a simplified NWSRG-style dry-salt sum from air Tmin; coastalness is distance from the area centroid to the nearest coastline.

Every dot is a postcode area — size = how many were tested there, colour = share failing. These figures are not age-adjusted: coastal and rural areas often run older fleets and salted roads, so higher raw fail rates do not automatically mean a worse local stock of this model.

Map labels: Scotland · N. England · Wales · London · S. West. Colour scale: 20%–33% failing.

Volkswagen Golf survival: which cars are still on the road

By age 19 stands out at 48.1%, against 100% at the ages 4–6 reference level.

higher = larger sharecompare shares in percentage points

Chart axes: car age, years · survival vs ages 4–6 · scale 0–105%.

By age 19, only about 48% as many cars of this family present for MOT as at ages 4–6 (reference n≈65,928); survivors still fail at 43.7% vs 7.7% at age 4. Older cars that still appear are a selected subset — weaker ones have already left. DfT licensed stock is 88% of its 2017 Q4 peak (1,020,501 → 900,884).

How many miles does a Volkswagen Golf last?

100,000+ miles is highest at 38.6%; 200,000+ miles is lowest at 2.0% — the gap is 36.6 percentage points.

higher = larger sharecompare shares in percentage points

Of 1,321,431 tests of this family with a usable odometer in 2024, 38.6% had already reached 100k miles, 11.5% reached 150k, and 2.0% reached 200k — a cross-section of cars still presenting for MOT, not a full birth cohort.

How many years is a Volkswagen Golf likely to last?

Bought at age 5 is highest at 13.5 years; Bought at age 15 is lowest at 3.5 years — the gap is 10 years.

bigger number = longer remaining lifeyears on the observed survival curve

estimateEstimate only — not a warranty: a typical 10-year-old example of this family (~83,449 miles median) has about 8.5 expected years of further MOT-present life on the 2024 survival curve. Derived from how many cars of each age still appear for test, not from a longitudinal panel.

Volkswagen Golf faults that come together

Emissions & leaks + Seatbelts & SRS stand out at 10.47×, against 1× if the faults were independent.

above 1.0× = more than expectedbelow 1.0× = less

Defects in «Noise, emissions and leaks» and «Seat belts and supplementary restraint systems» appear together 10.47× more often than chance (2,037 co-occurrences in the same test; lift≥1.3, n≥300).

Lift = how many times more often both groups appear on the same test than if independent. Thresholds: lift ≥ 1.3, co-occurrences ≥ 300.

Is low mileage better for a Volkswagen Golf?

Cars driven under 5,000 miles a year stands out at 11.3%, against 20.7% for cars driven 8,000–12,000 miles a year.

higher = larger sharecompare shares in percentage points

At matched ages 5–14, cars averaging under 5k miles/year fail at 11.3% vs 20.7% for 8–12k miles/year — 9.4 pp lower (n_low=133,626, n_normal=350,428). In this family low annual use is associated with fewer MOT fails, not more — the ‘garage queen’ penalty is not visible in overall fail rates. Diesel alone: 14.6% low-use vs 23.7% normal (-9.1 pp).

Volkswagen Golf real MPG against official fuel economy

petrol/electric (2022) real-world fuel use stands out at +310.5%, against 0% for an exact match to the official figure.

higher = larger sharecompare shares in percentage points

Since 2021 every new car in Europe records its own fuel use, and the regulator publishes what those meters say next to the laboratory figure the car was sold on. For this manufacturer the worst gap is 16.73% — diesel cars of 2022 burning 6.41 litres per 100 km against an official 5.49, with CO₂ higher by the same proportion.

This is the manufacturer, not the model. The regulator aggregates by legal manufacturer (VOLKSWAGEN), so these figures are context for the badge rather than a measurement of this particular car — a point we would rather state plainly than let a reader assume otherwise. The gap itself is not a scandal, it is the known distance between a standardised laboratory cycle and real roads, but it is the distance you actually pay for at the pump. The plug-in hybrid rows are not a data error and deserve a sentence of their own: an official figure of 1.4 litres assumes the car spends most of its life on the battery, and the meters say these ones burned 5.48. Owners are not plugging them in, and a plug-in hybrid that is never charged is simply a heavy petrol car. Source: European Environment Agency (EEA), Real-world CO2 emissions from new cars and vans, reporting year 2024, v3, 2023_Cars_Aggregated.csv, CC BY 4.0 — copyright holder: European Environment Agency (EEA).

Which Volkswagen Golf engine is best?

petrol under 1.0 is highest at 124.3; petrol over 2.0 is lowest at 64.7 — the gap is 59.6 points.

below 100 is betterabove 100 is worse100 = normal for this model at that age

Splitting this model by fuel alone hides the thing that actually differs. Within the same badge, the petrol under 1.0 runs at 124.3 against this model's own age norm while the petrol over 2.0 runs at 64.7 — a spread of 59.6 points between two cars that share a name.

Read this one with more care than the rest of the page. Age is held constant, but who buys which engine is not: the largest capacity in a range is often a performance or executive variant that is garaged, cherished and lightly driven, while the smallest is frequently a fleet or entry car worked hard from new. The gap below is real in the test records; how much of it is the engine and how much is the owner, this data cannot separate. Bands under 5,000 tests are not shown.

Best year for a Volkswagen Golf, and the years to avoid

Build year 2004 is highest at 110.8; Build year 2020 is lowest at 66.3 — the gap is 44.5 points.

below 100 is betterabove 100 is worse100 = normal for this model at that age

build year vs the same model at the same age · 100 = as expected

The best build year here is 2020 and the weakest is 2004. Comparing build years sounds simple and is not: within a single year of testing, a car's build year and its age are the same number, so a naive comparison of years is a comparison of ages wearing a disguise. Each year below is measured against the same model at the same age in other cohorts, across six years of testing — 19 of 23 years come out distinguishable from their own model's age norm.

Two limits worth knowing before you act on this. Age, test year and build year are locked together by arithmetic, so the test-year effect is taken from the observed national failure level of each year rather than estimated — it is an adjustment, not a solution. And for cars built before about 2005, a good score says more about which examples survived to be tested than about how that year was built: survivors are self-selected for care.

Volkswagen Golf MOT repair costs

The MOT-failure repair budget per test runs from £104.55 to £198.64, against £54.85 for the test itself.

bigger number = costs morepounds per test, MOT-failing faults only

Multiply how often each component fails on this model by what that repair usually costs, and you get the amount an owner should expect to budget per test for MOT-failing faults only — roughly £105 to £199, on top of the test fee itself. It is a budgeting figure, not a quote: the spread between a set of pads and a set of discs is most of the range.

Prices are modal quotes from a hand-kept catalogue of published UK garage guides — we do not scrape the booking sites, their terms forbid it, so this list is short and updated by hand with a source and date against each line. Each line prices one typical repair in that group, not a full remedy, and only groups our catalogue actually covers are counted. The contrast worth noticing is the second card: the legal maximum for the test has been £54.85 since 2010, while the repairs it uncovers have followed inflation the whole time.

For scale from a different direction: Warrantywise, which sees used-car warranty claims rather than test failures, puts the average claim at £2,052 across the ten highest-claim models it ranks and £539 for Honda, the lowest-claim make in its index (Warrantywise index 2026, 1.6 million claims, cars aged 3–15). Those sit an order of magnitude above the figures in this table, and the gap is the point: a warranty claim is a gearbox or an engine, which almost never fails an MOT, while the faults that do fail one are corrosion, lamps and tyres, which no warranty pays for. The two numbers answer different questions and neither replaces the other.

Volkswagen Golf running costs for a fleet

Vehicles failing at age 10 stand out at 18%, against 10% at age 5 in the same 100-vehicle fleet.

higher = larger sharecompare shares in percentage points

The number a fleet manager needs is not the failure rate, it is the year the car starts costing more. Between age 11 and 12 the failure rate of this model jumps by 7.94 percentage points — the steepest single step in its life, and the natural planning point for replacement.

Repair spend uses the same modal prices as the section above, so it inherits the same limits: one typical repair per component group, not a full remedy. Downtime is the gap between the failed test and the passed re-test, which is the window a vehicle is legally off the road.

Volkswagen Golf rust problems: where they corrode

Age-standardised corrosion susceptibility stands out at 0.83×, against 1× for the national age-standardised rate.

above 1.0× = rusts more than averagebelow 1.0× = rusts less

This Volkswagen Golf fails on corrosion at about the average rate once age is held constant, and its owners live in areas that are about typical for the models we publish.

Corrosion is the one failure mode that is as much about geography as about the car. We keep the two apart on purpose: a model can look rusty simply because it is popular in Scotland, and a model can look clean simply because it is bought in Surrey. Blending them into a single score would hide which of the two is doing the work.

Exposure is this model's tests weighted by an area corrosion index built from frost days, salt load and distance to the coast — the composite validated at r = 0.42 against corrosion defects we actually record. It is published as a rank, not a raw score, because every mainstream model skews towards low-corrosion areas for the simple reason that most of Britain's cars are in the south. Susceptibility is standardised to the national age profile, so an old fleet does not get labelled a rusty model.

Does a Volkswagen Golf advisory become a failure?

Lamps, reflectors and electrical equipment are highest at 31.5%; Identification of the vehicle is lowest at 3.7% — the gap is 27.9 percentage points.

higher = more likelypercentage is the observed transition probability

chance the same component fails at the next MOT, given an advisory now

This is the question an advisory note actually raises, and it needs six years of test history to answer. We followed the same vehicles from one MOT to the next across 2019–2024 and measured how often a component that was merely advised came back as a genuine failure at the next test.

Read it as a budgeting horizon, not a prophecy: the most likely component above turns into a failure for about 32 in 100 cars, and the median gap is 12.1 months and roughly 5,453 miles. The rest were repaired, sold, scrapped, or simply held. Rows with fewer than 500 tracked advisories are not shown.

What will fail next on a Volkswagen Golf?

Lamps, reflectors and electrical equipment after a failure at age 15-19 are highest at 28.3%; Brakes after a passed at age <3 are lowest at 0.9% — the gap is 27.4 percentage points.

higher = more likelypercentage is the observed transition probability

The same chains, cut differently: given a car's age and how its last test went, which component most often causes trouble at the following one.

The pattern that repeats across every family: a failure at one test roughly doubles the chance of a failure on the same component at the next one, and an advisory sits about half-way between a clean pass and a failure. Age bands in years; probabilities are per car, not per defect item.

Volkswagen Golf ownership: what six years of MOTs show

The ten-day re-test pass rate is highest at 52.1%; The later re-test pass rate is lowest at 3.0% — the gap is 49.1 percentage points.

higher = larger sharecompare shares in percentage points

Failing is usually a same-week event, not a disaster. Of 1,112,885 failures followed through to their re-test, 44.99% were repaired and passed the same day and 52.05% within ten days; the median gap is 1 day. 2,887,199 defect items were cleared in the process.

Per mile driven, not per year. Dividing failures by the miles that actually went onto the odometers between consecutive tests gives 0.408 failures per 10,000 miles, on a median of 5,878 miles a year across 1,029,496 chains. This is the number to compare between models — an annual rate quietly rewards cars that barely leave the drive.

Two populations hide inside one model. 33.39% of these cars strung together three or more tests with nothing recorded at all, while 34.52% carried the same advisory three times or more — the same fault, noted, ignored, noted again. When you buy used you are buying one of those two histories, and the MOT record shows which.

Odometer anomalies: 0.955% of 1,030,710 chains show a reading lower than the previous test, with a median drop of 22,312 miles. We call this an anomaly and not fraud on purpose — instrument swaps, imports, unit mix-ups and keying errors all land in the same bucket. Obvious typos (drops under 100 miles or over 90% of the reading) are already excluded.

Volkswagen Golf technical service bulletins: what dealers are told

NHTSA manufacturer communications are 1,651.

bigger number = more recorded dealer communicationsnot affected vehicles or confirmed faults

These are the issues the manufacturer fixes quietly: service communications sent to dealers without a public recall campaign. They show what the service network has been told even when the issue never became news.

United States · NHTSA manufacturer communications

Caveat worth stating plainly: this is NHTSA data about US-specification cars. British cars sold under the same name often use different gearboxes, engines and equipment, so a bulletin may describe a component that was never fitted here. The file counts dealer communications, not affected vehicles or confirmed failures, and it cannot tell you whether an individual UK car has the issue. A communication can carry more than one subject tag, so topic counts do not have to add up to the total. Source: NHTSA Manufacturer Communications API (US-market vehicles; queried model years 2015-2022). Licence: U.S. Government work / public domain · retrieved 2026-08-03.

Volkswagen Golf owner complaints: what goes wrong

Engine (not MOT-tested) is highest at 116; Visibility is lowest at 28 — the gap is 88.

bigger number = more recorded casescompare like-for-like scope, not sample size alone

Two different populations answering two different questions: the MOT records what an inspector can measure once a year, the complaint file records what annoyed or frightened an owner on the road. A component high in one column and absent from the other is the interesting case — driver-assistance and electronics dominate complaints while almost never appearing as an MOT failure item.

US owners have filed 532 complaints about the equivalent Volkswagen Golf to the National Highway Traffic Safety Administration, including 14 that mention a crash, 3 a fire and 26 an injury. These are owner reports, not verified findings — but they are the only large public record of what drivers themselves notice, and they cover the years between MOT tests.

Caveat worth stating plainly: US-specification cars, self-selected reporters, no denominator — complaint counts scale with how many cars were sold there and how motivated owners are to file. The mechanical gap matters as much as the statistical one: American cars of the same name often carry different gearboxes, engines and equipment — automatics where Britain buys manuals — so a drivetrain complaint filed in Ohio may describe a part that was never fitted here. Read the ranking inside a model, not between models, and treat drivetrain entries with particular caution. Source: NHTSA ODI consumer complaints (US market, public domain) — US-spec vehicles, indicative only.

Volkswagen Golf inspection results in other countries

Norway PKK stands out at 29.5%, against 33.5% nationally.

higher = larger sharecompare shares in percentage points

The verdict is consistent across every regime we can measure, which is the strongest form this comparison takes: independent inspectorates, different rulebooks, same direction of travel.

Useful as a cross-check on the UK defect table above: the categories are named differently, but a component that dominates in both regimes is a genuine weak point of the model rather than an artefact of one country's testing style.

Worth dwelling on how mundane the top of that list is: tyre pressure, tread depth, headlight aim. The Dutch APK records 1.176 defects per inspection against a much higher national failure rate than Britain's, and the reason is not that Dutch cars are worse — it is that the two regimes draw the line between "note it" and "fail it" in different places. Codes are the RDW's own; the English wording here is our translation of the Dutch original.

Britain is not the only country that inspects every car every year, and the Volkswagen Golf is on the road in all of them. Each national regime writes its own rules, so the pass rates below are not comparable to each other — what does travel across a border is whether a model sits above or below its own country's average.

Rows with fewer than 500 inspections are not published. Sources: RDW open data — Meldingen Keuringsinstantie + Geconstateerde Gebreken + Gekentekende voertuigen (2024), CC0. Not an RDW endorsement.; Traficom — Periodic inspections of cars by model / by model and fault object 2024, CC BY 4.0; Statens vegvesen — Periodisk kjøretøykontroll (PKK) 2024, CC BY 4.0.

Volkswagen Golf MOT: quick answers

Volkswagen Golf MOT Risk Index stands out at 85.1, against 100 for a car with the same age profile.

below 100 is betterabove 100 is worse100 = normal for this model at that age

What is the Volkswagen Golf first-time MOT pass rate?

75.06% of first attempts passed in 2024 (1,323,641 tests). This is the first-time figure: retests after repair are excluded. Basis: first-time, excluding retests.

Why does this number differ from other sites?

Other sites often report a pass rate that includes retests or counts all attempts. Our 24.94% figure is the first-time initial failure rate, using 1,323,641 primary tests as the denominator (first-time, excluding retests). The age-adjusted MOT initial-failure ratio 85.1 is shown separately because raw rates compare fleet age as well as cars.

Compare the Volkswagen Golf with other cars

MINI Hatch is highest at 85.5; Audi A4 is lowest at 84.6 — the gap is 0.9 points.

below 100 is betterabove 100 is worse100 = normal for this model at that age

Head-to-head

Ford Focus head-to-headVauxhall Astra head-to-headAudi A3 head-to-headMercedes A Class head-to-headHonda Civic head-to-head

Same brand

All Volkswagen MOT statisticsVolkswagen Tiguan MOT statsVolkswagen Passat MOT stats

Similar risk

Honda Jazz (index 85.0)MINI Hatch (index 85.5)Audi A4 (index 84.6)

Same size class

how Ford Focus compareshow Vauxhall Corsa compareshow MINI Hatch compares