Classic car auction data methodology
Old Cars Data normalizes collector-car auction records across multiple sources so developers, appraisers, pricing tools, and AI agents can query comparable sales, live listings, bid trails, and market statistics through consistent REST and MCP interfaces.
Direct answer
Old Cars Data normalizes collector-car auction records across multiple sources so developers, appraisers, pricing tools, and AI agents can query comparable sales, live listings, bid trails, and market statistics through consistent REST and MCP interfaces.
Provenance: methodology is based on Old Cars Data normalized auction records, source URLs where available, REST API behavior, and MCP tool contracts. Data catalog freshness: August 3, 2026.
Pipeline
How records are normalized
- 01
Source records are mapped into common fields for make, model, year, price, status, dates, source, location, vehicle details, and listing text.
- 02
Make and model names are normalized so API users can query canonical values instead of guessing from listing titles.
- 03
Completed auction results and live listings are treated as different market signals because asking, bidding, and final-sale behavior answer different questions.
- 04
Bid trails are exposed separately where available because bid progression can show demand intensity beyond the final price.
Field conventions
How to interpret dates and prices
These fields preserve source context but do not imply timezone or buyer-premium conventions that the source did not provide.
- auction_end_at
- A timezone-naive wall-clock value. No source or venue timezone is identified or guaranteed, so it must not be treated as UTC or as an unambiguous instant.
- auction_end_precision
- exact means the time-of-day was captured; it describes granularity, not timezone certainty. date means only the calendar date is known.
- price
- The amount captured from the source listing. Old Cars Data does not add buyer premiums, taxes, or fees, but a source-published result may already include them. Premium treatment is not normalized or exposed as a per-record field.
When timezone or premium treatment matters, inspect the returned source and url and verify the original listing.
Market signals
Completed sales vs live listings
Different signals answer different questions, so they are exposed independently.
- Completed saleA resolved auction outcome with a final status and price when availableComparable-sale research and market history
- Live listingAn active auction with changing price, bids, and ending timeMonitoring current market activity and candidate deals
- Bid trailA sequence of bids over time for supported auctionsUnderstanding demand, late bidding, and bidder behavior
- Market statisticsAggregated pricing and sell-through contextSummaries, dashboards, and initial valuation context
Definitions
How current market statistics are calculated
These definitions make the figures reproducible and prevent a market summary from being mistaken for a formal appraisal.
- Eligible sale
- A completed record marked sold with a positive recorded sale price. Active listings and non-sold outcomes are not included in sold-price statistics.
- Current model window
- Model-page medians, ranges, and counts use the most recent 12 months of eligible sales.
- Median
- The 50th percentile of eligible sold prices. Half of the recorded results are below it and half are above it.
- Typical range
- The 25th through 75th percentiles of eligible sold prices. It is an interquartile range, not the lowest and highest observed price.
- API currency handling
- The stats endpoint reports USD records; blank currency values are treated as USD. It does not silently combine unlike currencies.
The REST API also supports make, model, and year filters. See the statistics endpoint documentation for request and response examples.
Guidance
How to use statistics responsibly
Statistics should be used as context, not as a blind final value. Sparse markets, unusual configurations, modified vehicles, outlier sales, premium venues, incomplete listings, and condition differences can all move collector-car prices materially.
For valuation workflows, start with canonical make and model data, filter comparable sales by year and configuration where possible, inspect listing details, compare against live auction signals, and treat aggregate statistics as a market summary rather than a replacement for expert review.
Reference notes
Frequently asked questions.
Are Old Cars Data values appraisals?
No. Old Cars Data provides market records, statistics, and comparable-sale context. A formal appraisal should consider inspection, condition, provenance, location, documentation, and expert judgment.
Why can two similar cars sell for different prices?
Collector-car prices vary by condition, originality, mileage, options, color, documentation, seller reputation, photography, timing, venue, and bidder competition.
How often is Old Cars Data updated?
Newly collected records are ingested and normalized on a 15-minute refresh cycle. Source publication timing and field availability vary by auction platform.