NFT Score & Rank

Most NFT platforms rank NFTs by rarity. However, in reality, many NFTs with rare traits remain at floor price levels because they lack visual appeal or desirable trait combinations (e.g. a mighty crown paired with a dainty princess dress). At the same time, some NFTs with non-rare traits become highly sought after due to aesthetic appeal, celebrity hype, or trending themes.

To address this issue, Catfish Market aims to replace rarity rankings with popularity rankings that more objectively reflect each NFT's true market value.

P.R.

P.R. (Popularity Rank) value : The rank of an NFT among all participating NFTs in its collection

Participating NFTs within a collection are ranked based on each NFT's score from all its completed battles.

NFT Score

Catfish Market uses the Elo algorithm to calculate each NFT's score, with the main variables being the outcome of each battle and the strength of the opponent.

Formula

NFT Score = Original NFT Score + K × (Battle Outcome − Expected Win Rate)

• K: K is the rating adjustment factor (set to 32).

• Battle Outcome Win = 1, Loss = 0, Draw = 0.5 (draws not currently supported)

• Expected Win Rate Expected Win Rate = 1 / (1 + 10^((Opponent NFT's Score − NFT's Own Score) / 400))

Initial value of NFT Score

The initial score of each NFT participating in battles is based on its highest bid on Blur, with 1500 as the midpoint.

For example, if the highest bid of all participating NFTs in a collection is 3 ETH (usually a trait bid), it corresponds to an initial score of 1800. At the same time, if the lowest bid is 1 ETH (typically the collection's top bid), it corresponds to a score of 1200. The remaining NFTs are assigned initial scores proportionally between 1200 and 1800 based on their individual bid prices.