A valuation you
can't verify is
just an opinion.
Six steps. Fully documented. Every piece of reasoning shown. This is how we get from a listing URL or a set of photos to an evidence-backed value range.
We collect real completed sales from multiple platforms — eBay, Facebook Marketplace, documented forum transactions, and dealer records. Asking prices are explicitly excluded at the data ingestion stage. A listing is not a transaction. The only number that matters is what someone actually paid.
Every transaction is recorded with: sale price, sale date, platform, bike configuration (make, model, year, colorway), condition notes, and provenance indicators where available. The database grows with every valuation processed.
Asking prices are excluded. Always. A listing tells you what a seller hopes to get. A sold transaction tells you what the market determined a bike was worth.
Vintage BMX bikes changed within production years — forks, components, and colorways were updated mid-run without model name changes. A 1984 Haro Sport with factory Araya 7X rims is a fundamentally different valuation than one with replacement wheels. Generic make/model matching misses this entirely.
The TrueSpec factory spec registry covers 101 bike variants spanning 1979–1988 across 19 makes. Each entry documents frame material, fork material, bars, stem, headset, cranks, bottom bracket, sprocket, chain, freewheel, pedals, wheels, tires, seat, seatpost, and brakes — with year-specific and mid-year variant notes where documented.
We name the bike, the year, the component. A correct spec match is the foundation of every valuation.
When you submit a bike for valuation, we extract the configuration from your photos or listing URL. Every component is identified and compared against the factory specification for that make, model, and year. Components are classified as: factory correct, period-correct aftermarket, incorrect for year, or replacement/unknown.
The configuration match drives two outputs: an originality score that tells you how factory-correct the build is, and a comp weighting that prioritizes comparable sold transactions with similar configurations. A fully original 1986 GT Performer is not directly comparable to one with replacement wheels and incorrect stem — both are real, but the market treats them differently.
Configuration matching at the component level is what separates TrueSpec from generic collectibles tools. It is the core of the product.
We identify the comparable sold transactions that most closely match the bike being valued. Comparables are weighted by three factors: recency (more recent sales weighted higher), condition proximity (condition grade of the comparable vs the submitted bike), and configuration similarity (how closely the sold bike's components match the submitted bike).
Each valuation output shows 3–5 comparable sold transactions — not asking prices, not estimated values. Actual transactions. You can see what each comparable sold for, when, and where. The comparable analysis is fully transparent because a valuation you cannot verify is just another opinion.
Every comparable is a completed transaction. The evidence is always shown. Never hidden.
We never present a single number as the value of a bike. The vintage BMX market has genuine variance — two nearly identical bikes sold $800 apart in the same month, one to a collector who knew exactly what it was and one that sat on eBay for 60 days because the photos were bad. That variance is real. Collapsing it to one number would be dishonest.
The range model calculates: Low (rough or incomplete examples, replacement components, poor presentation), Mid (clean original build, good condition, normal documentation), High (all-original, exceptional condition, documented provenance, top-tier presentation). Explicit adjustment factors are applied for originality, condition, provenance, and recent market movement.
Never a single number. Always a range. The width of the range is itself information.
The confidence score reflects how solid the valuation estimate is, based on two factors: data depth (how many comparable sold transactions exist for this variant) and comparability quality (how closely the available comparables match the submitted bike's configuration, condition, and recency).
A confidence score of 90%+ means strong data — multiple recent comparables closely matching the submitted configuration. A score of 60–75% means the data is thin — few comparables, older transactions, or limited configuration matches. A score below 60% means the valuation is directional only and should be treated accordingly. We say so explicitly.
The confidence score is never buried, minimized, or spun positively when it is low. If the data is thin, we say so. A low confidence score shown honestly is more valuable than a high score fabricated from insufficient data.
Where the data comes from.
All transaction data is sourced from completed sales. No asking prices, no estimates, no forum opinions about what something "should" be worth.
101 bikes.
1,746 data points.
The factory spec registry is the foundation of configuration matching. It covers 101 bike variants from 1979–1988 across 19 makes — each entry documenting every factory component at the time of production, including mid-year variants and production changes.
Sources are classified by confidence level: primary magazine tests (0.97–0.98), secondary catalog sources (0.88), and enthusiast documentation (0.80–0.85). Every entry carries its source citation and confidence score.
What we don't know, we say so.
Transparency is the only credible position. These are the known limitations of the TrueSpec method.
See the method in action.
Paste a listing URL or upload photos. Every step of the TrueSpec method runs in under 30 seconds — and every step is shown in your output.