What is the Gera Vehicle Risk Index (GVRI)?

The GVRI (0–100) is a Gera Systems composite index measuring the mechanical reliability and defect risk profile of a UK car make+model combination. It combines two components from real DVSA data: MOT fail rate (60% weight) and normalised average defect severity (40% weight). Higher = higher risk. The index is computed transparently and independently reproducible from the publicly available DVSA dataset.

Why weight MOT fail rate at 60% and severity at 40%?

Frequency (fail rate) is the dominant driver of mechanical risk to an insurer: a car that fails MOT tests more often is more likely to experience mechanical problems mid-policy. Severity adds incremental signal — a model prone to dangerous defects (brakes, steering) carries more safety risk than one failing on advisory-level items. The 60/40 split reflects this intuition while remaining clearly communicable. The weights are documented and fixed; any reader can substitute alternative weights using the same published data.

Why is the "insufficient data" threshold set at N < 30?

A sample of fewer than 30 tests is insufficient to produce a statistically stable fail rate. With N = 10 tests, a single additional failure would shift the rate by 10 percentage points — well above the noise level. N = 30 is a standard minimum for proportion estimation and is commonly used in reliability and quality-control contexts. In practice, the top-600 by test volume (used for the GeraSure leaf pages) all have many thousands of tests.

Is the GVRI a premium quote or a guarantee of reliability?

No. The GVRI is an objective, data-driven risk indicator based on historical MOT outcomes across the UK fleet — it is not a premium quote, a prediction for any individual vehicle, or a guarantee of future reliability. A high GVRI means the make+model has historically shown a higher fail rate and/or more severe defects across many vehicles in the DVSA dataset. Individual vehicle condition, age, mileage, and maintenance history also matter.

Can I reproduce the GVRI myself?

Yes. The DVSA Anonymised MOT Test Results are published at data.gov.uk under OGL v3.0 (no key required). The four-step process above is the complete calculation. The 2016 file is available at the source URL below. Gera invents no numbers — every GVRI value can be independently verified from the published CSV.

Gera Vehicle Risk Index — Methodology

The complete, reproducible formula behind the GVRI, computed only from real DVSA Anonymised MOT Test Results (2016, OGL v3.0, 26,016,787 tests).

The formula

raw_gvri = 0.6 × MOT_fail_rate + 0.4 × avg_defect_severity_norm

GVRI = round( (raw_gvri − min) ÷ (max − min) × 100 )

where min and max are taken across all 600 qualifying UK make+model combinations (N ≥ 30 class-4 normal tests, 2016 DVSA data). The highest-risk combination scores 100; the lowest scores 0.

Step by step

Download and filter the DVSA source data. Obtain the DVSA Anonymised MOT Test Results file for the target year from data.gov.uk (dataset ID e3939ef8-30c7-4ca8-9c7c-ad9475cc9b2f, OGL v3.0, no API key required). Filter to class_id = 4 (cars) and test_type = 'NT' (normal test, excluding retests). The 2016 file contained 26,016,787 qualifying rows.
Compute per make+model MOT fail rate. Group rows by (make, model). For each group, MOT_fail_rate = count(test_result = 'F') ÷ count(all rows). PRS (Pass with Rectification at Station) rows count as passes. Exclude any make+model group with fewer than 30 tests ("insufficient data" threshold) to prevent statistical noise from micro-volume combinations from inflating the index.
Compute per make+model average defect severity. Join to the test_item file (same dataset, also OGL v3.0). For each failed test (test_result = 'F'), collect all 'F' (failure) item rows. Score each item: dangerous_mark = 'D' → severity 3; all other failure items → severity 1. Compute the mean severity per test (sum ÷ count of 'F' items). Then average per-test means across all failed tests for each make+model to get avg_defect_severity. Min-max normalise avg_defect_severity to [0, 1] across all qualifying pairs → avg_defect_severity_norm.
Compute raw GVRI and normalise 0–100. For each qualifying make+model: raw_gvri = 0.6 × MOT_fail_rate + 0.4 × avg_defect_severity_norm. MOT_fail_rate is already in [0, 1]; avg_defect_severity_norm is already in [0, 1] after step 3. Min-max normalise raw_gvri across all qualifying pairs: GVRI = round((raw_gvri − min) ÷ (max − min) × 100). The highest-risk combination scores 100; the lowest scores 0.

GVRI risk bands

Gera Vehicle Risk Index band definitions
GVRI range	Band	Interpretation
75–100	Very High Risk	Significantly above UK fleet average in fail rate and/or severity
55–74	High Risk	Above UK fleet average on both dimensions
35–54	Moderate Risk	Around the UK fleet average
15–34	Low Risk	Below UK fleet average
0–14	Very Low Risk	Substantially below UK fleet average

UK fleet average GVRI: 37/100 across 600 make+model combinations (2016 DVSA MOT data).

Highest-risk vehicles (2016)

5 UK cars with highest Gera Vehicle Risk Index
Make	Model	GVRI / 100	Risk band	MOT fail rate	Tests
Nissan	Vanette	76	Very High Risk	53.6%	3,196
Fiat	Multipla	73	High Risk	51.3%	6,905
Daewoo	Matiz	72	High Risk	50.9%	15,252
Ford	Puma	69	High Risk	48.7%	15,307
Fiat	Stilo	68	High Risk	48.2%	15,582

Lowest-risk vehicles (2016)

5 UK cars with lowest Gera Vehicle Risk Index
Make	Model	GVRI / 100	Risk band	MOT fail rate	Tests
Honda	Jazz i-vtec es cvt	7	Very Low Risk	2.9%	3,750
Skoda	Citigo	9	Very Low Risk	5.2%	11,466
Honda	Jazz i-vtec es	9	Very Low Risk	3.8%	4,166
Ford	Fiesta zetec auto	9	Very Low Risk	6.2%	3,565
Toyota	Prius t3 vvt-i cvt	9	Very Low Risk	6.7%	3,127

What the GVRI is not

Not a premium quote. The GVRI is a risk indicator, not a price. Insurance premiums depend on many additional factors (driver age, NCB, postcode, mileage, modifications, policy type).
Not a verdict on any individual vehicle. The GVRI is a fleet-level average across all registered examples of a make+model. A well-maintained individual car of a high-GVRI model may outperform a neglected low-GVRI model.
Not current-year data. The 2016 figures reflect the UK fleet tested in that year. Model refreshes, recalls, and age-distribution shifts mean newer data may differ. Gera updates the GVRI when a new DVSA annual release is available.
Not a DVSA publication. The GVRI is a Gera Systems index computed from DVSA open data. DVSA publishes the raw test results; the GVRI computation and this index are Gera's own.

Gera Vehicle Risk Index: frequently asked questions

What is the Gera Vehicle Risk Index (GVRI)?: The GVRI (0–100) is a Gera Systems composite index measuring the mechanical reliability and defect risk profile of a UK car make+model combination. It combines two components from real DVSA data: MOT fail rate (60% weight) and normalised average defect severity (40% weight). Higher = higher risk. The index is computed transparently and independently reproducible from the publicly available DVSA dataset.
Why weight MOT fail rate at 60% and severity at 40%?: Frequency (fail rate) is the dominant driver of mechanical risk to an insurer: a car that fails MOT tests more often is more likely to experience mechanical problems mid-policy. Severity adds incremental signal — a model prone to dangerous defects (brakes, steering) carries more safety risk than one failing on advisory-level items. The 60/40 split reflects this intuition while remaining clearly communicable. The weights are documented and fixed; any reader can substitute alternative weights using the same published data.
Why is the "insufficient data" threshold set at N < 30?: A sample of fewer than 30 tests is insufficient to produce a statistically stable fail rate. With N = 10 tests, a single additional failure would shift the rate by 10 percentage points — well above the noise level. N = 30 is a standard minimum for proportion estimation and is commonly used in reliability and quality-control contexts. In practice, the top-600 by test volume (used for the GeraSure leaf pages) all have many thousands of tests.
Is the GVRI a premium quote or a guarantee of reliability?: No. The GVRI is an objective, data-driven risk indicator based on historical MOT outcomes across the UK fleet — it is not a premium quote, a prediction for any individual vehicle, or a guarantee of future reliability. A high GVRI means the make+model has historically shown a higher fail rate and/or more severe defects across many vehicles in the DVSA dataset. Individual vehicle condition, age, mileage, and maintenance history also matter.
Can I reproduce the GVRI myself?: Yes. The DVSA Anonymised MOT Test Results are published at data.gov.uk under OGL v3.0 (no key required). The four-step process above is the complete calculation. The 2016 file is available at the source URL below. Gera invents no numbers — every GVRI value can be independently verified from the published CSV.

Explore the data

Contains public sector information published by Driver and Vehicle Standards Agency (DVSA) and licensed under the Open Government Licence v3.0. Source: DVSA Anonymised MOT Test Results — DfT (2016, published 2016).