The key to a great long-form generative art collection: the fat middle.

Much is said about the best pieces any given collection has to offer and that’s not without reason since they in a way are the “face” of the collection, the show stoppers, the ones that will always be at the tip of the tongue when you talk about those collections: No one can think of Fidenza and not imagine “The Tulip” and one can’t think about Ringers without having “The Goose” come to mind.

Ringers #879, The Goose.

But long-form collections are not 1/1s, if the best outputs are great but the average output is weak then by definition, this would be a weak collection and the artist would have been better off using his algorithm to release a few curated 1/1s and not a long-form collection. This is to say that the key to judging the quality of a long-form generative art collection is not its outliers (both positive and negative), but the fat middle that lives between them.

A great collection is in a way like a championship-winning team in the sport of your choice: You have your superstars, the faces of your franchise and behind them, you have a strong core of players, a “supporting cast” that makes or breaks your chances. Championship winning teams do not have the best superstars but the ones that have the strongest supporting cast.
In the context of a long-form collection, this “supporting cast” is represented by the hundreds of outputs that are a tier behind the ones you always remember. By the nature of their large numbers, they are usually not remembered immediately by anyone but their owners (and the few who dig deep at the collection), and the larger their numbers and higher their quality, the better the collection is gonna be.

I call this supporting cast the “fat middle”. If we imagine the whole range of outputs any given collection has and try to split them into tiers, we usually get four groups:

1. The superstars: You know who they are.
2. The strong core: A tier down from the superstars. In a way, they would be superstars themselves if they were on a “weaker team”.
3. The weak core: The line between these outputs and the ones on the tier above is the most subjective and nuanced of all, but idea is that they are the ones that while key to any championship winning team, you don’t really think they would ever be considered stars, even on a “weaker team”.
4. The bottom tier. These are the outputs you look at and think you would rather have another from the set if possible.

All those distinctions are obviously subjective and personal. Another way to think of them is that you would, for the most part, trade up from one tier to another while being somewhat indecisive between options on the same tier. It’s important to understand that those are relative values within the collection and not absolute judgment values. A bottom tier Fidenza is still more desirable to me than most other NFTs.

I would love to own Fidenza #766 even if I consider it a “bottom tier”.

In principle, every collection has pieces on all groups, even if we disagree where each piece fall (and I’m sure I’ll get a lot of flak for choosing #766 as the example above).

But collections are not really made by those “outliers”, they are made by the pieces on groups 2 and 3. How many of them are there? How good are they? The answer to those questions is what makes or breaks a long-form collection.

The fat middle concept

For a long-form collection to be structurally good, it needs groups 2 and 3 to have some specific characteristics:

• They need to be the vast majority of the outputs.
• A collector needs to be a little bit indifferent when choosing between them.

The first item is intuitive, you can’t have a large amount of group 1 by construction, so the difference comes to the bottom tier vs the middle. You don’t want a collection where a significant part of the outputs rely on the quality of their peers to be considered good. How many pieces in a collection do you wish were a little bit different? How many would you trade (1 for 1) against most of the rest of the collection and consider the trade a step up? If the number is too large, the collection does not work so well the way it was released (either mint size, should have been 1/1s, etc). Once again, we are not talking (yet) about the absolute quality of the outputs, only their relative value inside the collection.

The second item is the most counterintuitive because it sounds like an argument on the fungibility of NFTs, an oxymoron. But that’s not it, it’s more nuanced: the items are obviously not fungible, but they need to be consistent enough at eliciting an emotional response so that any given collector would not know how to choose between them. A prospective collector would be satisfied with either option. Their “fungibility” comes from the fact that you consider a large number of outputs somewhat equally desirable, a 1 for 1 trade is a lateral trade where what you get is not better or worst, just different.

When a collection has those 2 characteristics and the fat middle is of high quality, that’s when you hit the jackpot of long-form generative art. You have a strong, hundreds of pieces deep pool of options from which you have a hard time choosing and they all look incredibly appealing. The size, relative and absolute desirability of this fat middle is what turns your collection into a championship winning team.

I personally like exploring out of band outputs for the collections I like and usually, my experience can be summarized in a few ways:

• I lose the desire to continue to explore because too many outputs are not interesting enough (proportionally too many from the bottom tier).
• I keep the desire to refresh but skip to the next output too fast, I enter a kind of “search mode” for the best outputs (the fat middle is not interesting enough compared to the showstoppers)
• While I skip some outputs fast (bottom tiers are inevitable), I still stop to appreciate most of them. And I continue refreshing for a long time. Those are the true grail collections. A strong and large fat middle that keeps me engaged.

Important to note that I only do that for the collections I already like a lot, so I never “don’t like” the outputs. But being forced to see random ones gave me a much deeper understanding of the value of the collection as a single entity vs the sum of their pieces. The better collections are truly more than the sum of their outputs, and that’s what became clear to me.

I would like now to make a comment for those who got mad at me for using Fidenza #766 as the example of a bottom tier piece within Fidenza: The truly great collections are the ones that not only have the characteristics I said above but also there’s strong disagreement on which pieces are part of the bottom tier. Every collection has pieces that you would somewhat gladly trade for most others, but in the great ones, there’s no consensus on them. If we agreed on which pieces are the weakest, that would be a sign that Fidenza is not as good. Our disagreement is not a sign that one of us is wrong. It’s a statement of the quality of Fidenza. Our disagreement is a source of celebration, not frustration.