Raising the Bar: Incrementally Boosting POKT Staking Requirements

Ideas
1

Hey POKT Network folks,

I’ve been batting around an idea in a few other threads, and it’s generated some interesting discussions. I think it’s time to give it a thread of its own, so here we go.

We’ve seen how the cost of becoming a node runner has fallen dramatically. Initially, the staking requirement was above $15k; now, it’s under $500 in tokens. This change, combined with Lean Pocket’s fantastic work on slashing hardware costs, has been great for accessibility.

But it seems we might be paying a hidden price. With the average stake per node hovering around $1400 I worry we’re undervaluing our network and diluting our token value.

So, I propose we slowly roll back the tide and gradually reintroduce higher stakes, kind of like the good old days, but in a steady, manageable way. We might pull this off using a “ServicerStakeWeightMultiplier”-like system that rewards keeping stakes above a certain threshold. To avoid causing a price surge, we’d need to move cautiously, increasing stakes in small steps to accommodate limited market liquidity.

Here’s a paradox: Higher stake requirements might initially reduce ROI, leading to some runners opting out. But a reduction in competition means better rewards for those who stay and might actually attract others to join.

Furthermore, this gradual increase could push “overstaked” LeanPockets to consolidate. For example, someone running 20 nodes with an average $1k stake would face a $10k additional investment per $500 increase, or a total of $180k to reach a $10k stake requirement. This could promote decentralization and further reduce node count.

Although it may seem counterintuitive, I firmly believe that decreasing the node count is crucial, especially during times of falling prices. After all, most, if not all, node runners need to cover their bills in USD. With an average operational cost of about $50 per node, POKT network is currently shouldering a hefty $1M per month while at the same time value of the stake is going down the hill.

Final thoughts - with higher stake there comes greater responsibility. That currently does not seem to be a case. For instance, becoming a validator costs roughly $2k at current price. That means (does it?!) you can overtake first 1000 validators just by staking a little above $2-3M in total, which is less than 10 % of currently staked POKT.

I’m no expert on this, so do take my thoughts with a pinch of salt. I wanted to spark a discussion without crowding other threads. All feedback is appreciated!

image
image500×756 94.8 KB

4 Likes
2

Thanks for setting up the discussion, Spenc.

The idea of increasing the min amount of POKT staked per node to lower node count had been proposed before and my recollection is that the main stake can’t be changed without a fair bit of technical work that was deemed not worth the deviation from v1.

However, you bring up an interesting idea, can we work with ServicerStakeWeightMultiplier to penalise lower staked nodes, forcing consolidation?

So anything under 30k would be penalised with a <1 multiplier, encouraging consolidation at 30k or above?

The other general rebuttal is that the cost of the nodes is not great anymore, but the cost of serving chains is now the bulk of the cost. I assume lower nodes = lower chain access = lower costs, with the Altruist network still available for any severely underserved chains, but I’ll let those more technical than me comment on that.

2 Likes
3

This is an interesting topic to discuss but I think it will be better to think it in terms of v1. Some of your points are valid, and the current servicer stake seems low. There also the issue that the stake is not really an stake, because you never get slashed, so it is only immobilized capital (not sure if the term stake applies here also). This is corrected in V1, where is slashing risk for servicers.

Nope.

PIP-22 should die in V1. Lest not add more complexity there. Moreover, with the portal migration the Cherry Picker is gone, we don’t know whats coming and stake weighting is extremely sensitive to QoS sorting and profiling models.

No, please not again.

2 Likes
4

This is actually great point. I gave it 2nd thought with v1 in mind and I think some of the problems might be addressed in v1 thanks to Fisherman (quality over quantity). For instance investing into providing high-quality service in v0 carries huge risk as you can be easily outperformed by low-quality operators.

Is the stake weighting implemented in v1? For instance if there are 2 qualitatively equal service providers the one with higher stake earns more rewards?

2 Likes
5

It is not defined yet, I’m advocating for removal of it here (the thread touches several points on staking).

1 Like
6

I’ve read the other side’s points. But don’t you think it makes sense for the little guys running nodes to step back to improve service quality and network durability? Here’s what I’m thinking:

  • To deliver the best service, we need pros running the show - folks with the means to keep upgrading their setup. This basically rules out most of the small-time operators.
  • The market wouldn’t be ruined even if we only had like 10 big players (think c0d3r, poktscan, etc.) running all the nodes. Like, if Infura or Quicknode had extra capacity, they could whip up a similar product for the average POKT holder and maybe give better returns.
  • These big operators could be from the old-school web2 or the new decentralized web3 governed by token holders (= decentralised while centralised)
  • Being big also means easier access to bulk discounts on stuff like bandwidth, storage, and computing power or just have way better utilisation of these resources, so they can optimise costs very well

So, even though these big players make things feel more centralized, the market itself remains decentralized, right? This setup allows us to squeeze costs while still offering top-notch service. I’m not sure the small node operators can pull this off - they might hang around in areas that the big guys aren’t interested in, but I think they’re headed for the exit in the long run.

However I agree stake weighting is a tricky problem. What if we calculate some limit after which staked amount doesn’t provide any benefit (using median in the formula or sth)?

Random idea: what if we introduce staking parameter like “DiscountRate”, so nodes would be able to voluntarily reduce reward per relay in order to improve their scoring? So when there would be operators with high stake, high QoS, they would be able to compete for bigger market share by lowering their margins → minting less rewards globally, while it can be more profitable for node operator (eg if discount 10 % brings 12 % increase in relays).

1 Like
7

I guess the question is: how do we incentivize Nodes to consolidate to 30k+ POKT for v0.

Agreed, v1 is a different beast and adds Fishermen/Watchers etc so ignoring the economics of v1 for now.

My understanding - and correct me here if I’m wrong, is that an economically ‘simple’ min stake of 30k is too time consuming to implement for v0, so we are reduced to your old friend, PIP-22? Do we have other alternatives? What would be possible to help drive the above outcome? Slashing nodes with less than X POKT attached?

Wondering if there’s a possible quick fix here that can be reliably implemented, or if it’s not worth working on till v1.

2 Likes
8

I understand it the same way. While increasing the minimum stake within the protocol might be too big overhead, it should be still possible to implement it somewhere like here portal-api/src/services/cherry-picker.ts at a53c4952944041ba2749178907397963d7254baa · pokt-foundation/portal-api · GitHub so it wouldn’t need any protocol changes, right? That way we don’t force it on protocol level, however relays to nodes with low stake can be significantly reduced on the gateway level. Am I missing something there?

If we want to push average stake on more meaningful levels in v1, it certainly make sense to start exploring this in v0. For example what happens if we “deprecate” 15k stake? How many nodes will drop off? How many consolidate? How it affects market conditions - there’s not enough liquidity for large buys and earned tokens might not be enough to cover 100% increase (15k → 30k), so it’s better to do it incrementally rather than one big bump.

1 Like
9

This is dangerous to QoS. The apps gets, per protocol, a list of 24 nodes. If you segregate the 15K/30K nodes you will have a session with fewer nodes. From the rest of the nodes you will need to get those with high QoS, this includes removing nodes that are not geomeshed for example, or that are simply bad.
You will end with an smaller pool of nodes from where to choose.

Even when this is not a protocol change, I still believe that it is not worth the problem in V0. Changing this in V1 will be more organic, as the staking paradigm will probably change (due to regional or chain limitations).
Changing this will not have the expected effect (in V0):

  • Stake holders with less than the required POKT will move to pools. There are a couple of services from where to choose. So, total POKT staked will probably remain constant.
  • The cost of a node in a lean setup is negligible for node runner operations at scale. I hardly think it will make any difference.
  • Even if the cost reduction was real, the big node runners would be the ones that can take more profit from consolidation. Changing from 200 nodes to 100 nodes on small node runners will probably mean the same hardware cost, on the other hand, moving from 3000 to 1500 might be worth. This means that raising the stake will not help decentralization, for a small node runner it is the same.

In v0, after the introduction of LeanPOKT, you can think of the 15K POKT that a node costs, as the price of a “ticket”. The more tickets you have the more POKT you make (given avg. QoS). This is so because the base cost of a LeanPOKT node will remain constant from a single 15K node to hundreds of 60K nodes. So, your single 15K POKT node has the same USD cost as a cluster of 200 60K POKT nodes (12 M POKT).

2 Likes
10

What if we start reducing nodes from that list incrementally until there’s very minor portion of 15K nodes staked?

Wouldn’t it be easier to start increasing stakes incrementally in v0, so the migration is more seamless in v1?

This is great point and I realised it reading the other thread you mentioned. As I mentioned above I don’t actually believe that small runners can survive in this decentralised market. Imo it’s enough to have decentralised protocol, where everyone has the opportunity to compete freely and the fact is that big operators will most likely provide the best performance x cost ratio and in the end that’s what everyone wants, right?

Note that any DAO can decide to make such investment, hire experts, build performant chain infrastructure and then compete with other node runners. So it can be centralised operator, while decentralised at the same time.

The thing with current staking is that the big players are actually incentivised to push price down, because then they can accumulate at lower price and repeat, which gives them the advantage (= more “tickets” for cheaper value) over the node runners who staked at higher levels and are held hostage with negative ROI. Although this is the feature of free decentralised market and it’s perfectly fine, it seems we reached a point when becoming a node runner or even validator is just too cheap imo.

If v1 is going to address it, shouldn’t we try to prepare all node runners for that stake migration incrementally?

1 Like
11

Yes, agreed. Small node runners wont make a living out of Pocket, even only very few big ones could make it (but not sure tho).

I’m not sure if I understood this correctly. Big stakers will accumulate faster no matter the price of the “ticket” or their stake level. What I mean is, 12 nodes of 15 K will produce the same as 3 nodes of 60K in the long term. Yes, the first strategy will have 12 tickets and the second only 3, but the fist one is only paid x1 and the second one x4, so they are the same (almost). Both of them will reach 15K gains at the same time and can increase their number of nodes by one at the same time, then the fist strategy keeps creating nodes while the second first up-stake them.

I’m not sure if this is in line with what you say, but I see that the price of the “ticket” (15 K POKT) controls the speed of the staking (and to some extent the speed of tokens flow?).
Higher minimum stakes (much higher than 60K POKT) could create some sort of additional scarcity. The POKT will take longer to accumulate, creating longer waiting periods on the iddle POKT. In order to put your idle POKT to work you will need more POKT and at some point you will be pressured to buy at higher prices if you feel that you are loosing an opportunity to create more.
However this could also be fertile ground for pools, that could just allow the speed of the tokens flow to continue…

Agree and it is not good. But also there is the problem of exchange rate. POKT is very cheap now and if the stake goes too high and price recovers then it would be a restriction to protocol growth (not a problem for V0, but it is for V1).

This could be a valid thing, if we achieve some clarity on V1 economics first.
I would not agree to an increase of stake that is just a new higher value, because by itself is not worth it IMO.
On the other hand if we agree on some goals to V1, like:

  • new stake levels (really big ones)
  • minimum number of nodes required to ensure all chains and regions will have enough variability in sessions.
  • restrictions on V0 (i.e. if the POKT price goes up and you cannot stake any more nodes, don’t complain until V1)
  • staked POKT distribution (taking into account other actors)
  • more that I cannot think of right now

We might achieve something by modifying the stake minimum and it might be worth to do incrementally, allowing rollover of nodes.

Disclaimer, this post contains too much of me just thinking out loud, don’t take this too seriously…

2 Likes
12

I think that’s true only if we forget about USD value. For instance if someone builds a small infrastructure and stakes 50 nodes at avg price $0.35, the initial investment is $262k + operational costs let’s say $500/mo. Let’s assume high node rewards, $50/mo per node. That’s $2k/mo net, $24k/y. Nearly 10 % annual yields, not bad.

Now if someone wants to enter this business, he can figure out these numbers (roughly) and manipulate the market price down (not that hard tbh). So at 1/10 of the price ($0.035), same investment gives you same annual yields even if the rewards are $5/mo per node, however the ones who staked at higher price levels are facing significant disadvantage as they are no longer able to cover operational costs by earned rewards, therefore they can either shut down or take cover losses every month. I know it’s a bit of speculation right there, however it’s the way how you can squeeze operators who are not able to keep up (regularly invest when price goes down) while you can still make a benefit from such price drop.

Mechanism that would bind minimum stake to USD value doesn’t really solve this, however it makes it significantly harder as with dropping price you would need more and more tokens to stake which would lead to significantly higher demand → price goes up.

For instance, if an operator that bought his tokens at $0.1 decides to exit and realise the loss, I can buy it from him for a 1/3 of his initial investment and run the same amount of nodes. Which is obvious, if someone realises a loss, other party can profit on it. In scenario where the price is somehow bonded to USD value, I would still be able to buy his tokens for one third of his initial investment, however I would need 3 times more tokens in order to stake a node.

That’s why I think starting in v0 makes sense as these new staking requirements can create big demand for tokens while there’s no real liquidity. That can help with v1 launch as things can be more stable? Also if there’s weighted stake in v1 it basically solves this naturally.

Another idea for non-protocol option how to increase stake sizes it to do an air drop for node operators, so we can capture snapshots every 20 days and if you stake more than X for defined amount of months, you will receive POKT airdrop when v1 launches.

I should have mentioned the same disclaimer on each of my posts haha :smile: Just throwing ideas

1 Like
13

Oh I think I understand what you are saying. I you dont keep investing into nodes your returns will go down and you wont be able to compete with new actors that enter the market with the same investment as you. I guess this is true, but only due to the POKT price action. I dont know if there is a way to solve this with higher stake minimums…

Would this be solved with solid tokenomics? if POKT had more trade volume the price manipulation would not be so large.

I’m against airdrops, that’s just printing and seems like an scam almost always…

1 Like
14

I think that would certainly help. If there is some floating minimum / weighted stake, it would support this - with price falling down, we could effectively increase minimum of staked tokens which means removing liquid supply from market, so this spiral when price is pushed down would not be possible as there is not enough tokens.

Example I mentioned in the other thread is if all node runners agreed on min stake $5000, this would effectively push token price at $0.07 vaporizing all liquid tokens.

While that sounds silly, in real world where POKT market has very little liquidity, its more likely that the price would rocket through the roof. Thats why I mentioned we should think of ways how to do it incrementally, so its possible to measure its impact.

The question is how to avoid the scenario you mentioned - setting a min stake limit eg 150k, then price goes to $1 and nodes would be staking $150k. I think if we find a way how to make these increases incrementally, it would work the opposite way too (incrementally decreasing).

Besides binding min stake to usd value there might be the other options like some formula that incentivizes higher node
utilization, for instance staking new nodes to regions where is low utilization rate would be proportionally more expensive.

1 Like