Indices built from asking prices and realized auction results across DE, FR, IT, UK, NL and ES. Rebased to January 2019 = 100 and updated weekly.
Our indices only move when at least five constituent models have comparable price observations in consecutive months. We do not yet have that depth of repeated observation, so every index holds at its 100.00 baseline. We could publish a moving line — an earlier version of this calculation showed a 37% decline that turned out to be sampling noise from as few as two models per step. We would rather show a flat line and explain why. The data is being collected monthly; the indices will start moving once the history supports it.
Index level tracking begins as month-over-month history accumulates. Today we show current composition.
| Model | Years | Median price | Change | Liquidity |
|---|
Change is measured over the longest period available for each model; coverage is still building.
Models are assigned to an index by marque origin, era and segment (e.g. Japanese classics, Italian classics, seventies superbikes). Every tracked model with sufficient listings contributes.
For each constituent we compute a monthly median from observed European listings and auction results, requiring at least three observations in a month before a median is published — thin months are left blank rather than estimated.
Category indices are equal-weighted across constituents with published medians, rebased to 100 at each index's base month. The index level moves via chain-linking: month-to-month it reflects the price change of constituents present in both months.
Where an index has few months of overlapping data, its level stays near its base and we say so. Index history deepens as our weekly collection accumulates — we show current composition until the series is meaningful.
Full methodology available in our Reports.