Aesthetic Rankings for Doodles

#1

Doodles are beautiful

The thing that first drew me to Doodles was how freakin’ good they look. They’re distinct, inclusive, fun, and really cute. Within the collection too, there are some wonderful combinations of traits that produce really unique outcomes - I’m personally hoping to one day pickup a Coffee Head + Sunglasses Doodle, as that combination just speaks to me.

In the NFT space though, the prevailing platform is OpenSea, and the prevailing metric to help guide people in their purchases are rarity-based solutions. The former doesn’t allow for “custom” trait searches (such as searching for all Doodles that have 2 traits of the same colour). The latter takes the outcome out of the hands of the community and gives it to a statistical calculation instead. Additionally, it doesn’t allow us to recognise some of the beautiful coincidences that arise from a generative art project.

Looking at the #doodle-market channel, I see that many people are using rarity when trying to market their Doodles. We believe that while rarity will always be valuable knowledge when deciding what to buy, that aesthetics should be the primary way to judge what is an aesthetic entity. As such, the existence of an aesthetics-based ranking system for Doodles I believe would help people find the Doodle that speaks to them, and in addition the ranking game itself would provide for a really fun form of community engagement.

Proposal

Myself and another Doodle holder have been building an aesthetics-based ranking tool, https://poprank.io , and I really wanna get Doodles on board. This proposal is to ensure that both the Doodles team and the community are on board with the idea that any and all questions can be publicly discussed.

The tool is already live (for other collections), with trait filtering being implemented by the time this proposal is voted on, if it’s accepted by the community. As such, Doodles could have everything outlined below instantly upon the acceptance of this proposal.

Unique Doodle traits

As part of this proposal, our ask would be for the Doodles team, and the community lay out any custom “meta” traits that they’d want to be searchable. For example “2 matching colours” would show any and all Doodles that have 2 traits with the same colour. I think that the ability to search for these custom traits that aren’t necessarily defined in the metadata will help people find their forever Doodle.

How it works

There’s a head-to-head game where 2 Doodles would show up, and you’d pick your favourite

image
image1920×962 58.3 KB

After enough rounds have been played, there will be publicly visible rankings based on the rounds that have been played so far - these will continue to update as more rounds are played, and are based on the Elo rating system.

image
image1920×875 155 KB

This is an example of “meta” traits too. I’m really interested in having both the team, and members of the Doodle community define what “meta” traits they would be interested in seeing.

image
image1920×1055 146 KB

The team

The two developers are myself ibrosen (Doodles #6540, #6140) and shando (Doodle #1626). We’re both engineers and are responsible for every line of code that’s gone into the tool thus far!

Future

There will be a Discord bot where the rankings are queryable + the game can be played from Discord, as well as an embeddable version of the game that the Doodles team could put onto any site, if they so chose.

Challenges

Everything outlined in the proposal so far, other than filtering by traits (which will be live in a day or two), is already up and running. Getting the Doodles on board is more a matter of having this proposal be discussed and voted on, to ensure that it aligns with the Doodle’s vision.

The one technical challenge that exists is that of data integrity. What’s to stop a nefarious user from scripting our site and sending their assets to the top?

This is one of our core concerns, as the rankings are only as valuable as the data is trustworthy. We’ve already invested a lot of time into solving this problem. At a high level, only rounds from people with connected wallets will count. Having a wallet connected allows us to uniquely identify the rounds a user has played, which we use to run additional checks that ensure that a user is playing rounds in good faith. I can go into more detail with this with the team privately, as we believe there to be some security in obscurity and don’t want to publicly detail our security measures.

This is our main real challenge, and one we’re actively working on.

Conclusion

I believe that NFTs as we know them today are defined by their aesthetics. It’s our judgements, opinions, and connections to NFTs that should be the most important factors when making decisions on what to buy. Furthermore, Doodles as an NFT collection with a unique and incredible aesthetic, amplify the need for a tool that focuses on this, instead of rarity.

I’m keen to hear whether the community is interested in something like this, and what the team thinks about it!

9 Likes
#2

Really like this idea a lot. Seems like a fun way for our doodles to have a little competition!

3 Likes
#3

Thank you! Yeah if the team / community are on board, we could have them up and running in just a couple of days!

1 Like
#4

+1, I like this. Would bring exposure to otherwise low-rarity doods

2 Likes
#5

I like this proposal. Most people pick their doodle on aesthetic rather than rarity.

2 Likes
#6

I really liked this idea - in fact, I also tried to implement this algorithm but failed on this question: how do you ensure fairness that all NFTs in a collection get seen/rate equally (in an equal amount of times)? For example, if a project has 10k NFTs, one user might only be able to rank 50 of them, if 100 users participate in it, there are only 5000 different NFTs been seen.

#7

This is interesting, and could eventually be used for any other project. Just plug in the collection and then let people do the sorting.

Ranking by aesthetics may eventually be more prevalent than by rarity, or a combination of the two.

Aesthetics rankings may evolve over time as tastes change over time.
Executional consideration - This ranking methodology may be vulnerable to manipulation. I’m sure that it is solvable.

Thank you for the interesting proposal!

1 Like
#8

Since all our avatar data is public data (OS api is free), do you need permission to set it up? You can share with us the link to the doodles collection and we can have a go at the ranking game.

Or is it more of getting legitimacy/stamp of approval? (does it matter as much since there will be official rarity tools)

Seems like a fun thing to do, browsing doodles :slight_smile:

#9

It’s an interesting one! So there are two parts to how we solve this.

One is that when serving a player with 2 NFTs, we ensure that we give them NFTs that have been roughly seen the least, so that if 10,000 rounds are played, we can expect that each NFT has been seen around 2 times (1 round = 2 NFTs being seen).

Secondly, we only reveal our rankings after a certain amount of rounds have been played. We’re still playing with this parameter, but at the moment we have it such that the number of rounds that must be played for us to reveal rankings is 7.5x the number of NFTs in the collection. This means that each NFT would have been seen, on average, 15 times by the times rankings reveal.

It’s a fair few rounds needed. but when divided by hundreds of users, it becomes a pretty manageable amount!

#10

Yeah the manipulation part is the biggest challenge, and will be an ever evolving challenge. We have measures in place already, but will continue to improve them. We also have the ability to “rollback” rankings, to basically replay all the rounds and rebuild the rankings, if we find that there has been manipulation.

Thank you for checking it out!

#11

We dont need permission per se, but I wanted to go through the proposal process firstly to make sure it aligns with your vision. For example, I spoke to Creature world about this a week or two ago, and they explained that for their NFTs, they aren’t a big fan of rankings of any kind, so I’m glad I checked.

It sounds like you’re keen on the idea though, and the community too has expressed interest in it, so I can set it up tonight and send you a link tomorrow! What’s the best medium via which to send you the link once it’s ready?

#12

For clarity, I’m not one of the founders nor team members, just a fan and a fellow dood. So let’s speak as fellow community members :slight_smile:

As a doodler, I think it’ll be a fun game and that reflects the way I’ll treat the aesthetics rank too. Just for fun.

Any ranking system, even aesthetics-based has its downsides. Some might feel bad if they find out that their moustache + long hair + skelly avatar ranks low on looks. Which would mean there can never truly be an official ranking system based on aesthetics unless it was the intention of the project from the onset to be a project about that.

In doodles, what matters is people vibe with and love their doods in their own special way. I think if this ranking is seen in that way - a game, it’ll be fine. Pushing for people to place importance on ranking risks breaking the culture (inclusiveness, everyone’s dood is awesome). Notice that few, if any, ever discuss rank in the discord.

To the non-doodlers, it maybe helpful for them to reduce purchase dissonance though I think its a non-problem in most cases. As you said, Doodles look freakin amazing :smiley:

Just my 2-cents.

#13

Ah I appreciate the clarification!

I fully agree re: rankings should never take precedence over an individual’s connection with their NFT. Is that not also a criticism against rarity-based rankings though? That being said it’s a very valid point, and that’s I think where the benefits of the game outweigh the drawbacks of a ranking system, as the game will serve to increase one’s connectedness with the Doodles + help them find one that truly speaks to them.

I’ll set the doodles up soon and post it in the #think-tank chat. Thank you for your input so far, I really appreciate it!

For any founders reading the thread too, if at any point you have any questions / qualms / want to officially support aesthetic rankings, please reach out to me!

#14

Yes, in the case of rarity-based systems, Doodles seem to defy that convention. Oh boy do people love their traits! I remember there was talk about whether to have trait based #channels. Trait tribalism.

Poprank looks like a professional project. Kudos on the slickness.
If you’re building this as a startup, I suggest taking some time to better define the problem you’re that solving and get more validation from potential users.

If you’re taking the same approach, see if you can build it without requiring people to rank in order to give utility to users of the ranking. Crowdsourcing models are notoriously difficult to find scale and it looks like you might try a token incentive system to solve that. Be careful in adding new layers of solutions to solve the weaknesses in the model.

Omg, I sound preachy. :sweat_smile:

1 Like
#15

Yes! Love this competition

1 Like
#16

Great idea. It has worked very well for the z-huge collection mountvitruvius.art

1 Like
#17

Thank you everyone for the feedback! I’ve added them to the site and I’ll reach out to the team to get their feedback now, it feels like the best way to move the conversation forward.

PopRank

#18

Still looking for the OK from the lead team. If not, I’ll take the Doodles off the site to make sure I’m not doing something that goes against the team’s wishes.

#19

I just did a handful of rounds on the PopRank site and see so much value in this tool. The aesthetics matter. And they should be factored in the rankings! I hope more people will take a look at this proposal.

#20

Thank you for the support @dannyddg and everyone else! I’ve taken Doodles off the site for now, as I haven’t heard back from the team yet and don’t want to do something that runs counter to their vision, but I’m open for them to reach out to me!

2 Likes