top of page

Quantitative approaches to NFT trading

Whether you are new to the NFT metaverse or an experienced jpegs trader, we are sure this article will present you with some interesting new ideas and operational insights, enjoy your reading...

Illiquid markets mean high risk

NFT markets are illiquid. This means that if you own an NFT you may not be able to sell it until a buyer comes along who is willing to buy your very NFT. The risk of not being able to sell your NFT is high, and the minimum price at which the last sale was made may not mean much. When trading NFTs, the trader has the possibility of losing the entire initial investment. An analogy of an illiquid market can be understood with cryptos that are traded in exchanges with central limit orderbooks and low liquidity. A buy or sell in the market can move the price considerably, just as floor prices can be unbalanced if supply and demand on a collection are disproportionate. Not all collections are equally risky. As with altcoins with large or small market capitalization, higher market capitalization usually means greater liquidity and lower volatility. In general, projects that have performed well over a period of time are the most likely to hold their price. The more time that has passed in a collection that has maintained value, the more it is a sign that investors consider it valuable in the long run.

Most NFT collections fail

For the indicators we will show in this article, we analyzed all the declines in NFT collections that occurred in October on the Ethereum blockchain (where 80 percent of NFT volume is traded). The amount of collections launched each day was so high (more than 30 per day) that we had to filter somehow to choose only the most popular ones. So the filtering criterion was to include in the analysis only NFT collections that had more than 5k users on Discord and more than 5k followers on Twitter. We ended up with 102 collections. We followed these projects prior to their launch until they reached the secondary market, capturing data such as the amount of the collection bid that was minted, or the minimum price. Baseline price data were taken just 2 days after the mint date, as our interest was to measure short-term performance, more akin to trading than investing.

In fact, if we analyze the successful collections that are now ranked among the most traded and the most expensive, we find that it took more than 2 days for all of them to appreciate in price several times from the initial mint price. The life cycle of an NFT collection that appreciates over time usually lasts several months, with short-term hype/mania cycles that can marginally and rapidly increase the price. But what is the case with the 90 percent of collections that do not gain traction and the price drops steadily over time? We have found that most of these collections lose trader interest within 1-2 days and then the base price continues to decline gradually over time.

The first sign that an NFT collection is about to hit the secondary market at a price above mint is if the collection is fully minted. This means that all the supply in the primary market has been covered by higher demand. Since the demand is greater than the supply, it is very likely that the secondary market prices are higher than the mint price and therefore, if you wanted to sell a piece of that collection, you would make a profit.

So, of these 102 collections, how many were fully minted? Only 36, or just over a third. In the next graph you can see on the X-axis all the collections analyzed, and on the Y-axis the percentage of the supply coined. Highlighted in green are the 36 collections that managed to be fully coined, while highlighted in red are the majority of those that failed to coin even 25 percent of the supply.

The problem with these collections that fail to sell most of their supply is that if we were to undermine a piece from that collection and try to sell it later in a market, we could probably incur a loss by trying to sell it at a lower price than mint or even fail to sell it at all. The latter case is due to the nature of NFTs, highly illiquid markets where demand has to come organically and on the other side you need someone to specifically buy that NFT from you. The problem is compounded with rarity. If your NFT is not very rare in the collection ranking, there is a high probability that a buyer interested in the collection will find it more tempting to buy a rarer piece, even if at a slightly higher price.

There are several reasons why most of these collections fell below demand. One is that we found evidence (see next chart) that many of these collections used bots to fake social media activity and fool investors into thinking that these collections had a lot of interest before the mint began. Another reason why demand is not being met is that the mint price is too high, or the available supply is too high, or both. Usually this is because new projects try to set mint prices and supply equal to those of projects that were very successful before them. This may not be a good benchmark, since the demand for each project before the mint date is completely different.

Analyzing the "Roof of the Mint" quantitatively measures the supply of each collection in the primary market. To calculate this metric, we multiplied the mint price by the supply. This metric can help us identify popular and highly anticipated collections, or collections that are aiming too high and may be eager to oversell while their real demand remains low. In the next graph, we can see that most collections aim to have less than $5 million in mint market cap, a good estimate to consider for future NFT launches:

Returning to profit, of the fully minted collections (36) only 20 revalued after two days. This is another consequence of illiquid markets: even if a collection has been fully minted, demand for the collection must continue to grow for the price not to fall. Because many traders are looking for short-term appreciation when nfts come on the secondary market, the underlying price can easily fall if demand is absent (a continuing race to try to set the lowest price among traders looking for an exit while demand does not come). In the graph below you can see the returns of the NFT collections analyzed. There are two bars for each collection because we included fees in two scenarios, one with low gas prices and another with very high gas fees. We took the lowest and highest gas prices we saw over a one-month period. Few collections performed massively, while most failed.

Commissions are critical to keep track of when trading, and even more so in the NFT world. There are two types: blockchain commissions and marketplace commissions. Blockchain commissions play an important role in Ethereum, but not so much in other blockchains for a buyer. Marketplace commissions vary widely from market to market. In this example we took OpenSea's commissions, namely sales commissions (2.5 percent), royalty fees (variable), and initialization fees (paid once per collection). We included both in our analysis. For OpenSea, the person who initiates the transaction is the one who pays the gas fee. This has a direct implication on profit, so choose wisely whether you decide to sell by direct listing or by auction.

Indicators to measure demand and momentum

Let's look at some indicators that can be used when analyzing a collection in the secondary market. One of the most popular indicators for comparing different collections is examining trading volume over time. Using this indicator as a demand signal could give misleading results. In the next example we can see the traded volume over time of the Bored Apes collection. As can be seen, although the trading volume decreased dramatically at one point, the minimum price of the collection remained relatively stable over time. This is logical considering that when a collection increases in price it becomes less accessible and therefore less volume is expected to be traded, but this does not mean that investors have lost interest in the collection and the floor price can certainly be maintained.

Intuitively there is a better indicator for measuring the demand for a collection, and although it is fairly obvious, in our experience we have not found many people looking at it. It is the total number of bids for a collection. This figure is only measurable in marketplaces that allow bidding, such as OpenSea. As can be seen in the next graph, taken from the Crypto Punks collection, a sudden increase in the number of bids served as a leading indicator of floor price appreciation. This makes sense when one considers that when a buyer is willing to spend the amount that a Crypto Punk costs, he or she will first try to send some bids on certain pieces and, if they are not accepted, will buy at the market price. This is an analogy of how orderbooks work, with limit and market orders. Some bots simulate really undervalued bids to try to fool sellers, but we believe this activity is relatively minimal.

Measuring the pump of a collection can be very complicated. Many collections may have low trading volume and prices a floor may drop, but rare pieces may have high value and be actively traded. In this case, we decided to analyze the number of transactions that occurred with the Loot collection over time and rank them by size. In light color are shown the cheapest transactions, which occurred mostly at the beginning of the project, while in darker color are shown the most expensive transactions (some of them over 50 ETH). It can be seen that over time the high amount transactions have been missing, a sign that the collection has lost interest. Otherwise, we believe that the number of expensive transactions that would have occurred over time would have somehow remained.

Long-term vision and supply shock

Knowing that NFT appreciation is a long-term game, we find it very useful to follow metrics that can indicate the long-term perspective of NFT collection holders. Our preferred metric is simply the product of calculating the percentage of the collection that is priced in the markets. A percentage below 20 percent, in our experience, is a positive sign that speculators are mostly out of the game and that most holders are trying to hold their piece for the long term. This indicator can be difficult to obtain historically for many collections, but it can be calculated instantly in any marketplace simply by looking at the number of pieces listed. Super affordable.

How to play with rarity rankings

Most popular NFT collections contain attributes that allow each NFT to be ranked from rarest to least rare. Several websites include these rarity rankings, and the rare pieces at the top of the rankings are always more expensive. The methodology for calculating this rarity ranking varies greatly from market to market, because the rarity rating is always something subjective that can be somewhat agreed upon and modeled mathematically through a normalization or weighting method. Even the most popular website for tracking rarity ranking among Ethereum profile image collections,, has listed 4 different methodologies:

In their case the "Rarity Score" is the default one, which is expected to be the most correct in most collections. In a collection usually the agreement for rarity ranking tends to be the methodology that is being followed by the website being used. In our analysis, we have found that most collections follow a rarity distribution in which the price increases exponentially as soon as one approaches the 5-10% rarest pieces, as can be seen in the next graph, which shows the current selling prices of all pieces in the collection that are listed on the market. While there may not be much difference in price between a piece with rarity less than 90% and one with rarity less than 50%, the rarest pieces can vary by several orders of magnitude. This can also help identify pieces whose price is lower than the distribution of the price curve that the collection is following.

Ideas similar to fundamental analysis

Estimating the expected future value of NFT collections before they are minted can be a daunting task. We have found online groups of people and influencers who have developed a framework for assigning different scores to certain characteristics that may contribute to the success of an NFT collection. This is just one example, and is an exercise for the reader to think about how to evaluate different attributes. Do you think the utility of the NFT is more important than the investors behind it? How important is it? Can the anonymity of the team behind the project influence some investors? How much do the advisors matter? How can the attractiveness of different NFT collections be measured and compared?

The future of NFTs

Public opinion about NFTs is still convinced that it is a silly trend that is close to disappearing. But the reality is that interest in NFTs continues to grow: the amount of searches containing the term NFT is certainly not at an all-time high, according to Google Search Trends. However, volumes have recovered from the activity seen this summer. As has happened several times before, the NFT craze is occurring in waves, and so far each new wave has been an order of magnitude larger than the previous one. This cannot be expected to continue to happen forever, but there is a greater chance that interest in NFTs will increase or at least persist when much of the retail demand can be met (Coinbase marketplace whitelist registrations have exceeded 3M). Currently many of these users have never used DeFi before and will have to overcome the hurdles of learning how to use custody and wallet, or pay blockchain fees. We think that many users who have so far found investing in cryptocurrencies with an environment of charts, candlesticks, and terminals daunting or tedious may find their way paved by these "Jpegs." After all, most assets traded in the "real world" are not fungible, and an increasing part of our daily lives are found online.

The analyses and ideas provided here are for educational purposes; none of this can be considered financial advice, and we expect our readers to do their own research when deciding to purchase any assets.

Datas powered by IntoTheBlock


Iscriviti alla Newsletter


Grazie per l'iscrizione!

Latest Articles

bottom of page