Data methodology
About our data
A plain-language account of what our numbers mean, how they are produced, and where their limits are.
Where the data comes from
For the Latvian catalogue, we use publicly available SS.lv listings. It is the catalogue’s only listing source in Latvia.
autopase.lv is not affiliated with SS.lv. The prices shown are sellers’ asking prices, not completed transaction prices.
How many listings are in the catalogue
The catalogue currently contains 25,973 active listings with an image. This count uses the same active-catalogue population as search and refreshes once a day.
How often the data is updated
Different parts of the catalogue run on different schedules:
- new-listing feed — every 10 minutes;
- listing details — every 2 hours;
- photos — every hour;
- price snapshot — daily at 06:00 UTC;
- deal-score recalculation — daily at 06:30 UTC;
- automated direct-URL checks — every 3 hours (a diagnostic signal, not a basis for removing a listing);
- overnight brand check — from 01:00 UTC (currently without listing removal).
A direct-URL check is unreliable because SS.lv can return the page of a removed listing as active. We normally remove a listing when the full overnight check of its brand no longer finds it; until the next brand check, it may remain in the catalogue for a while. Removal of sold listings is currently paused while we restore this check. As a result, some cars in the catalogue may already have been sold. We will mark them as removed as soon as the overnight check starts confirming removals again.
How the deal score is calculated
We compare each car with similar listings by brand, model, year and mileage; when the exact group is too small, the comparison widens to the brand. The result uses a 0–100 scale. A value of 50 can mean market level, but it is also the internal result when there is not enough comparison data. That is why a deal badge appears only from 70 points. The named levels begin at 70, 80 and 90 points. The calculation uses the price difference plus small adjustments for positive or warning signals found in the listing text. Removed listings are also included in the comparison set; they are currently the majority of the sample, including listings first seen around six months ago, because the comparison has no listing-age limit. The score is therefore a guide rather than a technical inspection of the car.
Price history
A price-history entry is written only when a listing price changes. We show the reductions that we recorded during subsequent checks, but we do not use this table to construct an “average price for the month”.
What we do not know
- Without a separate check, we cannot confirm the VIN, the car’s condition, or its ownership history.
- We do not know the completed transaction price after a sale.
Data for press and research
For a citable market cut or an explanation of a specific metric, visit our press data page.
Data for the press →