It is notoriously difficult to price NFTs. Low liquidity, high asset diversity and subjectivity of Art make it impossible to have a single price-point for any specific Item. This is not only a problem for individual investors who want to know how much to bid or ask for NFTs, but also for institutions or DAOs who need to have an up-to-date valuation for NFT portfolios. We kept facing this problem over and over across many of our projects like 1of1.works and our portfolio tracking software. We, therefore, developed The NAVgator as a first attempt to create a tool that can give a more nuanced and informed overview of pricing, specifically for Collectibles.
At Dialectic, we differentiate most NFTs into three main categories: Art, Gaming and Collectibles.
Usually, for Art, the pieces are 1:1s or low-count editions, and the tokens don't have many overlapping traits; this means that each piece or edition needs to be assessed separately and on its own terms, here the driving factor of prices are: aesthetics, theme, artist recognition and other external factors. This makes it rather difficult to use an automated or algorithmic approach for pricing. Therefore Art is usually best priced by Art professionals or crowdsourced appraisals such as Abacus.wtf.
For Gaming, it’s quite the opposite, where often it’s possible to compute a rather precise value for NFTs based on either the expected revenues that can be generated by using the token in-game (Axies for example) or if there is enough overlap between tokens (Lands with 3 size-types for example) strictly comparative pricing can be used. While there may still be some variation based on other factors, so far it’s not the biggest pain point.
The opportunity comes when we are looking at Collectibles, here there are usually many traits with differing rarity and aesthetic appeal and each token has a unique combination of traits and sometimes even a differing number of traits. The sale prices of tokens vary by many multiples in a collection across tokens and traits. Since there is often some level of comparative data for individual traits, it is possible to gather and segment this data into a more cohesive price profile for individual tokens. Making collectibles the most approachable category for automated pricing.
To achieve this we have developed a few different pricing strategies that aim to cover all angles through which investors value NFTs and then have a simple metric to define a “fair” price. This method is of course not able to cover 100% of the tokens as in highly diverse, very young or low volume collections the accuracy will be low.
This is usually the first avenue investors go down to establish the price of Collectibles, tools such as rarity.tools give users an easy overview of rarity ranking and some investors will purely base themselves on comparative rarity pricing, for example, if they need to choose between buying two tokens at the same price they will naturally buy the “rarer” one.
There are many problems with purely using rarity, firstly the correlation between prices and rarity is weaker than most people expect. This could be due to the fact that a larger percentage of investors look at aesthetics more than a rarity. Or there may be other external factors that are not included in trait rarities like memes and meta-culture.
Since many collections have different ways of doing rarity and have extra mechanics such as staking, burning etc. rarity is not a silver bullet that gets us to precise pricing, but it is one aspect of the overall price function.
This is a very basic way of defining the value of a token, we identify the trait that has the highest rarity (i.e lowest number of occurrences in the collection) and then we look at the lowest listing price (i.e the floor price) of any other token with that trait.
Pros: It gives us an indication of the lowest value anyone is willing to sell at. And if anyone wanted to buy a token with that trait, this would be the least amount they would have to spend. Given that there is certainly a rarity component to pricing this gives us an indication of what the market values this given rare trait at.
Cons: The token may have another trait that is less rare but more sought after and therefore of higher value. Especially for very rare traits, the lowest listing may never be filled, if the seller is not actually willing to decrease the listing over time, but no one is interested in buying it, this doesn’t give us any useful information. Additionally, if the token has multiple rare traits only the rarest is considered and the others are ignored.