NFT rarity and traits

Most generative collections assign each NFT a set of traits, such as background, clothing, accessories or eyes, drawn with different probabilities. An item with traits shared by only 1% of the collection, or an unusual combination, is considered rarer than one with common traits. Rarity tools rank every item using methods such as trait frequency or statistical rarity, and marketplaces display these ranks.

Rarity strongly influences pricing. Collectors pay premiums for one-of-a-kind traits or top-ranked items, while the floor is usually set by the most common pieces. During the 2021 boom, rarity ranking sites became essential tools, and some traders specialised in sniping underpriced rare items listed by sellers who had not checked their rank.

Rarity is not the same as desirability: an ugly rare trait can be worth less than a popular common one, and different ranking methods can disagree. Teams that know trait distributions before reveal could abuse that knowledge, and leaked metadata has let bots mint rare items. Rankings can also change if metadata is edited after mint.