70s Pop Series One, probability is my paint.
At 6 pm on the 7th of April 2020, I launched the 70s Pop Series One collection on the Art Blocks platform. Here I'm going to explain what that all means, what it is, and what it isn't. Then a bit about the artistic and technical aspect of it.
But first I'm going to cover the money side of it because it's probably the reason why many people are here, the part they have the most questions about, the most attention-grabbing, and the bit most likely to make people then read the part I really want to write about - the thinking behind it.
💰 💰 💰
70s Pop Series One is a collection of 256 different artworks, all created by the same bit of generative art computer code (explained below). After claiming the first one for myself, I pushed the "publish" button to make the project go live. Two hours and two minutes later, all remaining 255 pieces had sold.
Each piece sold for 0.128Ξ (Ethereum), which at the time was worth $251.42 (£181.20, €207.92), for a total of $64,109.55 (£46,198.30, €53,016.35), of which Art Blocks gets 5%, and I get the remaining 95%.
On the secondary market since then the highest re-selling piece went for 1.3Ξ equivalent to $2,133.82 (£1,537.96, €1,764.42). From the secondary market so far, I've received a further 1.314Ξ, which was $2,928.71 (£2,110.51, €2,422.51).
Giving a grand total today of $67,038.26 / £48,308.81 / €55,438.86.
All these figures were correct at the time of writing. Because this is all magic sky money, a.k.a. cryptocurrency, the figures above will be changing on a minute by minute basis.
Before I get into the details of the project, it's worth saying what they are and what they aren't.
What they are
They are NFTs, which I've written about here: "NFTs and crypto-art, from a practising artist, who thinks, on balance that (ultimately) it's a good idea.. The short version is, they are digital images, like gifs, pngs, svgs and jpgs (although they could be many other things, music, video, poems etc.), and the NFT is like a certificate of ownership saying that person X owns artwork Y.
Owning a unique copy of a digital thing may sound like an oxymoron, but that's all explained in the post I linked to above
In most reports about NFTs, the articles are accompanied by brash, harsh images from some meme hell hole pits (I'm being unnecessarily judgey here for effect). But I feel we're starting to move beyond that.
What they aren't
A lot of the conversation around NFTs can be simplified if we just pretend that a digital artwork is the same as a physical artwork and ignore the differences for the moment.
For the 70s Pop project, I didn't create 256 artworks first and then let people look at them all and pick which one they want to buy.
Instead, the artwork is created after one of them is purchased.
The buyer may have seen the artist's previous work, and they may have seen examples of what the artwork could be like, but they don't know exactly how the final piece will look.
How weird this sounds depends on two things.
"I'm going to paint a series of works based on my Stripes/Spots work, and there will be 100 pieces in the collectionm would you like to buy one now?".
Then I expect the prospect of buying an artwork sight unseen is less daunting because you already know roughly what it's going to look like.
The second thing is how familiar you are with the field of Generative Art (GA).
Which is artwork created, generally by computer, by a set of sometimes simple, other times complex rules.
I've written a post that explains GA in more depth here: Generative Art with Art Blocks, a new edge.
But, for the moment...
A quick introduction to Generative Art.
I'm going to use the phrase " Procedural and Generative Art " interchangeably, even though there are important differences.
This may be going back too far in time, but in the 80s, there was a computer game called Elite, quoting the wiki page.
The Elite universe contains eight galaxies, each with 256 planets to explore. Due to the limited capabilities of 8-bit computers, these worlds are procedurally generated. A single seed number is run through a fixed algorithm the appropriate number of times and creates a sequence of numbers determining each planet's complete composition (position in the galaxy, prices of commodities, and name and local details; text strings are chosen numerically from a lookup table and assembled to produce unique descriptions, such as a planet with "carnivorous arts graduates"). This means that no extra memory is needed to store the characteristics of each planet, yet each is unique and has fixed properties. Each galaxy is also procedurally generated from the first.
However, the use of procedural generation created a few problems. Braben and Bell checked that none of the system names were profane - removing an entire galaxy after finding a planet named "Arse".
From a single seed number, whole galaxies were created. If a different original seed number were chosen, then the planets would have other names, prices for goods, etc.
Bringing this to modern times and a game like Minecraft, the same is true; the worlds created are a lot more complex, and the amount of code used to make them magnitudes greater, but the idea of creating it all from the same seed is still there.
With Minecraft, if you pick a starting seed, which can be a phrase, string of numbers or any random gibberish, and then give that seed to me, we will both start in the identical world. As we both head north, the same mountains, valleys, lakes, and oceans will be found.
These things are said to be deterministic.
I must stress at this point though, that just because something is procedural doesn't mean the results have to be simple. You could have a computer program chock full of algorithms that you put a single seed into, and it then spends the next 72 hours working out everything down to the smallest pebble and leaf, and then 48 hours rendering the result.
Generative art works in a very similar fashion; instead of creating a galaxy or a world, it creates an artwork. Being deterministic means that if you and I both feed the same starting seed into the algorithm, we'll get the same artwork out. If we both use a different seed, then the results will be different, but within the parameters defined by the algorithm.
One way to see this is to pop over to the main page of the 70s Pop website and have a play with the "Pop-o-matic". As you watch it change, or as you hit the "Random" button, you'll see both the image change and the "Hash" change in the URL, and below the design, the "hash" looks like this...
0x52e87025336a ... 42447e15323368
...which gives us this image...
That you can recreate any time by following this link.
Deterministic, repeatable, shareable and also scalable, which we'll come onto in a moment.
Keeping things on the blockchain
Back once more to different "types" of NFT art.
As I mentioned above, you could have some procedural art that took days to compute and render. That's the type of art that you would create first and then put onto one of the NFT market places to sell.
But, people do have concerns about this. One that keeps getting discussed is that you are buying an NTF (certificate of ownership) that says that you over here owns that artwork over there. However, the artwork the NFT is pointing to could cease to exist over there, in which case, what good is your certificate of ownership?
Art Blocks get around this by putting the code that generates the artwork onto the blockchain itself.
Or rather, I, as the artist, use the Art Blocks platform to burn the code into the blockchain, an act that cost me around $1,840.85 to do. This is the identical Ethereum blockchain that keeps track of your buying of the artwork.
This means that your record of buying the artwork, your transaction "hash" for buying the artwork, the "hash" you get in return for that transaction, and the code you put that hash into to generate the artwork all live for as long as the blockchain exists.
The artwork can't go anywhere because it can always be recreated at any time.
Back to the artwork.
Now, if you have a look at the projects on Art Blocks, you can see that although there's a range of them, there is a similarity in style. One of the factors behind this is that all the artwork is created in real-time, so gloriously rendered scenes are out.
Another is that the code has to be burnt to the blockchain, which costs actual money; every character of code costs $$$. Meaning everyone is keen to keep the code small.
So rather than the Minecraft worlds of endless hills and valleys, we have the 80s worlds of Elite.
I'm not saying that is a bad thing, and several of the artists have pushed hard in various ways to move beyond the limitations of the constraints — just an explanation as to why some styles are more apparent while others are missing.
But, once again, referencing Riley and Hirst in this post, simple rules based work is already firmly grounded into the art world.
Tying all the above up, the selling and buying process goes something like this.
The artist, if they are allowed, creates a project on the Art Blocks platform. This involves deciding how many art pieces will be in the collection, the price of each piece, and lots and lots of testing.
Before the project "drops", the artist will press a button to create the first piece in the set; this will be #0. My number zero is here: https://artblocks.io/token/46000000 (all 256 pieces are in the gallery). Later the artist will make the project go live, and other people can then buy an artwork.
When they purchase one of the artworks, the transaction takes place, which takes time to be processed through the blockchain. Once compleated, a token "hash" is created, fed into the code (pulled from the blockchain), and the final piece is created with ownership given to the buyer. The creation of a work is called "minting".
People can see all the previously minted works in the collection. If the project hasn't sold out, they can choose to "mint" a new artwork. Or, if they want, they can look at already minted artworks and see if any are for sale on the secondary market. These are the 70s Pop Series One works that are currently for sale on OpenSea.
If the collection has sold out, then the only way to get hold of one is on the secondary market. Even though, as seen on the Pop-o-matic billions of combinations are possible, only 256 offical artworks exist.
Now onto the actual artwork.
The 70s Pop Series is based on my previous pen plotter artwork...
...which is in turn based on Truchet tiles. The beauty of Truchet tiles is how the simplicity can build up to complex patterns. For example, all of the 70s Pop designs are based on just two tiles.
The magic comes in how they are placed.
When I first knew I would be working on the Art Blocks platform and choosing what to do, I turned to all the work I'd been doing on Instagram for the pen plotter. The 70s Pop pen plots have always been some of my most popular posts; if I review my top posts, they are there at the top.
To sound non-"arty" for a moment, I already had business intelligence that people liked this design. I can guess at various reasons why that should be, but I also didn't need to. I knew even before starting the project that it would most likely be successful; I had two years of previous daily posts on Instagram and feedback to tap into.
Being slightly defensive, I sometimes see NFTs being referred to as a gold-rush with any old crap thrown in there. While this is undoubtedly true, the one thing that all successful Art Blocks projects have behind them is that the artists have been working in this Generative Art field a time.
While due to the constraints of the platform and technology, the results can often look simplistic, and GA may seem like an easy way to make a lot of art, and therefore NFTs as an easy way to make lots of money. That always comes through; you can see (and collectors can see) the difference between someone attempting a cash grab and someone who knows what they're doing.
(End of slightly defensive bit 😆)
Painting with probability
One thing that historically, Generative Art has had as part of its process is a curation step. You plug values into your code, look at all the results, reject the bad ones and keep all the good ones.
In this instance, you don't get that luxury. If you're going to produce a collection of 1,000 pieces, you want all of them to be good. All the curation happens at the start.
At its heart, 70s Pop is two tiles, repeatedly (and deterministically) pseudorandomly placed and rotated. But that wouldn't be interesting.
Instead, there are several ways those "random" placements can be influenced. Here is a selection of tiles where the curved ones are biased towards the middle...
...meanwhile, here's three that have the format of "middle" but are also reversed...
Others are biased from top to bottom.
There are other features too, colour palettes, number of tiles and lines, decorative elements, patterns, etc. The whole list is on the Rarity page.
For each of these elements, I had to consider the chances of them showing up. My role as an artist is not only to write the code but also to try and shape the collection.
There's a couple of considerations to take into account, much like working with any other artistic material.
Wide vs Narrow
One of the "grains" of the material you have to work with is the total number of works that are going to be in the final collection.
For example, I had a lot of plans for what to put into the collection; features, special cases, and so on, but I decided to limit the collection to 256 pieces. If I wanted all the different things to show up, I'd need to set the probability of them reasonably high to make sure they all appeared.
But with "only" 256 pieces having them all have so many different things going on starts to make the whole thing look too disparate.
Something I've seen happen in other projects, mainly ones with more artworks in, is people like to start grouping them based on similar looks and traits. Collectors picking up several pieces to put together is quite common, so hitting a sweet spot is quite important.
The reverse is also true, a collection with a thousand pieces in, but only a small number of features won't have enough variety.
Matching the number of features and variety to the size of the collection is one crucial thing. In the end, I decided that instead of having an extensive collection with all the things in, I'd split it up and start with "Series One" with a handful of the features and save other features for "Series Two" (and three, but no more).
One quick example of deciding on probability values was the "Doubled" feature. While other features had simplier base percentages (which got more complex because of certain combinations) like 5%, 10%, 20% and so on, the "Doubled" feature has a percentage chance of turning up of 2.34375%
2.34375% of 256 is exactly 6.
I wanted that feature to ideally show up six times in the whole collection. In the end, it turned up seven times.
Flat vs 'spikey' rarity
The second grain to work with is how you distribute that rarity.
For instance, say you have ten colour palettes to choose from, and you give each one a 10% chance to be selected; that's fine.
Then you have another feature with five options, and you give each of those a 20% chance of showing up — the third option with 20 different values and 5% for each of them appearing. You get the idea.
Each artwork will be statistically unlikely to appear. A single piece with a particular colour palette, feature and other feature to occur would be 5% of 20% of 10%, a 1% chance of appearing overall. Throw in a few more features, and this can quickly shrink down to 0.01%.
But, even though that combination of features has a super low chance of showing up, it has the same super-low chance of showing up as all the other combination.
The artworks don't need to vary in statistical probability, but it's certainly more interesting if they do. There's something magical about discovering you have one of only a handful of a particular type.
5% of the time, one of the 70s Pop tiles will have a pattern appear, but which pattern shows up is controlled by the following code...
pattern = 'scales' if (patternChance < 0.76) pattern = 'feather' if (patternChance < 0.55) pattern = 'flow' if (patternChance < 0.38) pattern = 'weave' if (patternChance < 0.24) pattern = 'tiles' if (patternChance < 0.13) pattern = 'zig' if (patternChance < 0.05) pattern = 'zag'
The chances of the different types of patterns decreases as you go down the list, the last one only has a 0.64% chance of appearing in 256 pieces. Turns out that zag pattern appeared just once.
In 70s Pop the colour palette is "flat", while the pattern is skewed down to one end, the "Phased" feature is spiked up at the other and affected by the palette. While the number of tiles and lines are on a bell curve, tending towards the middle values. This interlinking and overlapping variety is (I think) more interesting than lots of different features all having 10% chance of being picked out of their group.
The collection as a whole
I think taking all the things above into consideration is what made 70s Pop one of the more successful Art Blocks "Factory" drops.
Working out what was a good starting point based on knowing it was already a popular design. Choosing the size and price of the collection based on how well previous "drops" and projects had gone. Adjusting the number of features to match the size of the collection and then carefully shaping the probability of those features to create an uneven spread of "rarity".
There are a few more things to throw into the pot, such as surprise "hidden" features, and how they work aesthetically (obviously) for people to display in online galleries. I also made mine intentionally tile well so people could use them as textures.
People also seem to love using the artwork they bought as their avatar; mine don't work on that front, but I've seen others that do, and it's great.
I think that just about covers it as an introduction to 70s Pop; if you have more questions, hit me up.
I'm planning on making a quick roadmap for where this is going, which I'll link here. But until I've done that, it's basically.
- Series One: 256 pieces @ 0.128Ξ - sold out.
- 1.5 "70s Pop Super Fun Summertime Bonus Pack": 64 pieces @ 0.064Ξ, sweet summer vibes, all primary sales for charity - coming soon.
- Series Two: 256 pieces @ 0.128Ξ, new features and palettes (old features and palettes are retired).
- 2.5 "Revolution": 64 pieces @ 0.064Ξ Bauhaus all the way, all primary sales for charity.
- Series Three: 256 pieces @ 0.128Ξ, new features and palettes (old features and palettes are retired).
- 3.5 "A series of tubes": 64 pieces @ 0.064Ξ all primary sales for charity.
For a total of 960 pieces spread across three series and three bonus charity "packs".
The best way to keep track of when those will drop is by following me on Twitter @revdancatt.
One last technical note
This isn't a technical guide for making project drops for Art Blocks (although I will be doing that). But I did want to mention one thing.
If you do a view source on this page: https://api.artblocks.io/generator/46000000 you
can see that it's a basic HTML template, with the "hash" placed at the top, minimal CSS, one
<canvas> element and then a
The code burnt onto the blockchain gets injected into the
<script> node. The size of
<canvas> is based on the browser size.
Here on the 70s Pop site, in the Pop-o-matic, you can submit the hash from the above page, and you'll get a far higher resolution image back out that you can download, 7,200 by 7,200 pixels (read my Printing & Framing guide for more information about that).
As mentioned before, as long as the blockchain exists, so does your token and the code to put it into. Browers will, of course, keep evolving and moving forwards and will also find their ways into other hardware and software.
At some point, browsers may become obsolete, but I can almost guarantee emulation will cover whatever was happening in 2021. As long as something can pretend to be 2021 browser software, the artwork exists.
And as resolutions and pixel densities increase, so will the size and quality of the artwork. Art Blocks code is designed to be deterministic and resolution agnostic. You want your artwork big; then, you can get your artwork big. 😆