In an NFT collection, the floor price is a commonly understood point.
It’s the lowest price at which a seller will list one of the collection’s NFTs. And, short of bartering or sending over a custom offer, it’s the lowest price you’ll pay to get into that collection.
Floor price is everywhere. It’s become one of the key data points we look at when it comes to quickly assessing an NFT collection’s value. When the floor price is up, the collection is doing better. When it drops, it’s doing worse.
Though it may be commonly understood and easily found, it is a flawed way to evaluate NFTs.
So, why is the floor price wrong?
Because it doesn’t actually represent the floor price of a collection.
The Two Main Flaws In Floor Price
To find a new way forward from floor price, we need to first take a deeper look at its flaws.
There are two ways in which the floor price can be easily manipulated.
- Collusion — When an owner (as discussed in this thread about Goblintown) or owners in a collection agree to affect the price of their listed NFTs, they can use the changing floor price to their advantage.
- Rogue Sales — When someone needs to sell quickly, they can sell below floor price. This drops the floor price, despite there being no significant changes in the collection.
These are our two main concerns because they unfairly skew the floor price as an assessment of an NFT collection’s value. What the floor price represents is the lowest point at which a particular seller publicly says they’ll sell at, not what a collection’s value is at any given moment.
In a sense, the floor price is often given a similar amount of credence as a stock’s price. Though comparing stock investing to buying NFTs is imperfect (no one share is worth more than any other share), the parallel here would be remarking at a skyrocketing share price, only to find out that all the shareholders just decided that the price should be higher.
On the other end of the spectrum, someone selling their stock for half of the price for their own personal reasons shouldn’t be an indicator to the whole market that the business is now valued at half of the price.
Though imperfect, these trades work on OpenSea in the following ways:
- Owner creates a listing with a price and Buy Now option -> Buyer acquires it for that price.
- Buyer offers a price for the NFT -> Owner can accept the offer and sell for that price.
- Owner creates an auction.
- Owner creates a private listing available only for users the owner chooses.
How To Determine A New Floor
We see the value of having a number like floor price, even if we believe floor price as it is now isn’t delivering on the value. In theory, it would help us to understand a collection’s value — and how that value changes over time.
What if floor price could do that, though?
We’re setting out on determining a new floor price, one that takes into account actual trades. This is different because this historical flow price actually shows the minimum trade price that happened while excluding outliers.
Below, we have three listings: BAYC, Azuki, and Wabi-Sabi Collective Genesis. We are going to look at all of the listings for the collection over time (represented by purple dots) and all of the trades (blue dots).
What we see is that the actual trades are lower than listing prices.
Collection Link: https://opensea.io/collection/boredapeyachtclub
Collection Link: https://opensea.io/collection/azuki
Wabi-Sabi Collective Genesis
Collection Link: https://opensea.io/collection/wabi-sabi-official
While there are outliers in trades with lower prices, these are likely offline deals that are not relevant for the market analysis. What’s more, how much should these deals be impacting our assessment of a collection’s value?
What looking at trades in the past reveals is not just the owner’s perception of price, but also a demand from the buyer side.
By validating the buyer side demand, our concerns about the two ways the floor price can be manipulated in the present are no longer relevant.
How To Move Forward With Price Change
As an investor in some asset, you likely want to know how that asset performed in the past.
For instance, in the OpenSea marketplace you can find this information:
- Number of items
- Number of owners
- Floor price
- Total volume
You can also find this graph, which represents the 90 Day Average Price and 90 Day Volume — https://opensea.io/collection/wabi-sabi-official/activity
It’s easy to look at this graph and take away that the collection’s price is fluctuating wildly.
Does price really fluctuate that much? No. It’s more about averaging trades with different rarities.
Though buying one of the top 10 rarest NFTs for 3–5x the average price may be common, how much impact should that have on the value of a collection?
This is the issue with relying on a mean price — it’s too easily swayed by these outliers.
When floor price is relied on, that’s the outcome.
How Dakko Is Using Median For Price Change
So long as we continue to stick with floor price as our shorthand for a collection’s quality, we are held hostage to bad actors and volatile markets.
Instead, if we take the traded minimum price, remove the outliers, and consider the median instead of the mean, we get a floor price that’s more honest.
This is one of the primary motivations for developing Dakko, a machine learning-based algorithm designed to highlight valuable NFT collections.
Currently, we are running trials using medians for price changes to create an ever-more accurate Dakko Score, which helps us to understand an NFT collection’s viability over the long-term.
Let’s look and see how a median-based price compares to the current way.
Here we’re considering the same collections above. The blue line represents the collection’s median value, the green line represents the mean value.
If we’re taking the mean value as significant, we can see how easily swayed it is just by a few rare items or outliers. The truth is that these few sales don’t — and shouldn’t — cause a lasting impact on the collection’s real price.
With Azuki, we can see a similar clarity from using this new system. Following the mean, we see a number of spikes that suggest the collection is more valuable than it actually is, while it rarely suggests the collection is less valuable. Alternatively, the blue median line reflects a more consistent value.
Wabi-Sabi Collective Genesis
It is in collections like Wabi-Sabi that the value of this new floor price is more impactful. Not every collection will have the esteem of a BAYC, yet it’s important to be able to understand its performance. Here the median gives a clear sense of the collection’s value, even as frequent outliers affect the mean.
When considering just the floor price or just the mean price, we are susceptible to fluctuations and inaccuracies. That’s why with our next iterations of Dakko, we are taking the minimum (except outliers, 10 percent) for floor price (which is called the 10th quantile).
Will it work? We believe it will, and we want you to join us to find out.
Where We Go From Here
One problem that has plagued the NFT community is pump and dump schemes. Like other periods of fast growth in the internet’s history, there are always going to people taking advantage of an evolving space.
These schemes specifically see a community artificially inflate a collection’s price, then sell out of it. We recognize that even at Dakko’s best, we can’t prevent this level of collusion.
Instead, we are developing a way to avoid it completely.
By taking into account the number of top investors — that is, wallets with the most profit over the past 30 days — involved in a collection, we can begin to get a sense for the seriousness of the community and the collection. In collections that are seeing an intense level of activity, but very little support from established investors, we believe we can better identify and separate the valuable collections from the rest. We’ll explore this further soon.
To see how we are developing Dakko and apply to join our Closed Beta, click here.