2. Comparative overview of rating structures

Euro NCAP, C-NCAP, and Bharat NCAP all publish consumer-facing star ratings, but the three programs arrive at those stars through different scoring architectures, different weighting philosophies, and different minimum-performance gates. For an engineer working on a vehicle program that targets multiple markets, the practical question is not "how many stars does each scheme award" but "which physical test results move the score, by how much, and what hard floors can veto an otherwise strong result." This section compares the three structures at that level of detail.

Star ratings: a label, not a score

In all three programs the star rating is a discrete output derived from a continuous or semi-continuous score, and the mapping is non-linear. Euro NCAP awards 1–5 stars per tested variant, with the overall rating built from four assessment areas — Adult Occupant Protection (AOP), Child Occupant Protection (COP), Vulnerable Road User protection (VRU), and Safety Assist (SA) — each expressed as a percentage of its available points. C-NCAP likewise reports a five-star scale, historically with a five-star-plus tier, and aggregates across occupant protection, VRU/pedestrian protection, and active safety. Bharat NCAP, launched on the basis of AIS-197 and aligned in spirit with the Global NCAP protocols, uses a five-star scale for Adult Occupant Protection and Child Occupant Protection, with the two scores reported and combined into an overall star result.

The key structural difference is granularity of the published output. Euro NCAP publishes four area percentages plus the star rating, so a vehicle can be diagnosed area by area. C-NCAP publishes sub-scores by test item and an overall score before the star mapping. Bharat NCAP publishes AOP and COP star ratings separately, which makes the child-protection performance visible rather than folded into a single number. When benchmarking, always compare at the area-percentage level; comparing stars alone hides large differences inside the same star band.

How the overall score is assembled

Euro NCAP's overall score is a weighted sum of the four area percentages, with weights that have shifted across protocol versions. In recent versions the weighting has been approximately 40% AOP, 20% COP, 20% VRU, and 20% SA, and the weighted result is then mapped to stars. The important consequence is that a strong AOP score cannot fully compensate for a weak Safety Assist area, because the weights cap the contribution of any single area.

C-NCAP uses a points-based aggregation: each test item (frontal offset, side impact, whiplash, pedestrian head and leg impacts, active safety tests) carries a defined maximum, raw points are summed, and the total is normalised and mapped to stars. Because the item maxima are fixed, the marginal value of improving one test is directly readable from the protocol tables — a useful property when allocating engineering effort.

Bharat NCAP computes AOP and COP scores from crash-test measurements (HIC, chest deflection, femur loads, and restraint-system checks) using the Global NCAP-derived scoring sheets, then converts each to a star rating. The overall result reflects both, so a vehicle with five-star adult protection and three-star child protection does not present as a clean five-star vehicle.

Rating architecture comparison across Euro NCAP, C-NCAP and Bharat NCAP Three parallel columns showing how test results flow into area scores, weighted aggregation, minimum performance gates, and final star ratings for each program. Euro NCAP C-NCAP Bharat NCAP Test results Area scores Aggregation Gates & stars Frontal ODB, MPDB, side barrier, pole, VRU head/leg, AEB, LSS AOP % · COP % · VRU % · SA % each normalised to its own available points Weighted sum ≈ 40 / 20 / 20 / 20 weighted % → star band (non-linear mapping) Minimum thresholds per area e.g. SA floor for 5 stars 1–5 stars, per variant Frontal 50 km/h, MPDB, side, whiplash, VRU, active safety Raw points per test item fixed item maxima from protocol tables Sum → normalised total score item-level marginal value directly readable Score thresholds per star tier 5-star-plus tier above 5-star item-level minimums apply Frontal offset, side impact, child restraint installation AOP score · COP score HIC, chest deflection, femur loads, restraint checks Separate AOP and COP stars combined into overall result both must be strong Per-area star floors weak COP caps overall 1–5 stars Common principle: a high weighted score does not guarantee the top star — minimum performance gates in each area can cap the result. Compare vehicles at the area-percentage level, not at the star level, to expose differences hidden inside the same star band.

Minimum performance requirements and the veto effect

All three programs impose floors that operate independently of the weighted total. In Euro NCAP, each assessment area must reach a minimum percentage for a given star rating; a vehicle with an excellent AOP result but a Safety Assist score below the threshold for five stars will be capped at four stars regardless of its weighted total. C-NCAP applies minimum item scores alongside the total-score thresholds, so a single weak test item can prevent a higher star tier even when the aggregate is sufficient. Bharat NCAP's structure makes the floor explicit: because AOP and COP are rated separately, a weak child-protection result directly limits the overall outcome.

The most common benchmarking error is to compare weighted totals across programs. The totals are computed on different point bases, with different test inventories and different normalisation, so a "78%" in one program is not equivalent to a "78%" in another. Compare area-level percentages and the specific test items behind them.

Worked comparison: reading a score breakdown

Consider a hypothetical mid-size SUV evaluated under all three schemes, with the following illustrative area results (these are teaching values, not measured data): AOP 88%, COP 82%, VRU 71%, SA 64%. Under a Euro NCAP-style weighting of 40/20/20/20, the weighted total is 0.40×88 + 0.20×82 + 0.20×71 + 0.20×64 = 35.2 + 16.4 + 14.2 + 12.8 = 78.6%. If the five-star gate for Safety Assist sits at 70%, the 64% SA result caps the vehicle below five stars despite the respectable weighted total. The engineering conclusion is unambiguous: the SA deficit, not the VRU result, is the binding constraint, and AEB and lane-support performance should be the priority for the next iteration.

Under a C-NCAP-style points aggregation, the same physical performance would be re-expressed against that program's item maxima. Because the item maxima are published, the team can compute the point delta required to cross the next star threshold and map it to specific test improvements — for example, the additional points obtainable from improved pedestrian head-impact performance at the specified impact points. Under Bharat NCAP, the same vehicle's COP result would be examined against the child-protection scoring sheet, with restraint installation checks and dummy readings reviewed item by item.

Practical implications for multi-market programs

Three structural facts should shape test planning. First, the test inventories differ: Euro NCAP includes a broader active-safety and VRU set, C-NCAP has its own impact configurations and speeds, and Bharat NCAP's current scope is concentrated on crash performance and child protection. A single physical test does not map one-to-one onto all three scoring sheets. Second, the aggregation logic differs: weighted percentages in Euro NCAP, fixed item points in C-NCAP, and separate area stars in Bharat NCAP. Third, the gate logic differs, and gates are where programs most often diverge in outcome for the same vehicle.

A workable approach is to build a cross-program score model early in the program: one row per physical test, one column per program, with the item maximum, the achieved or predicted score, and the gate threshold recorded for each. Populate it from protocol tables rather than from memory, and update it after every sled or full-scale test. The model then answers the questions that matter — which single test improvement buys the most stars in each market, and which gate is currently binding — without re-deriving the scoring logic each time.

When a vehicle sits one star below target in any program, first identify the binding gate rather than the lowest area percentage. The lowest percentage is not always the constraint; the gate that the vehicle fails is.

The comparison above establishes the structural vocabulary — area scores, aggregation method, gates, and star mapping — that the following sections apply test by test. The essential discipline is to keep the three layers separate: raw test measurements, area-level normalised scores, and the gated star output. Confusing them is the root of most cross-program benchmarking errors.

← Previous 📖 Contents Next →