LowMileageInsurance.co.uk — mileage distribution study, version 1.0 Research draft dated 14 October 2026. Named editorial approval is pending. Question What proportion of cars fall below meaningful annual mileage band boundaries? This is original secondary analysis of published aggregate statistics, not a new survey, insurance pricing study, or analysis of individual vehicle records. Primary source Department for Transport, National Travel Survey, NTS0904, released 10 September 2026, annual mileage band of cars in England, 2002–2025. The downloaded ODS and source.json contain the exact source URL, retrieval date, byte count and SHA-256. The landing page is https://www.gov.uk/government/statistical-data-sets/nts09-vehicle-mileage-and-occupancy DfT lists the next workbook update as summer 2027. Check the landing page for revisions before using a newer release. The build never fetches replacements. Population, measure and exclusions Weighted share of cars in the England household survey sample after the table's exclusions. This is not a count of drivers, policyholders, households, or the whole UK registered vehicle fleet. The unweighted sample is 11,918 cars in 2025. Total annual mileage is estimated by the respondent, not an odometer reading. Three- and four-wheeled cars and SUVs are included. Cars are excluded where age, ownership or estimated trip-purpose mileage is missing. See the ODS Notes sheet. Pre-2013 figures in this release have been revised to include unreported and other fuel types, consistently with post-2013 coverage. All years use the same pinned release; older editions are not spliced into the time series. Calculation and precision Read numeric office:value values from NTS0904 in content.xml inside the ODS. The visible whole-number percentages are display rounding. Do not sum those rounded labels or renormalise them to 100. All underlying band shares must sum to 100 within 0.00001 percentage point. Sample sizes remain unweighted counts. Below 5,000 miles = sum of B:G; below 7,000 = B:H. In 2025 these are row 30. Below 3,000 = B:E; below 9,000 = B:I. Use complete bands only. Each result means strictly less than the threshold, not less than or equal to it. Results use underlying precision and are rounded once, to 1 decimal place in tables and whole percentages in headlines. A difference is calculated before rounding and expressed in percentage points, not a percentage growth rate. The exported threshold calculations are floating-point numbers; comparisons should allow a numerical tolerance of 1e-10, not exact decimal string equality. The tidy band CSV preserves original numeric values as decimal strings. Every row includes its source sheet/cell and the applicable unweighted sample size. Do not sum sample sizes repeated over mileage bands to estimate a sample total. The 5,000–6,999 band cannot identify how many cars are below 6,000. We do not interpolate, invent an exact median, or infer policy eligibility or savings. Uncertainty and comparability The survey is subject to sampling, measurement and non-response error. We have not calculated confidence intervals or tested statistical significance. The unweighted base alone does not support simple-random-sample confidence limits for a weighted complex survey. Small differences are descriptive, not evidence of a reliable population change. 2020–2022 reflect pandemic behaviour, collection changes and smaller samples. Samples expanded from 2023. DfT says statistics based on interview questions are unaffected by the 2025 digital diary change; do not substitute diary-based miles per person for this annual car measure. Quality: https://www.gov.uk/government/statistics/national-travel-survey-2025/nts-2025-quality-report Reproduce 1. Unzip research-bundle.zip into an empty directory. 2. Run: python reproduce/study.py --output reproduced Python 3.10+ standard library only. No internet or personal data required. 3. Compare reproduced/mileage-study.json, mileage-bands.csv and mileage-thresholds.csv with the copies at the root of the bundle. Same input and code produce identical bytes. Each output has a SHA-256 in checksums.json, supplied alongside the bundle on the website. 4. Optional figure regeneration: install matplotlib in your own environment, then run python reproduce/plot.py. Figures appear in reproduce/charts. Figure rendering may vary with matplotlib/font versions; underlying numbers do not. The supplied SVG and PNG images include attribution and limitations. To reproduce inside the Site repository: python research/study.py --output var/reproduced Market report method The initial purposive set contains seven records: specialist propositions, conventional annual information, a comparison route and an archived brand. Counts describe this set, not the number of UK insurers or market share. Records are sorted alphabetically; no ranking or price comparisons are calculated. The selected local review catalogue supplies all commercial facts. Each export preserves source URLs, observation dates, uncertainty, and a review label. Published information does not prove a quote can be obtained. An unverified journey is not evidence of closure. An official closure notice is retained as a historical record. No personal quotation or purchase has been completed. Automatic fetch results are separate from editorial observations. A failed fetch preserves the last success and does not rewrite availability. Checks of provider evidence identify this report as an affected page. The report is rebuilt when reviewed source data or the monitoring snapshot is updated. This hosted edition is static between deployments; the prior monitoring automation does not automatically redeploy this Site. Archive snapshots carry a content revision and are retained in versions.json and the bundle's history directory. Machine-readable status market-availability.json and mileage-study.json are review data. They do not assert canonical IMR approval. approved-facts.json is deliberately empty until a verified approved canonical release is connected. A future read-only registry feed requires that connection; this download is not a substitute API. Reuse and attribution Contains public sector information licensed under the Open Government Licence v3.0: https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/ Suggested citation: Department for Transport, National Travel Survey NTS0904, 2025 data (10 September 2026 release); analysis by LowMileageInsurance.co.uk, version 1.0, 14 October 2026. Reuse the DfT aggregates under OGL v3.0 and preserve the source credit, geography, year, definitions and uncertainty notes. Do not imply DfT endorsement. Provider website material retains its existing rights. The charts and machine-readable files are supplied for review and reuse with this attribution; they are not yet an editor-approved public release.