Price vs. market
How the asking price sits against the comparable-market median.
Methodology / v1
A score should show its work.
MileMosaic turns a vehicle listing into a 0–100 market read. It is a consistent way to ask, “How does this listing compare with the cars around it?” It is not a crystal ball, a certified valuation, or a substitute for looking at the car.
The readout
0—100
Weighted, bounded, explainable.
The overall result is the weighted average of four factor scores, rounded and clamped so it always stays between 0 and 100.
The equation
Four signals. One transparent blend.
Price gets the largest voice because it is the question most shoppers are trying to answer. The other three signals keep that answer in context.
40%
Mileage vs. market
Whether the odometer reading is favorable for this peer group.
Age fit
Whether the price makes sense for the vehicle’s model year.
Condition
The submitted condition, adjusted against the peers’ average condition.
overall = price × .40 + mileage × .20 + age fit × .20 + condition × .20
Inside the factors
The details behind each number.
01
Price starts with the middle of the market.
We compare the asking price with the peer median. The 25th and 75th percentiles create guardrails: 90% of P25 is the favorable anchor and 120% of P75 is the unfavorable anchor. Between those anchors, the score moves linearly. That gives an unusually low or high listing room to stand out without pretending the median is a precise appraisal.
02
Mileage is relative, not absolute.
Mileage uses the same P25/P75-derived anchors as price: P25 × 0.9 and P75 × 1.2. Lower mileage is favorable within this group, but the read is always relative to comparable vehicles. A 70,000-mile car does not mean the same thing for every make, model, and year.
03
Age fit follows the local price curve.
We fit a straight price-versus-year line to the comparable listings. With five or more peers, the highest and lowest 10% of the points are trimmed before fitting. A listing at 70% or less of the fitted expected price scores 100; at 130% or more it scores 0; the space between is linear.
04
Condition is a calibrated input.
The submitted label maps to a starting value: excellent 100, good 80, fair 55, and poor 30. We then adjust it against the comparable group’s average condition and clamp the result to the 0–100 range. It is a structured signal, not an inspection or a promise about mechanical health.
A deliberate boundary
Drivetrain is currently informational, not a weighted input. We show it in the listing details so you can interpret the result, but FWD, RWD, AWD, and 4WD do not add or subtract points in the current scoring policy.
Comparable selection
Start close. Widen only when needed.
The engine removes the listing itself, then looks for a usable group. It normally needs three or more matches before escalating to a looser tier.
- 1
Exact
exact tierSame make and model, trim-compatible, within one model year.
The tightest read when the listing has enough close peers.
- 2
Close
close tierSame make and model, within three model years.
A broader view when the exact group is too small.
- 3
Wide
wide tierSame make, any model year.
The fallback when the model-specific market is sparse.
Freshness
A timestamp, not a freshness promise.
What we can say
createdAt
When the listing was stored.
generatedAt
When this score was generated.
A result’s generated-at line tells you when MileMosaic ran the calculation. It does not guarantee that every source was refreshed at that moment, and we do not claim a source-by-source freshness SLA. Market data can change between a listing being stored and a score being read.
Known limitations
Useful context has edges.
Sparse peers soften certainty.
Missing or insufficient peers return neutral 50 factor values and low confidence. A number is not statistical certainty just because it is precise.
A listing is not an inspection.
The model sees submitted price, mileage, year, make, model, trim, location, drivetrain, and condition. It cannot verify title history, accident damage, options, or mechanical health.
Markets move.
The read reflects the comparable pool available to the engine. Inventory, seasonality, geography, and unusual vehicles can all make a comparison less representative.
The policy is intentionally simple.
This version uses four weighted factors. Drivetrain is visible context today, not a score input; that policy may evolve only with a clearly documented change.
What this score is not
It is not the final word on a car.
It is not a guarantee of value, a purchase recommendation, a financing decision, an inspection report, or a prediction of what a specific buyer will pay. It is a transparent starting point for asking better questions.
Read the factors, check the vehicle, and make the call with the full context.
See the policy in motion
Start with a real, documented read.
The seeded Camry example shows the score, its factor breakdown, the comparable count, and the time the result was generated.