Thread by @SfermionR

The way NFT collections are created and traits distributed varies across all projects. Some collections feature more gradual shifts between rarity levels within the collection while others possess more dramatic shifts. Others lack definition between rarity levels at all.

We believe that a greater understanding of the pricing structures within collections is necessary for the greater financialization of NFTs.

We found that not all rarity is created equal across the 12 collections we examined.
sfermionresearch.com/research/not-a…

Rarity distributions across the collections broadly followed the same pattern, skewing heavily to the left and tailing off as rarity increases. The distributions below illustrate this.

However, there were a couple notable exceptions, mainly #Doodles and #moonbirds. Each of these collections feature a few significant humps in the distribution as it tails off indicating higher pockets of rarity concentrations at levels where they should be diminishing instead.

In order to determine the impact of rarity on a collection's performance and trading behavior, we segmented the sales data for each collection by rarity quartile and examined prices, their standard deviations, and trading frequency.

Coming into this exercise we assumed that trade counts would vary by rarity quartile with rarer assets trading less frequently than more common assets. This assumption turned out to be the case broadly across the collections we examined.

More interestingly, the differences between quartiles turned out to be more aligned than expected with the first quartile making up ~30% of transaction volume, 2nd and 3rd quartiles ~25%, and the 4th quartile around 20%.

The relationship between trading frequency by quartiles proved to be remarkably consistent across time and collection.

How about pricing by quartile? We came into this analysis with the assumption that asset rarity would impact pricing in a fairly uniform manner with pricing increasing in conjunction with relative rarity.

While technically correct, we found that the differences in price between the first three quartiles were so minor that scalping based purely on rarity would prove to be a challenging exercise in the best of times.

Assets with rarities in the fourth quartile stood out in both the premium over the other quartiles and increased standard deviation. This pattern held true across all of the collections we examined with the exception of two.

The first was #deadfellaz which lacked significant definition in rarity and resulted in muddled behavior across quartile.

The other was #BAYC which, behaved much more like Deadfellaz than the other collections analyzed in that pricing differences between quartiles and standard deviations were minor and largely consistent across the collection regardless of rarity.

In this way, the BAYC collection is very much blue-chip through and through regardless of rarity. Other high-flying collections did not conform to this behavior making BAYC a standout in the collections examined.

Broadly speaking, assets with rarities occurring within the upper quartile trade at a fairly consistent premium to the next highest quartile which suggests buying opportunities should one of these assets be mispriced, at least in an appreciating market.

The consistency of behavior between the first three quartiles suggests that the financialization of NFTs is underway with benchmarks, more accurate pricing metrics, and lending markets able to materialize around these assets on a much larger scale than we previously thought.

The bad news is that your NFT with middling rarity isn’t going to suddenly spike in price relative to others. The good news is that the middle actually makes up the bulk of all collections, which suggests that you really should just buy NFTs that you like.