Pre-proposal: 🧙 GANDALF (Decrease MaximumChains)

[shane]

Thank-you @Qspider @iannn @msa6867 and @whyclosedsource for the recent comment. Lots of interesting ideas.

@iannn

I also think that we should see an expansion of chainIDs to represent different indexes and other value add components with different to represent different qos checks or whatever, and allow the traffic per chain to distribute into those as DEMAND requires

@Qspider

We should add a new element to the relays to tokens multiplier to account for single node runners and start unwinding the tokenomics that have been precedent the last 2 years. It’s just not working. If it takes a hard fork then it takes a hard fork.

@whyclosedsource

is it possible to have ‘node pools’ similar to mining pools in the backend.

These are intriguing solutions to try and address reward imbalance, independent node running, QoS, and rare-chain burden, but they would add significant complexity to tokenomic and node running from what I currently see. The simplest solution is using the existing MaxChains parameter. Many of these ideas would take reworking of v1’s architecture to add more on-chain elements, parameters, and result in substantial complexity to understand.

@iannn

I believe this campaign is about capturing more stake per chain/per entity and not about helping indie node runners, at least that is my perspective as an indie node runner competing in this ecosystem.

You are partially misunderstanding my motives. Since May I have been publicly vocal that I believe it would be wise for POKT to make node running more attractive for the purposes of increasing buy pressure. Please see this thread on Telegram for starters. Most of POKT’s past buy pressure, from the retail side, came from folks wanting to join the node running economy (much like yourself).

To make node running approachable, I say simplify and balance out our reward ecosystem. Node runners, like yourself, frequently talk about how things are getting to a point where running nodes is a net loss operation. That doesn’t have to be the case if any users can spin up a POKT node and chain node to generate yield on hardware from home (POKT original vision and message which is now completely lost).

Stimulating growth/interest in the node ecosystem has been my message from the start, which I have been very consistent about. Even to the point of building out Burn And 🥩 Harnessing (BASH) Deflation Economic Model to try an increase node rewards, while also increasing buy pressure through more node staking. I’m still hopeful that we can evolve beyond ARR to hit those goals, but we do want to pace things out as ARR only recently passed.

@iannn

at least that is my perspective as an indie node runner competing in this ecosystem.

We also may be defining independent node runners differently. When I say independent node runner I mean folks that are staking their own POKT, or POKT of folks they are associate with. These are not folks that have a public node running businesses, like yourself, who are paid to run nodes for what would be strangers. Last year there were many VCs, POKT community members, hobbiests, etc, which were independently running their own nodes… which is mostly not the case today.

We all agree that it makes zero sense to try and run your own nodes when MaxChains is 15. However, it does become possible with something closer to 1, which opens the door for hobbiests to join our ecosystem again, and potentially bring some of that buy pressure back. Folks like yourself that joined POKT to run your own nodes is where much of the buy pressure came from back in the day, and many of those folks have stuck around and contributed in major ways to the ecosystem.

MAIN POINT

It is a net loss for POKT to alienate would be hobbiests from joining POKT because the economics are stacked in such a way that they cannot meaningfully participate, without breaking the bank, or signing away their soul The other RPC projects, competing with POKT, are seeing community growth around node running, and to me it is foolish for POKT to just do nothing when we already have the parameter in place that will greatly address current imbalances.

Most of POKT independent developer community, which has brought many contributions to the POKT ecosystem, came from folks interested in the node ecosystem. We are shooting ourselves in the foot if we let that source of talent dry up.

@iannn

I also think that the conditions you describe as optimal around 5 or 3 or whatever is something we could grow into by next year at 15 if we don’t stop making arbitrary changes that will achieve outcomes other than what’s stated, and stop jumping the gun before other things mature.

Yep, it’s been suggested a few times that this should be a progressive processes. I’m taking that to heart and am looking to make my proposal have more a progressive plan.

@iannn

Also, we already have conditions now where there are already multiple “sets of 15” if you will that could be staked as a large provider to maximize rewards, and those larger providers that will outperform will be staking fleets into multiple sets of 15 nodes. If the maxchains were reduced, these same savvy providers, some of whom built machine learning tools to do this very thing last year, will grossly outperform anyone else in the aftermath of a chain reduction.

Check out the the Dissenting opinations, where I explain how when you decrease the max chains, movement from large node runners would actually bring more balance to more chains.

For the better part of 2023, you have been the top provider, week over week, in terms of rewards generate per node. You have mastered the 15 MaxChains ecosystem, which I commend you for. Even if MaxChains is reduced, you can still run 15 chains to maximize your rewards, but it will also have a side-effect of helping balance out the network of a whole.

Final Thoughts

Through this proposal, and through private conversation this is what I’ve learned:

1. The vast majority understand POKT is imbalanced with MaxChains at 15
2. Providers are hesitant to doing any changes in v0 because many providers are non-custodial, and rotating user’s nodes to different chains will be the main challenge.
3. Providers have many systems and processes in place for the current MaxChains 15, which would require reworking for a new MaxChains. For some, this is more of a burden then others, depending on how their business is structured.

Regarding 2, I understand. How will providers choose which user will be staked on what chain? Currently the “Money Chains” provide +80% of the POKT reward per day, but with reducing MaxChains, especially to 1, providers will have to place users on specific chains and then likely put in place a rotate to provide them with as constant rewards as possible. For providers don’t use non-custodial and operate more as a pool, they will not see much change, but it does complicate things for non-custodial providers.

Possible Solutions

I’d like to see if PNF or the DAO could put out an RFP for a chain rotation tool for nodes. Providers will need to address this elephant in the room for v1 anyways, so makes sense for the ecosystem to be proactive with building solutions in the place, pre-v1.

I’d also encourage someone to take this on through a Socket, where you can simply open it and start building this out with compensation for the DAO. Sockets are permissionless to start, and only require posting a forum post (more info here), so any developer can open one up to start building this kind of tool. I see this clearly fitting under the V1 Development And Launch category of ERA.

For context, I was confused about Socket up until a week ago, so check out my comment here if you want some clarity on how they are started. I think a chain balancer tool for nodes would be amazing and could help folks in v0 and v1.

Next Steps

I do plan to make a proposal that can go to a vote. I think making adjustments in v0 to balance out rewards and re-open the door to hobbyists will help POKT compete against our competitors and attract talent to our community. It also opens up the opportunity for POKT to increase its buy pressure as the node economy becomes interesting again to outsiders. I don’t think we should justify known issue, but instead be proactive with making POKT’s economics to be the best it can be. I understand that doing too much, in too little of time, can be unhelpful, so I think a progressive path is best way forward.

Thanks for the feedback

[iannn (Ian)]

We do not all agree that it doesn’t make sense to run because of the 15 chains, and just because you place that in your final thoughts as bulleted point, and frame it was “we all agree”, doesn’t make it true.

Spin it how you want to, but increasing min stake and reducing max chains are analogous, and this is a lot of circumlocution and hand waving to try to distract from that fact - this is all about duping voters into increasing the ratio of stake to rewards, ie just trying to capture more stake per pokt minted.

I came in as an indie with my own stake, and I have never been able to make money by running nodes for myself let alone for other people (but not because of 15 max), which I have only done trying to chase expenses to keep doing what I’m passionate about, please don’t try to characterize me as a capitalized business profiting off of stakers, and not an indie still just because I stubbornly refused to die and built my way to have this many nodes. I have remained consistent that people choose to trust me to stake. But I’m still an individual node runner, no capital no employees no outsourcing.

I rode my stake from $1 to the atl every time and still haven’t unstaked. Some entities in this discussion cut and dumped their entire sizeable stake before SOTU last year if I’m not mistaken. Whatever the circumstances don’t matter, not my business, only bringing it up because my perspective is much different as one who rode the cliff all the way down. There are others with diamond hands like me, and I hope that they look at a proposal like this with sadness in their hearts like I do.

We need to increase demand and increase value-add, not this, this isn’t going to help that one bit other than a slapdash attempt to capture more stake per pokt minted. Those who look at pokt and avoid it are going to look at this as more of the same old tricks. It’s going to hurt pokt’s image as bad as anything else has so far.

I’m not arguing against in fear that it’s a cumbersome change to reoptimize around, I will surely adapt even if it passes, even if it takes me time. Though having been around since last year I do know of several entities who claim to have already built the tools internally to satisfy a large fleet that needs to optimize staking across multiple sets of chains. Some entities aren’t going to have to adapt to this change, and those will be the ones who benefit most, not indies.

[shane]

@iannn

this is all about duping voters into increasing the ratio of stake to rewards

Can you please show how this is at all what will happen if MaxChains is reduced? This makes no sense as POKT still has the same amount of rewards per relay.

Each chain still has the same amount of relays from customers, and there are still the same amount of nodes on the network. All that would be different is to get network average you run less chain nodes per stake, which means you service more relays on the fewer chains.

From the Spreadsheet, which is data from POKTScan, Nodes were serving this many relays per day (Column J), and generating this much reward per day (Column M):

When you set the MaxChains to 3, nodes service MORE relays per chain, and generate more rewards per chain

Nothing in GANDALF, or MaxChains, is changing “the ratio of stake to rewards”, and suggesting it would is objectively false, unless you provide some kind of reasoning. I’ve gone through great lengths to show the opposite is actually true since you can decrease infra costs, while still maintaining the same amount of stake.

I’m open to anyone providing clarity here.

@iannn

this isn’t going to help that one bit other than a slapdash attempt to capture more stake per pokt minted

Again, this is objectly false unless you explain to us how the MaxChains parameter at all requires more stake per POKT minted.

[Qspider]

@shane

Can you please show how this is at all what will happen if MaxChains is reduced?

Let’s say I’m an independent node runner and I’m just running 8 chains on my home infra. An Eth archival node and a few smaller chains then I’m using a pool for the others and they’re using my chains. All of us have to stake new POKT nodes for every chain you reduce because we’ve all been used to this ecosystem and it brings the community together. Limiting the chains has damaged the ecosystem. If max chains is reduced each of us have to unstake earn less rewards 21 days and then consolidate our stakes to account for the new limit. It’s a lengthy process and not just for big service providers it’s for anyone unless you’re 1 stake then you still have to undo your entire infra that you’ve been doing or whatever other methods you used but I think most of these little guys probably dropped off a year or 2 ago when node averages dived with the price.

Let’s say I’m completely new to the space. I only use 1 chain eth. I still have options to use other chains and participate but it may be more valuable to stake a different less heard of chain that’s not as big as eth so I have to search and make a deal with the community and we can switch chains. Not with the perfect world balance apparently. I would just stake the smallest resource chain and earn the same 4.68 POKT on a Tier 4 node ~60,000 POKT.

Regardless you’re making 4.68 POKT to run the least intensive node cause they all make the same if this whole limit Max Chains is concerned about full node costs. There is no incentive to go out in the ecosystem then and participate really as the rest is centralized somewhere because there is no incentive to run a full node unless you need it for dev. This seems like a completely different vision from when I started a couple+ years ago. Each single node runner literally has to go buy more POKT to earn anything like they were earning a few months ago. 4.68 POKT avg on 60,000 token stake. That is less than 15000 stake rewards in April, June? Am I reading that chart correctly?

When it comes to QoS and rewards and cost of node running let’s stick to the stuff we are already doing and have been doing rather than limiting Maxchains further. I would rather wait till after wPOKT has been around for a few months and see about unwinding some other policies before visiting this again. Node averages dived after SER happened. Let’s go back to SER like we changed our minds. Over 900 nodes unstaked yesterday that is following the thousands that unstaked after SER passed. 10-20% network dropped off after this stuff continues to do so with no complaints of max chains I have heard but we have liquid, minted and base all increasing because of bad policies implemented in preceding months.

[Qspider]

Excuse me I meant liquid and base increasing not to bring up mint since that needs to be changed as rewards are too low for excitement compared to how much needs to be put up for staking.

[shane]

Thanks for the reply @Qspider. This communication is great as it allows me to address some misunderstandings.

@Qspider

If max chains is reduced each of us have to unstake earn less rewards 21 days and then consolidate our stakes to account for the new limit.

This looks to be a misunderstanding. To change the number of chains a node is staked to, all that is required is a regular stake transaction. No unstaking is required.

Every node runner at some point has changed the chains they are staked with by submitting a stake transaction with the RelayChainIDs listed. If MaxChains is reduced, nodes would just submit a stake transaction that include the amount of chains that MaxChains allows. It’s a very simple transaction.

Many folks using Community Chains, have done update stake commands to change the chains they are staked on, so it’s a common practice and is no barrier for node runners.

@Qspider

Let’s say I’m completely new to the space. I only use 1 chain eth. I still have options to use other chains and participate but it may be more valuable to stake a different less heard of chain that’s not as big as eth so I have to search and make a deal with the community and we can switch chains. Not with the perfect world balance apparently. I would just stake the smallest resource chain and earn the same 4.68 POKT on a Tier 4 node ~60,000 POKT.

@Qspider

Each single node runner literally has to go buy more POKT to earn anything like they were earning a few months ago. 4.68 POKT avg on 60,000 token stake. That is less than 15000 stake rewards in April, June? Am I reading that chart correctly?

This is another misunderstanding so thanks for bringing it up. You missed that in my example above where nodes where making 4.68 POKT per day, the MaxChains was set to 3, meaning that a node runner would need to run 3 chains to make network average. In your example, you only have the node runner running 1 chains.

If we take your example and instead have the node runner runs 3 chains, since that is what MaxChains allows, then they would be making 14 POKT per day. 14 POKT per day was exactly what was network average when this data was taken from POKTScan, though the POKT node had to be serving MANY more chains since MaxChains was 15.

If we take your example of a node runner serving only one chain, when the MaxChains is set to only 1, then the node runner would generate 14 POKT per day.

@Qspider

Over 900 nodes unstaked yesterday that is following the thousands that unstaked after SER passed. 10-20% network dropped off after this stuff continues to do so with no complaints of max chains I have heard but we have liquid, minted and base all increasing because of bad policies implemented in preceding months.

I believe MaxChains being 15 is a bad policy, and I’ve shown exactly how it effects the network. People unstake because of the economics of POKT, which includes many of the things you listed, but MaxChains is objectively part of that economics as well, but it’s always been a hidden imbalance, hence this work.

We both seem to agree that POKT’s economics need to improve, and this is why I’m suggesting addressing this as part of solution (not the whole solution by any means ).

[Qspider]

@shane

This looks to be a misunderstanding. To change the number of chains a node is staked to, all that is required is a regular stake transaction. No unstaking is required.

I was referring to the existing members in the pool of full nodes wanting to utilize more of their own infra. The community. The current staking amounts in total would need to be adjusted to account for the removal of chains so you would have to buy more pokt and stake more to have a pool exist but in fact it makes it to where it just eliminates the pool ecosystem so anyone using other chains they happen to stake to it makes no sense to even run a POKT node for most of the chains. Totally removes the pool and then it becomes some kind of centralized entity of full nodes that only app developers use the portal for. In essence becoming it’s own competition with centralization.

I don’t think we need to punish the entire network because of some poor earning chains that are over provisioned. It seems like for example with it’s passing all accounts can just move to Iotex which is a light weight node and make the same as any other chain.

[shane]

@Qspider

Totally removes the pool and then it becomes some kind of centralized entity of full nodes that only app developers use the portal for. In essence becoming it’s own competition with centralization.

Sorry, but I don’t know what you mean by “the pool”. We have Community Chains, which many independent node runners use, but I’m not sure how reducing MaxChains creates any kind of centralization or “it’s own competition”.

@Qspider

eliminates the pool ecosystem

The only public pool is Community Chains. As far as others having private pools, I can’t speak to those as those business models are not relevant to POKT’s economics. POKT’s model has never been around private chain pools… it’s always been decentralized node runners.

Yes, it is possible that some folks will stop using CC to run their own nodes if MaxChains is reduced… but I don’t see how that should drive POKT’s economics decisions. It helps POKT to have more decentralization even if it reduces reliance chain pooling.

[Qspider]

@shane

We have Community Chains, which many independent node runners use, but I’m not sure how reducing MaxChains creates any kind of centralization or “it’s own competition”.

I apologize but I thought I had explained this already. What is to stop everyone from just using a low resource node if all the averages are the same? For example, service providers that manage stakes for people just stake Iotex or maybe some other chain up to 2TB.

@shane

I’m not sure how reducing MaxChains creates any kind of centralization or “it’s own competition”.

With everyone moving to less full nodes that consume resources to have less impact since they all make the same the backend will have to be centralized somewhere. Perhaps PNI or Community Chains or a combination of service providers instead of the node runner community.

@shane

The only public pool is Community Chains. As far as others having private pools, I can’t speak to those as those business models are not relevant to POKT’s economics.

They should be relevant since they are investing in POKT because it helps with web3 infra. I don’t see how you can imply Community Chains is relevant but others private or closed are not. It is all relevant to POKT economics as these are buyers and partakers that are relevant and providing infrastructure to the space whether they are service providers or single node runners and I am thinking the reason why GANDALF was created because of the cost of it all.

Whatever happens I hope this isn’t it.

[shane]

@Qspider

What is to stop everyone from just using a low resource node if all the averages are the same? For example, service providers that manage stakes for people just stake Iotex or maybe some other chain up to 2TB

In that case, if easy chains gets over saturated, then the other chains go up dramatically in reward value. It would be similar like it is today, but those running the harder nodes will likely get much more rewards then the tiny change there is now.

If I set MaxChains to 3, so each node is suppose to get 4.68 when balanced, but then apply POKT’s actual current imbalance, this shows what today’s imbalance would look like in that scenario. You can see this by going to the “Gandalf w/ Today’s Imbalance”, where it take the MaxChains for the Gandalf page and merges it with today’s node imbalance.

As you can see below, the hard chains would likely still get higher rewards as more nodes would be on the easier chains.

This is by no means to be a perfect reflection, but it shows that suddenly chains with less nodes on them can produce substantially more rewards than saturated chains. This is what incentives balance… and in areas where the free market leans toward imbalance, those on the harder chains can get more relays.

Now look at the reward difference with today’s imbalance and MaxChains at 15… the rewards are hardly effected by the imbalance with Solana and BSCA

This is why GANDALF makes sense… as it more largely rewards those that commit to rare-chains, while also reducing the amount of chains they need to run.

@Qspider

With everyone moving to less full nodes that consume resources to have less impact since they all make the same the backend will have to be centralized somewhere.

This is a misunderstanding. It’s not that there are “less full nodes” on the POKT network… GANDALF means that each node runner runs fewer chain nodes, BUT there is much more spread of the chains those three full nodes are on.

Right now, there are 8 “Money Chains” where every node runner should run to maximize their reward. This means that every node on the POKT network must support those same 8 chains. That is way to many full nodes needed for those 8 chains.

Instead, GANDALF would means that node runners run only 3 chain node, but those 3 chain nodes aren’t the same nodes as everyone else. Reducing MaxChains requires full nodes to spread out to other chains, instead of basically requiring the same 8 nodes to always be run.

It is not that there are less full nodes on POKT, it is that there is more of a spread on the chains that those full nodes are on.

@Qspider

I don’t see how you can imply Community Chains is relevant but others private or closed are not.

I never said that CC was relevant and private pools are not… that is a misunderstanding. I’m saying that POKT’s economics can’t be determined by any pooling businesses (CC included)… and POKT needs to focus on what makes good tokenomic.

When I mentioned CC, I said that we would likely lose customers. I’m proposing GANDALF even though it means that less folks may choose to use CC. We can’t predict everything that will happen with every business model in POKT, but we can in-fact, say that POKT tokenomics around chains are currently imbalanced and causing most of the ecosystem to focus on 8 chains for the majority of their rewards, which produces a large infrastructure burden for anyone wanting to the join the POKT ecosystem.

Love the memes though @Qspider Adds some fun to what would be a dry discussion!

[msa6867]

@shane

@iannn

this is all about duping voters into increasing the ratio of stake to rewards

Can you please show how this is at all what will happen if MaxChains is reduced? This makes no sense as POKT still has the same amount of rewards per relay.

OK. So perhaps @iannn is not stating the matter precisely, but the point he’s making is obvious: decreasing MaxChain while holding minstake (and the PIP22 parameters) constant is the equivalent of increasing the “pay to play” stake required per chain to gain an entry into the servicer selection array. Any speculation as to how that will affect rewards per POKT staked on a specific chain given an unknown amount of pruning of each chain’s selection array as consolidation occurs is just that - speculation.

We can confidently say the resulting change to average reward per POKT staked is bounded on the low end to the ration of new to old MaxChain. (E.g., if MaxChain is reduced from 15 to 3, then the ration of the new average reward per POKT to old average reward per POKT staked is bounded between 0.2 (=3/15) and 1 (in the ideal). I am sure you are correct that it will be closer to the higher end of this range (albeit with greater volatility as others have pointed out). Nonetheless on average it will be less than 1, not exactly 1, so that " increasing the ratio of stake to rewards" remains a true statement. The end result will be very chain specific. Meaning that for some chains, the decrease in reward per POKT staked will be negligible, while for other chains it will be significant. (For yet other chains, reward per POKT staked may even go up as too many node runners abandon the chain leaving it vulnerable to insufficient support). Averaged over all chains, however, it will be some non-negligible number less than 1.

This is in keeping with what you yourself admit to being a secondary motive. Namely to generate buy pressure for the token. Any new buy pressure that comes in the form of incentivizing supply-side actors to increase tokens staked to the supply-side is, de facto, a decrease in average reward per POKT staked.

As I stated before, I am not commenting here as to whether this is good or bad, I just don’t want to see this aspect swept under the rug.

[shane]

I have to admit @msa6867, I am not following your comment. Thanks for bring it forward though and I’m happy to address your points with some more context.

@msa6867

decreasing MaxChain while holding minstake (and the PIP22 parameters) constant is the equivalent of increasing the “pay to play” stake required per chain to gain an entry into the servicer selection array.

I do not following what you mean here. How does MaxChains effect the rewards of someone that is staked at 15k more than someone who is staked at 60k?

@msa6867

E.g., if MaxChain is reduced from 15 to 3, then the ration of the new average reward per POKT to old average reward per POKT staked is bounded between 0.2 (=3/15) and 1 (in the ideal)

How do you derive that the average reward will be between .2 and 1? Average relays per nodes would be no different than it is today, other than they receive more relays on fewer chains. RTTM and number of nodes stay the same.

Is there an area of the spreadsheet that can be used to understand how you are calculating average rewards between .2 and 1?

Most of your point seems to be surrounding this comment about “new average rewards”. If you could clarify what you mean here than I can address it.

@msa6867

This is in keeping with what you yourself admit to being a secondary motive. Namely to generate buy pressure for the token. Any new buy pressure that comes in the form of incentivizing supply-side actors to increase tokens staked to the supply-side is, de facto, a decrease in average reward per POKT staked.

Yes, having node growth from independent node running does ever technically reduce relays per node, but it creates growth for more folks to join our ecosystem.

Framing independent node growth as “decreasing average rewards per POKT” is a framing I haven’t heard used I absolutely don’t agree with that framing, as it is antithetical to POKT’s message and mission.

We should not drive policies that tries to reduce POKT participants, IMO.

[msa6867]

Sorry for the confusion, let me see if I can clarify

@shane

How does MaxChains effect the rewards of someone that is staked at 15k more than someone who is staked at 60k?

This is a non-sequitur. I never made any statement that could possibly be construed to imply that MaxChains affects differently according to bin one is staked into. Perhaps if I had simply said “decreasing MaxChain while holding all other system parameters constant”???

So why did I specifically call out MinStake and all the PIP-22 parameters out of all the system parameters (you may ask)? Because these are the parameters that would need to be changed in tandem with changing MaxChain if the desire were to keep the required stake per chain per “node” instantiation the same as it is today.

For example. suppose we were to change MaxStake from 15 to 1. And suppose simultaneously we were to change MinStake to 1k, change ServicerStakeFloorMultiplier to 1k and ServicerStakeWeightCeiling to 4k. Then we will have left unchanged the amount of stake required per chain per node instantiation. Thus someone who currently has only has 15k staked to a single node supporting 15 chains can still stake the 15k to 15 chains (via 15 separate nodes) once the MaxStake is dropped from 15 to 1. And someone who currently has 60k staked in bin “4” to 15 chains via a single node can still stake the 60k to 15 chains via 15 separate bin “4” nodes once the MaxStake is dropped from 15 to 1.

If MinStake is kept at 15k, ServicerStakeFloorMultiplier is kept at 15k and ServicerStakeWeightCeiling is kept at 60k, then the required stake per chain supported has effectively multiplied up by a factor of 15x so that the the node runner with 15k POKT would either have to shed all but 1 chain or buy an additional 210k POKT if they desired to continue supporting all 15 chains (obviously they could also settle for some state in between).

Setting aside hobbyists and small node runners for a minute. If there were only large-well-capitalized node runners, Changing MaxChain to 1 while holding all other system parameters constant is the equivalent to keeping MaxChain at 15 while raising MinStake and ServicerStakeFloorMultiplier to 225k and raising ServicerStakeWeightCeiling to 900k. [Note: I am certainly not advocating such a course of action. Decreasing MaxChain is much simpler and superior to raising MinStake for lots of reasons… I’m just pointing out equivalence.]

@shane

How do you derive that the average reward will be between .2 and 1?

I think what you are trying to ask is, “How do you derive that the ratio of new to old reward per POKT staked will be between .2 and 1”?

Pretty simple. The upper bound is just the ideal case that you present in your spreadsheets, so hopefully you are not asking me to comment on the upper bound (eg the assumption that the amount of consolidation on a chain will be proportional to the change in MaxChain. Eg, if 15k nodes are currently staked to chain X, then in the “ideal” case, after changing MaxChain from 15 to 3, only 3k out of 15k of those nodes will have elected to keep chain X (while the other 12k nodes shed chain X in support of some other chains)

The lower bound is just the opposite case where all nodes that support chain Y when MaxChain=15 elect to continue supporting chain Y after MaxChain is reduced to 3. Prior to the change, the node had servicer opportunities on 15 chains, and so effectively could spread the capital cost of 15k (MinStake) over 15 nodes (on average 1k per chain). After the reduction in MaxChain that capital cost gets spread over only 3 chains (on average 5k per chain). But for this particular chain where all the competitors also decided to keep supporting this chain, there is no equivalent bump in probability of servicer selection. Hence the reward per POKT staked drops to 1/5th the previous value. Hopefully this is clear (and rather self-evident?)

@shane

Framing independent node growth as “decreasing average rewards per POKT” is a framing I haven’t heard used I absolutely don’t agree with that framing, as it is antithetical to POKT’s message and mission.

Come on @shane, stop putting words in my mouth. Buy pressure from “independent node growth” is not at all what I’m talking about. I’m talking about the manipulated buy pressure placed on exisiting node runners trying to navigate the waters of the change, just as would happen if MinStake were increased to 75k or 225k or whatever.

[shane]

Thank-you for the further context.

Your premise is that if MaxChains is reduced from 15 to 1, then node runners would have to stake 14 new POKT nodes to serve relays on 15 chains. We both viewing the goals of node runners differently, which is why there is misunderstandings.

My thesis is that the goal of node running is to serve relays with the least amount of infrastructure required. The purpose of the node is to serve relays, in as efficient form as possible.

What you are suggesting is that if a node runner’s goal is primarily to serve 15 chains (regardless of it’s inefficiencies), then reducing MaxChains would require them to purchase more POKT so that they can continue serving their goal of 15 chains.

That goal makes no sense to me. The purpose of POKT is not to “serve 15 chains per 15k POKT”, it is to serve as many relays possible with that 15k POKT, in the most efficient way possible. Regardless of if MaxChains is 1 or 15 the AVERAGE relays per node stays the same.

If MaxChains was reduced to 1 and MinStake was reduced to 1k as you have suggested, then this is the comparison to just reducing MaxChains and leaving MinStake at 15k

MinStake 1k 15 (1k POKT) Nodes = 15 chains = 14 POKT per day

MinStake 15k 1 (15k POKT) Node = 1 chain = 14 POKT per day

Clearly, MinStake at 15k is more effecient because you can run fewer POKT nodes AND on fewer chains, to serve the same amount of relays and generate the same amount of rewards. Reducing MinStake has no benefit to average rewards generated per POKT staked, so let’s not continue down this line of thinking.

@msa6867

After the reduction in MaxChain that capital cost gets spread over only 3 chains (on average 5k per chain).

But they are still serving the same amount of relays, because they serve more relays on each chain when it is only 3. The goal of staking POKT is to serve as many relays as possible. Reducing MaxChains does not change the amount of relays the average node serves.

@msa6867

But for this particular chain where all the competitors also decided to keep supporting this chain, there is no equivalent bump in probability of servicer selection.

Why would competitors stack on a chain? My entire proposal is about how staking on an over saturated chain will reduce rewards per node… but your arguing that node runners may still do that anyways. So for what reason would a node runner reduce their own rewards by stacking on chains that doesn’t need more nodes?

No economic model can’t accommodate that level of ridiculousness If someone is going to self sabotage themselves, then they are welcome to do so, regardless if MaxChains is 15 or 1.

[RawthiL (Ramiro | POKTscan)]

The argument around reduction of Income per 15 K POKT is not accurate.

Total relays per day : 1 B
Number of nodes: 18 K
Avg. Relays per node: 1B / 18K
Avg. Income per node : (1B / 18K ) * RTTM

As you see the number of chains per node is irrelevant (its not even part of the equations).
The only changes in incomes per node could come from the re-distribution of nodes among different chains. The same reason why today’s avg. node reward is not (1B/18K) x RTTM x SSW.

I agree with @shane that this re-distribution could mean (little) more income to most nodes in the network if node runners act in an economically logical way.
We don’t know how this change will play out but certainly if a reduction of incomes per node happens, its due to poor staking strategies and not a model limitation.

P.S.: Don’t mix stake weight in this discussion, it is irrelevant, adds bloat and it must be removed in V1.

[Breezytm (BreezyTM | Stakenodes)]

Thanks @Shane for putting this proposal together. I am in support of it. However, there are two items I would like to see covered

1. What would happens to a node with 15 node staked if the parameter has been decreased to 5 and the operator decided not to update the stake? How would it be enforced? I chose 5 because it is a number that has been discussed for the past two years as a potential ideal number of maximum chains per Pocket node rather than 15.
2. To level the playing field I think PIP 22 is a must go. I understand what this proposal is attempting to accomplish and I believe having a base daily reward is important for new comers. A Node should be rewarded for worked done rather than amount staked. It was terrible then and remains to be a terrible idea. At least IMHO. Would you be willing to submit a proposal to undo it as part of the attracting “hobbyists” into the ecosystem agenda or as an addendum to this proposal?

It is abundantly clear the network is over-provisioned. I have seen several individuals in the forum, discord, and Telegram attempting to redefine the term “over-provision” to fit their narrative. For a long time the entire community used to agree on the meaning. The total amount of infrastructure servicing Pocket Network currently has far outpaced the demand side. I believe this could potential solve this issue while maintaining quality of service.

Lastly, I will say this. There are quite a lot of chatter about v1 this and v1 that. The launch of v1 shouldn’t be where we test our hypotheses. No one tests their changes in BETA before first implementing them on ALPHA. Why is Pocket any different?

[shane]

Good questions.

@Breezytm

What would happens to a node with 15

I asked in Discord to the team and this is what was the response:

So if MaxChains is reduced, it won’t have any impact on existing staked nodes. Any new nodes would need stake within the new MaxChains. From my understanding, any new staking transactions with more than the max chains simply becomes a invalid transaction.

This basically means that nodes with 15 staked get grandfathered over. This is an issue, because node runners that don’t update their stake can serve 15 chains while others are serving a reduced amount. Providers will try and hold out from updating stake, to keep as many nodes staked at 15, while any new nodes will get significantly less rewards from having to follow a lower MaxChains.

From what it looks like on my end, there will need to be a chain upgrade then, as you can’t have some nodes running on 15 and others with less. All nodes need to be treated equal.

Not only does this create an issue with v0, but this also means that if v1 is going to have a MaxChains parameter, then v1 needs more mechanisms to ensure all nodes are staked at, or below, the MaxChains. I feel this fully solidifies that v1 needs to be 1 chain per servicers, and MaxChains parameter needs to be completely done away with.

@Breezytm

Would you be willing to submit a proposal to undo it as part of the attracting “hobbyists” into the ecosystem agenda or as an addendum to this proposal?

I too believe (at the time and today) that was a horrible call to add different weights to the same work. Part of POKT’s thesis was that RPC relays will become a commodity, and POKT will provide the best access… but what commodity on earth has a different value depending who it came from? This would akin to the value of milk depending on how many cows you have. 1 gallon of milk is worth $1 if you have 100 cows, but if you have 400 cows then 1 gallon of milk is worth $4 Every product in the POKT ecosystem is significantly more complex because a relay has a different value depending on who serves it. It absolutely needs to be eliminated in v1.

That said, I’m not sure bundling MaxChains and Weighted Stake is the best path. We now have a question on if MaxChains could be changed in v0, due to there being no enforcement on existing nodes.

If someone wanted to do a proposal so eliminate weighted stake I would totally vote in favor

[Qspider]

@shane

I’m not sure bundling MaxChains and Weighted Stake is the best path. We now have a question on if MaxChains could be changed in v0, due to there being no enforcement on existing nodes.

If someone wanted to do a proposal so eliminate weighted stake I would totally vote in favor

Hard choices. Wouldn’t eliminating weighted staking just increase the amount of pokt nodes? Would it also remove producer blocks? Only validator stakes get producer rewards so if you remove weighted staking everyone gets producer blocks? Would increasing stake requirements for tiered nodes accomplish the goal? What is the goal for removing weighted staking?

We should be having incentives to go the extra mile and increase stake amounts to achieve web3 infra market dominance.

I would also like to add that it seems that just adding more chains to the network rather then editing MaxChains could be a solution to over provision. There should be some incentives for node runners to support newly listed chains and there is since newly whitelisted chains don’t start with much support. In the ever-evolving web3 market space we should be constantly seeking out chains to support and whitelisting them. I hear it’s around 50 chains now.

[Qspider]

Just to answer my own question I don’t think weighted staking was referring to validators and producer blocks and there would still be 1000 even if we eliminated weighted staking.