What is the Gera Risk Comparison Delta (GRCD)?

The Gera Risk Comparison Delta (GRCD) is the signed difference in Gera Vehicle Risk Index (GVRI) between two UK car models: GRCD = GVRI(Car A) − GVRI(Car B). Positive GRCD means Car A has higher modelled insurance risk; negative means Car B has higher risk; zero means identical risk from DVSA MOT data. Range: −100 to +100.

What is the GVRI formula?

GVRI = 0.6 × MOT_fail_rate + 0.4 × avg_defect_severity_norm, min-max scaled 0–100. MOT_fail_rate = (failed tests) / (total class-4 normal tests) for that make+model. avg_defect_severity = mean severity of 'F' (failure) items per failed test (standard failure = 1, dangerous item = 3). Both components are first computed raw, then the weighted sum is min-max normalised to 0–100 across all qualifying pairs (N ≥ 30 tests). Source: DVSA Anonymised MOT Test Results 2016, OGL v3.0.

What data source powers the GRCD?

The GRCD is derived from the DVSA Anonymised MOT Test Results dataset, published by the Driver and Vehicle Standards Agency under the Open Government Licence v3.0. The 2016 release contained 26,016,787 class-4 (car) normal MOT tests covering 53 makes. Only initial (normal) tests are included — retests are excluded. The dataset is freely downloadable from data.gov.uk and the DfT open data portal.

Why use a minimum of 30 tests?

The N ≥ 30 threshold ensures statistical stability. With fewer than 30 tests, the fail rate estimate can vary wildly from the true population rate. All 600 qualifying combinations in the GS-I1 dataset meet this threshold. Pairs with N < 30 are labelled "insufficient data" rather than producing potentially misleading GRCD values.

Is the GRCD a premium estimate?

No. The GRCD measures the difference in modelled mechanical defect risk from real DVSA MOT data — it is a risk-context index, not an insurance premium or quote. Actual premiums depend on many additional factors: driver age, no-claims bonus, postcode, annual mileage, policy type, vehicle modifications, and insurer-specific underwriting. GeraSure's GRCD helps buyers understand the relative mechanical risk profile of two cars before they compare quotes.

Can I reproduce the GRCD independently?

Yes. Download the DVSA Anonymised MOT Test Results dataset (2016) from data.gov.uk. Filter to class-4 (car) normal tests. Group by make+model. Compute fail_rate = fails/total. For each failed test, score items: standard = 1, dangerous (D) = 3; average per failed test; average across all failed tests for that make+model. Weighted sum: 0.6 × fail_rate + 0.4 × avg_severity. Min-max normalise to 0–100 across all pairs with N ≥ 30. GRCD for any pair = GVRI(A) − GVRI(B). All steps reproducible from the public dataset.

Gera Risk Comparison Delta (GRCD) — Methodology

Full reproducible formula for the GRCD proprietary metric. Source: DVSA Anonymised MOT Test Results 2016 (OGL v3.0, 26,016,787 class-4 normal tests). All steps independently reproducible from the public dataset.

The formula

Step 1 — MOT fail rate

MOT_fail_rate = failed_tests / total_class4_normal_tests

Step 2 — Average defect severity

item_score: dangerous (D) = 3, standard = 1

avg_defect_severity = mean(item_score) per failed test, averaged across all failed tests

Step 3 — Raw risk score

raw_score = 0.6 × MOT_fail_rate + 0.4 × avg_defect_severity

Step 4 — Gera Vehicle Risk Index (GVRI)

GVRI = round(100 × (raw_score − min_raw) / (max_raw − min_raw))

min-max across all pairs with N ≥ 30 tests

Step 5 — Gera Risk Comparison Delta (GRCD)

GRCD = GVRI(Car A) − GVRI(Car B)

Positive → Car A riskier. Negative → Car B riskier. 0 → equal risk. Range: −100 to +100.

Input data (key figures)

DVSA Anonymised MOT Test Results 2016 — key input figures (OGL v3.0)
Parameter	Value	Notes
Source dataset	DVSA Anonymised MOT Test Results	2016 release
Total records used	26,016,787	Class-4 (car) normal initial tests only
Makes covered	53	UK passenger car makes in dataset
Qualifying make+model pairs	600	N ≥ 30 test threshold applied
Minimum N threshold	30	Pairs below this → "insufficient data"
Dangerous item score	3	rfr_type_code contains "D"
Standard failure score	1	All other "F" items
GVRI scale	0 – 100	0 = lowest risk, 100 = highest risk
GRCD range	−100 to +100	Signed difference of GVRI scores
Licence	Open Government Licence v3.0	Freely reproducible

Worked example

The highest GRCD in the 2016 dataset: Vauxhall Vectra vs Toyota Aygo.

GVRI(Vauxhall Vectra) = 60

GVRI(Toyota Aygo) = 23

GRCD = 60 − 23 = +37 pts

Vauxhall Vectra fail rate: 42.5% (211,347 tests, High Risk). Toyota Aygo fail rate: 15.3% (111,853 tests, Low Risk). All from DVSA MOT 2016, OGL v3.0.

Exclusions and limitations

Not a premium estimate. The GRCD reflects mechanical defect risk from MOT data — not insurance pricing. Premiums also depend on driver profile, postcode, no-claims bonus, annual mileage, and insurer underwriting.
No premium data. NAIC/state-DOI and UK insurer premium tables are not available in machine-readable key-free form. Gera publishes risk-context indices only; no specific insurer product is quoted or implied.
2016 data. The publicly available DVSA Anonymised MOT dataset dates from 2016. Newer vehicles or model updates may have different risk profiles. Gera re-dates this index when newer public datasets are released.
Class-4 (car) only. Commercial vehicles, motorcycles and other classes are excluded. Only class-4 normal initial tests are used.
N ≥ 30 threshold. Make+model pairs with fewer than 30 tests are excluded to avoid unreliable fail-rate estimates.

Data sources

CMP-INS data sources
Source	Publisher	Licence	Access
DVSA Anonymised MOT Test Results 2016	Driver and Vehicle Standards Agency / DfT	Open Government Licence v3.0	Free, no API key
GS-I1 GVRI cells (own computed index)	Gera Systems (from DVSA source above)	Derived from OGL v3.0 source	This site — methodology public

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).

Gera Risk Comparison Delta: methodology FAQs

What is the Gera Risk Comparison Delta (GRCD)?: The Gera Risk Comparison Delta (GRCD) is the signed difference in Gera Vehicle Risk Index (GVRI) between two UK car models: GRCD = GVRI(Car A) − GVRI(Car B). Positive GRCD means Car A has higher modelled insurance risk; negative means Car B has higher risk; zero means identical risk from DVSA MOT data. Range: −100 to +100.
What is the GVRI formula?: GVRI = 0.6 × MOT_fail_rate + 0.4 × avg_defect_severity_norm, min-max scaled 0–100. MOT_fail_rate = (failed tests) / (total class-4 normal tests) for that make+model. avg_defect_severity = mean severity of 'F' (failure) items per failed test (standard failure = 1, dangerous item = 3). Both components are first computed raw, then the weighted sum is min-max normalised to 0–100 across all qualifying pairs (N ≥ 30 tests). Source: DVSA Anonymised MOT Test Results 2016, OGL v3.0.
What data source powers the GRCD?: The GRCD is derived from the DVSA Anonymised MOT Test Results dataset, published by the Driver and Vehicle Standards Agency under the Open Government Licence v3.0. The 2016 release contained 26,016,787 class-4 (car) normal MOT tests covering 53 makes. Only initial (normal) tests are included — retests are excluded. The dataset is freely downloadable from data.gov.uk and the DfT open data portal.
Why use a minimum of 30 tests?: The N ≥ 30 threshold ensures statistical stability. With fewer than 30 tests, the fail rate estimate can vary wildly from the true population rate. All 600 qualifying combinations in the GS-I1 dataset meet this threshold. Pairs with N < 30 are labelled "insufficient data" rather than producing potentially misleading GRCD values.
Is the GRCD a premium estimate?: No. The GRCD measures the difference in modelled mechanical defect risk from real DVSA MOT data — it is a risk-context index, not an insurance premium or quote. Actual premiums depend on many additional factors: driver age, no-claims bonus, postcode, annual mileage, policy type, vehicle modifications, and insurer-specific underwriting. GeraSure's GRCD helps buyers understand the relative mechanical risk profile of two cars before they compare quotes.
Can I reproduce the GRCD independently?: Yes. Download the DVSA Anonymised MOT Test Results dataset (2016) from data.gov.uk. Filter to class-4 (car) normal tests. Group by make+model. Compute fail_rate = fails/total. For each failed test, score items: standard = 1, dangerous (D) = 3; average per failed test; average across all failed tests for that make+model. Weighted sum: 0.6 × fail_rate + 0.4 × avg_severity. Min-max normalise to 0–100 across all pairs with N ≥ 30. GRCD for any pair = GVRI(A) − GVRI(B). All steps reproducible from the public dataset.