How we verify our data

A pull rate is the published chance of finding a given figure in a sealed blind box. We read these off the manufacturer's own case sheet or the labeled lineup graphic that ships with a series, the same artwork printed on the display case. We do not average the contradictory numbers that circulate on resale blogs, and we do not estimate odds for a figure whose official ratio we have not seen. If a figure has no verified ratio, its page simply does not claim one.

Blind draws are independent. A stated 1 in 144 means roughly one in every 144 boxes, not a guarantee after 144 boxes. Our box-count answers are expected values from that ratio, and we say so.

Sizes are the official listed dimensions for a series, in centimetres and inches, shown next to a real iPhone for honest scale. Where a series lists a range, we mark the tallest figure. We do not invent a measurement for a series that has not published one.

Retail prices are captured per market straight from the retailer's own listing, stored with the currency, the market, and the capture date. Every price on the site shows when it was taken, because a collectible price moves and a stale figure is worse than none. Secondary-market value changes daily, so we point you to check a live listing rather than quoting a single resale number as fact.

A figure's popularity score is a 0-100 read of how sought-after it is relative to other figures from the same brand- blended from real demand signals, not a single vendor's opinion. It is not a sales rank and it is not a price; it is how this figure's measured demand compares to its siblings.

Blend v1 pins exactly one source per signal, so every input is traceable back to somewhere real:

We also capture orders, units sold, and buy-button clicks per figure, but blend v1 does not use them yet - they are stored for a future revision, not silently folded in.

Each signal is ranked against other figures from the same brand(a small indie brand's numbers are never drowned out by a giant one's), then the ranks are blended by the weights above. If a figure is missing some signals, the score re-weights across whatever we do have rather than penalizing it for a gap in our data.

No measured demand means no score, not a zero. A figure with zero captured signals shows no popularity score at all on its page, the same real-data-only rule as everywhere else on this site - we do not manufacture a low number just to fill the space.

If we cannot source a fact, we leave it blank rather than fill it in. Every page tells you where its numbers came from and when. That is the difference between a catalog you can cite and a page that just sounds confident.