Our Methodology

Data Confidence

How we measure and rate the completeness of our firearms data

What is Data Confidence?

Every firearm in the GunSpec database carries a Data Confidence score — a percentage from 0% to 100% that tells you how complete, verified, and trustworthy the specification data is.

This score isn't just a count of filled fields. It's a weighted evaluation across multiple dimensions: specification completeness, source verification, metadata richness, and cross-reference depth.

How Is It Calculated?

The confidence score is a weighted composite of five dimensions. Each dimension contributes a different percentage to the final score based on its importance to data quality.

Core Specifications

35%

The essential physical and mechanical attributes that define a firearm.

Docs:Weight, overall length, barrel length, height, width, action type, firing mechanism, magazine capacity, caliber, effective range, muzzle velocity, muzzle energy, rate of fire

A firearm with all core fields populated scores 35/35 for this dimension.

Materials & Construction

15%

Frame, slide, barrel, and stock materials along with finish and rifling details.

Docs:Frame material, slide material, barrel material, stock material, finish, rifling twist, number of grooves, barrel rifling direction

A pistol with no stock (N/A) isn't penalized — only fields relevant to the firearm type are scored.

Metadata & Context

20%

Contextual information that places the firearm in history and industry.

Docs:Manufacturer, category, country of origin, year introduced, production status, description, notes, safety mechanisms, features

Having a detailed multi-paragraph description with historical context scores higher than a single sentence.

Source Verification

20%

How many independent sources have been cross-referenced to verify the data.

Docs:dimensions.sources.fields

0 sources = 0% of this dimension

1 source = 25%

2 sources = 50%

3 sources = 75%

4+ sources = 100%

We verify against manufacturer spec sheets, military procurement documents, independent ballistic testing databases, and reputable firearms reference publications.

Relationships & Depth

10%

Connected data that enriches the firearm's profile in our database.

Docs:Caliber associations, variant linkages, military/law enforcement adoption records, game statistics

The FN SCAR-H scores well here: 1 caliber, 7 variants, 10 users, and full game stats.

35%

20%

15%

10%

Core SpecificationsSource VerificationMetadata & ContextMaterials & ConstructionRelationships & Depth

What Do the Scores Mean?

Here's how to interpret the confidence percentage you see on each firearm page.

90 – 100%Excellent

Comprehensive data with nearly all fields populated. Multiple independent sources verified. Rich context with detailed descriptions, variant linkages, adoption records, and game stats. This is reference-grade data.

HK416 (95%), Glock 19 Gen5 (93%), Glock 48 (94%)

80 – 89%Good

Strong core specifications with most fields filled. At least 2–3 verified sources. May be missing some secondary fields like trigger pull weight, sight radius, or a few materials. Very reliable for most use cases.

FN SCAR-H (87%), AK-47 (85%), M4A1 (86%)

70 – 79%Adequate

Core specifications are present but some gaps remain. Typically 1–2 sources. May lack detailed descriptions, variant data, or adoption records. Usable but with caveats.

Remington R5 (72%), QBZ-95 (78%), VHS-2 (75%)

Below 70%Incomplete

Significant data gaps. Core fields may be partially missing. Limited or no source verification. Typically newly added firearms awaiting further research, or obscure models with scarce public documentation.

These firearms are actively being researched and updated.

What Does 100% Look Like?

A firearm with 100% data confidence would need every one of these criteria met:

All 13 core specification fields populated with verified values

All applicable material and construction fields filled

Complete manufacturer data, category, country, production years, multi-paragraph description, and detailed notes

4 or more independent, authoritative sources cross-referenced

At least one caliber linked, all known variants connected, adoption records documented, and game statistics generated

Safety mechanisms, features list, and feed systems documented

In practice, few firearms reach 100%. Even well-documented models like the HK416 (95%) have minor gaps — some specs are classified, vary by production batch, or simply aren't published by the manufacturer.

Why Does This Matter?

For API Developers

Filter by confidence threshold to ensure your application only uses data that meets your quality standards. The Enterprise tier offers a dedicated endpoint to query firearms below a given confidence level.

For Game Developers

Game stats are derived from real-world specifications. Higher confidence means the underlying data powering game balance calculations is more accurate and complete.

For Researchers & Reference

Confidence scores let you quickly identify which entries are well-sourced reference material versus preliminary data that may need independent verification.

Accessing Confidence Data via API

The GunSpec API exposes confidence data at multiple levels:

GET /v1/firearms/:idEvery firearm detail response includes the data_confidence field (0.0 – 1.0)

GET /v1/firearmsList and search endpoints include data_confidence for filtering

GET /v1/data/coverageEnterprise: GET /v1/data/coverage — field-level coverage percentages across all tablesEnterprise

GET /v1/data/confidenceEnterprise: GET /v1/data/confidence?below=0.5 — paginated list of low-confidence recordsEnterprise

Docs

How We Improve Scores

Our data pipeline continuously works to increase confidence across the database:

Automated collection from manufacturer spec sheets and procurement documents

Cross-reference against 4+ independent sources per firearm

Community contributions reviewed by our editorial team

Regular refresh cycles to capture updated specifications and new production variants