Validator Reputation & the Future of Restaking: A Crypto-Additive Security Model

Transcript

Of course. Hey everyone. Uh, thanks for being here. Um, I am an independent consultant. I work with different protocols on different problem statements mostly on governance and interoperability. Um today's talk is going to focus a lot more on a mechanism like one of the subsets of governance called as reputation but applied to a very specific group which is node operators validators and um it jumps into a lot of numbers. So we'll see like different ecosystems, see how decentralized they are. Uh we'll define decentralization very specifically and hopefully I will

get all of you to participate with me a little bit. So I will show you some numbers and then I want you to start guessing what you know how many rewards each node operator could be holding and we'll see how good you are. Uh but yeah, I'd love if all of you can interact and just shout back at me. Uh we are Does it work? Sorry. Okay, I'll be back in a moment.

Thanks for the patience everyone. You want me to give you Okay, we are trying to sort it out.

Okay, maybe what we can do is while this sorts out, I can try and get started. Uh, does that work? It's perhaps for the good that we don't have visuals. So, I get everyone's attention. Um, okay, let's start. My first slide actually reads something that I don't think everyone here is going to agree with, and that's fine. It says, "Decentralization is dead. Everything we intended to do, we failed at that not everything has to be decentralized. I also know that decentralization can be very tricky, very expensive and perhaps very subjective.

Uh with that in the next 10 minutes I actually planned to take you folks through three Star Wars the the famous triology. It starts act one starts with the very famous a new hope. Uh, act two in the Star Wars triology takes us through any guesses. Are there any Star Wars fans here? Yeah, it's essentially the Empire Strikes Back. Yes. And the third one is the JD is back. Uh, and so what we'll do is you'll see that theme praying up playing out a new hope. Like think about it. Why are we all here in the ecosystem

when we have to go out and defend blockchain? Some are some of our arguments are about how decentralized is it? Um when I watched Vitalik's movie, the first screen that came was how they are these massive entities and centralized forces and how it was supposed to be the new era of internet. It brought a lot of hope where you didn't need these centralized powerful middlemen to connect service providers and users. you just have a massive peer-to-peer network that could, you know, be decentralized, could have smart contracts, could work upon rules, and could facilitate large transactions. And we weren't being controlled by centralized dopamine hits all the time over time. What do you think happened to

that? At the very core of our blockchain networks, the node operators, do you think they're sufficiently decentralized or are we recreating the same system that we had earlier? So, let's actually get into the Empire Strikes Back. Uh, let's take five foundations. I wish I could show you some graphs, but nonetheless, oh, great. We have it. Um, amazing. The wish comes true. Uh, let me go back this one.

Does it work? I don't see it changing. Sorry. Okay. Um, going to jump right into the Empire Strikes Back because that's where we were. Let's take couple of random but very important ecosystems. Let's start with Celestia, right? There's four analysis that we did. Uh number one is a parto principle. What percentage of validators hold 80% of rewards? The second analysis we did is normalized rewards. So we said

how much does the lowest 50% hold? So if I had to put them all in rank, what would that look like? We did a dynamic mo mobility which means you're not just looking at static numbers. You're also seeing over the one year. So I've taken the data of 2024 to 2025. Over that one year, do validators at the lowest rank actually move up top? Is there mobility in the network? And then fourth we saw some classic economic inequality indicators which is there's the Lorens curve, the Ginny coefficient. Do they also indicate some form of economic inequality?

Let's look at some of the numbers. Parto apparently in Celestia 27% of validators end up holding 80% of rewards. What about the lowest 50%. Unfortunately, they just hold 8% of total rewards. In fact, the top validator go captures an extraordinary 27.25 of n all normalized rewards. Now, rewind back a little bit. We started with a new hope. We said we wanted to replace these massive centralized infrastructures and replace them with a more decentralized internet. That's that was the promise of the new internet. So how is it that this is better when fewer validators end up holding such a massive

power over the network? uh we go into checking the economic indicators over time and we see that the ideal the orange line that you see it's the ideal curve right that's what a normal system would probably look like a more equal world would look like but the yellow line that kind of sweeps through down to up is still unfortunately the state of inequality today right uh the genie coefficient being 0.7 so the closer to one it is the more kind of realities you See um sorry let me just pull up something as we move um maybe and then we have the quintile transition matrix. So that matrix is essentially

kind of showing you how validators are moving from one quintile to another over the one year and you'll see that there's barely any mobility. If there is one that's in the very top 20% they continue to remain stagnant throughout the year. There's hardly any movement. Uh this is what we call as the sticky wealth where you continue to dominate your position of power for a long period of time in the network. Can it get any worse? What if we look at polygon? Uh what do you think is the state of art there? So we did the same analysis again just anchoring back to everybody. Can I get some numbers? What do you think are like if you had to answer parto what percentage of validators at Polygon do

you think are holding 80% of rewards? Random guesses. Just throw a number. Okay, 10 19 let's see. Well, it's 16% uh somewhere in the middle. They end up holding 80% of rewards. And what about the lowest 50%? It's absolutely like they're living off 2% of total rewards. Now let me clarify some definition here. What do I mean by rewards? So this is normalized rewards. So you get block rewards for producing blocks. You get attestation rewards for signing checkpoints. It's a combination of both. But when I say it's normalized, it means that it's calculated over the period of active time. So let's say you

joined one year ago, I join now. Of course, I'm going to have lower rewards than you. So we normalize the rewards based on your active epochs in order to be able to compare apples to apples. Um, go on further. We see six validators hold half of the entire network rewards over time. What do you think about mobility? Do you think the lower validators are able to climb up over time? Let's see. Um, the mobility index is 9.4%. Which means yearon year a maximum of 9% validators tend to move up. Uh, and when I say move up, move up by just one quarantine. So you're not going to see the body bottom 50 directly go right up. So sticky wealth uh genie

coefficient being just as bad. Laurens curve showing similar patterns of economic inequality. U and the bright line that you see that's an indicator of people staying in their quantile and not moving up a lot. U and that bright line is a thing to worry because maybe we start systems with centralization. The promise was progressive decentralization. I don't think that's happening. What about dimension? Now, uh we have few few examples so bear with me. What do you think? Throw numbers at me. What percentage of validators do you think are holding 80% of total rewards? Any guesses?

Okay. Can get worse. Well, it's still at the you're seeing a very clear clear parto seeing 10 to 20% validators holding over 80% of rewards. There is an example where it gets worse. So, hint hint. Dimension has 12.2 validate percent of validators. The lowest 50 hold 2.5%. Again, really, really low. Shockingly, two validators hold nearly 60% of all rewards and 10% command three quarters.

Um when you design governance and you give voting power based on how much stake they hold, you have to be careful because this is where uh giving one token one token staked is equal to one voting power can get dangerous really fast, right? Um same economic indicators show us really bad uh gen coefficients. The bright line in that quantile transition matrix, if you remember, shows us that people tend to stay where they are. Finally, the final example. I'm not going to bore you guys too much, but um here it is extremely shockingly low.

7.1% of validators hold 80% of rewards. The lowest 50 literally hold 0.5% of their normalized rewards. Um the top 10% validators in Kronos hold 90% of all rewards. I don't know if it can get more centralized or dangerous. And this exactly brings me back to the question, what are we really doing? If our goal was to create the new internet to replace these centralized entities, have we not now created a node system, operating system that replicates the same traditional system in our real world. Um, so what is the solution? I think we all understand the problem. Uh,

this is where the return of the Jedi, uh, we get into the third part, the final part where We're researching something called as validator reputation score. Like the term suggests, it's a score that depends on meritocracy, right? It looks at three things. It looks at your attestation effectiveness score. It looks at your proposal effectiveness score. And it looks at your social effectiveness score. Now, think of these as Lego blocks. This is not a one-sizefitit all, right? You could say, I'm going to give 70% weightage to block production because the block rewards are much higher than at station. And you could kind of give weightages of your choice. Social effectiveness score can be broken down into parameters that matter to you as a

network. So if governance is important to you, that would be one of the social effectiveness score that you could incorporate. If diversity and reducing single point of failures is important to you, that could be your social effectiveness score. The goal is empower, recognize, and reward validators in a fair in a way that's fair, meritocracydriven, and accessible. Um these are the formulas. Please let me know if you have any questions towards the end. I'm not going to get into the the thing of it. But it can be used in three big applications. Number one, reward distribution. So we all need stake. Why do we need stake? Because if there's any form of security concerns, slashing can take place. So stake becomes security. But should stake be

privilege where if you have more stake, you get more rewards thereby ending up dominating the network? Perhaps not. So reward distribution can be tied to reputation. Stake can exist for network security. Uh it can be used for validated admission. I have something very exciting to show you here. And third, it can be used for shared security orchestration. I'm sure you a lot of you have been hearing about shared security restaking validator sharing. How do you do that? How do you decide which validator secures which network? Why not use something fair like a reputation in that system?

This has this been tried? Well, it has in some sense. Um, in Polygon, we had a validator admission framework that was stake only, which means that if you met the minimum stake criteria, you'd get a validator slot. We moved that to stake and reputation. So, while there is a minimum stake, which is not very inaccessible, a lot of your ad admission criteria now below is kind of determined based on your kind of reputation. uh we use experience score, expertise score. We look at your onchain data, see what you've how you've kind of validated in previous networks and then determine it.

What we found over 12 nodes were onboarded using this admission framework with zero performance issues and high network health. So I'm not preaching this for the philosophy of it. I'm preaching it because it actually works for the health of the network. um in a veil. We're now bringing this with AIP4. This is called as aail nomination program where we're trying to tie reputation with rewards. So the more reputation you have, the more opportunities you get, the more rewards you end up making if you perform well in those opportunities.

Finally, I started off with a very big statement. I don't think it's completely dead. I think decentralization is waiting. It depends on us on how much we want to act how long-term we want to think about the networks and the rebellion lives and it lives within us. Yeah. So let's work to bring meaningful decentralization to every network and happy to answer any questions. Thank you Tanisha. That's a