I really hope you get a really lucky one soon!
Today on Reddit you could read from cardano Key numbers that a stake pool needs about 1M to have a fair chance to mint a block - this is a bug in the ecosystem since it will lead to a concentration of multi- pool operators instead of a real decentralization. And it is completely in contradiction to the headlines of the project ābank the unbankedā - in a community that favorizes those who own a bigger portion of the cake there canāt grow a balance. This is - for someone who still is not deep inside the secrets of Cardano - a disappointing setup and I well followed the complains of small SPO that face the same problem. Money wants to fall against gravity and unfortunately Cardano enhances this unnatural behaviour. My personal opinion is, that people overestimated their cleverness and see their limitations right now.
Solution? Change the algo to a homogenious distribution of elected pools for mining at least for a certain percentage of new blocks- regardless of their pledge or saturation.
4 Likes
I donāt think itās a bug, itās how the system is designed. Itās Proof of Stake, you need stake to get blocks to mint. You are putting together a lot of different topics here, I donāt think that ābank the unbankedā has anything to do with stake pool operation.
You say that youād like a equal distribution among elected pools, what do you mean by elected? Who says if a pool is elected?
I operate a small pool that so far have minted zero blocks, so I cannot be accused of serving the big pool groups.
1 Like
elected by the random algo that of course has various parameter to set propabilities. Not elected by people or a commiteeā¦ If you have low stake on your pool you will most likely NEVER mint a block- and of course you are not responsible of favorizing big pools- was not my intention to say that
1 Like
I donāt like you proposal. This would lead to an explosion of operators spinning up multiple no pledge no stake pools, just to cash out on those free blocks.
There is probably nothing wrong with the election of block leaders based on stake, plus little bonus for pledge. It could be improved by taking the pledge leverage in account. The biggest stumbling block for small pools is the ridiculously high fixed fee.
2 Likes
Ok- I really do not claim to have full knowledge here but we seem to agree that it is desireable to avoid that small pools end up with having nothing instead of best ambitions and costsā¦ do we ?
1 Like
some sort of adjustment has to be done for the SPOās . I would also say a minimum of 1Million Ada (currently 1.2million dollars) is needed before minting 1 block, or being close to some reward percentage, so now I ājustā delegate to pools.
1 Like
However because such a large number is required to mint blocks, by your delegating means that you will chose a bigger stake pool anywayā¦because you want to get rewardedā¦and since the small stake pools donāt mint they will be entirely skipped and wonāt mint further anywayā¦so itās a vicious circleā¦ Iām not claiming i have any solutions, itās quite closer to the opposite of thatā¦ I know nothing (John Snow).
Anywayā¦after many sleepless hours recently (i can only investigate at night cuz i work during the day) i found that i did indeed not mint 2 blocks assigned to my poolā¦
Iām not sure exactly what is happening but thereās an error on the network or somethingā¦it has something to do with this:
Or this:
There was no warning or no implemented failsafe that i could observ what was going onā¦
I will go to work todayā¦come back home do further investigationā¦try not to die cuz i havenāt slept the last 2 nights and will try to figure this out and update everyoneā¦but as of right nowā¦ This was my issue:
In summation i lost 2 blocks: 2021-03-04T13:14:41.04 & 2021-03-12T05:48:49.05
block hash a1fdcd3737857071dea847df4dfa311ba677c0d1c0360c03fe4ad71ebe323a3d
slot 23297390
block hash 37ff0e99a9474e3fff0a7e48e638f0f77972d20152c64bb1d523b9ba4770deef
slot 23961838
Reason: InvalidKesSignatureOCERT
Somehow the node was able to start with invalid keysā¦and i donāt know how at this pointā¦i will update everyone but due to thisā¦i lost about 1800ADAā¦not thrilledā¦at allā¦imagine i had delegators ā¦this could really hurt my pool to the point of no repairā¦i made a lot of money from Cardano and Iām gratefulā¦but honestly Iām working to the bone for that sumā¦cuz itās a jungle in here as it is out thereā¦Iām really hopeful that things will stabilize at the end of this year and parameters will make senseā¦but now im questioning if wether Iām to hopefulā¦before i didnātā¦not a good feeling.
@SirFraggle You may want to have a look at this, perhaps it can deepen your understanding.
Please also consider, even if we had k=100.000 and out of those lucky ones, every pool were to make 1 block on average per month. It would be a lot more difficult to reach consensus among 100k participants. As a result, you would have much longer forks on chain that eventually are discarded. As a result, and because Bernoulli still holds, if would happen quite frequently that your pool wins a block after 6 months of waiting only to find out later, that this block gets discarded because it was on the wrong fork.
Iād say, be careful what you wish for.
I looked at it and I got the point. I also looked at other statements in that thread so a slight feeling of something is still not optimal balanced is not my USP. Can you give me a closer hint to your Bernoulli link? As mechanical engineer I am interested in such comparisons and probably my understanding of moving molecules is deeper than that of mass behaviour / game theory.
Thanks for your effort
cheers
Bernoulli says ā¦ if you have a jar with 22.630.000.000 lots (i.e. the total stake) of which 200.000 lots are winning (i.e. the poolās active stake) and you draw 21.600 times (i.e. number of blocks per epoch) - there will be a chance of almost 80% that you draw 0 winning lots (i.e. 0 blocks in that epoch). Only on average, you will win 0,2 blocks per epoch. Bernoulli also says ā¦ I donāt care what append in previous epochs. The likelihood of one or more blocks would look like this ā¦
Therefore, if you homogeneously distribute 21.600 blocks over 100.000 pools, each pool has a chance of 80% of not winning a block. To this you add the negative effect of not reaching consensus in time, so the long awaited block is more likely to get discarded because it is on the wrong fork.
In the end you might find yourself in a situation with less satisfied pool owners and not more.
Thereās a lot of pools with no history after e253ā¦strangeā¦like mineā¦
Only pools with block have full history
Iām really scared that as the network evolvesā¦the smallest of pools will slowly dieā¦and that begs the questionā¦what have we accomplished if itās the same as everywhere elseā¦the one most important reason I was excited about Cardano was about the prospect of minting a block with little to no money just the technical ability and the necessary gearā¦Iām sure they are thinking of those parameters and figuring them out but honestly thereās like one more chance to set them right at the end of this yearā¦after thatā¦it either makes you or breaks you as a network
4 Likes
In my humble opinion, the decentralization, security and availability of a blockchain that has the power to support a countryās identification system should not depend on people doing marketing on youtube or being favoured and ranked top by getting very early to the party.
Many other proof-of-stake blockchains assign the staking randomly to a set of stake pools that have passed some approval procedures, mainly about performance, availability and tech expertise.
Users that want to stake should not need to think about āwhat poolā. The system should split the available stake evenly among the existing valid approved pools. For the end user should be just a button āstakeā and an average APY. This happens already in other PoS blockchains.
Proof of the inefficiency is that there are millions of ADA being staked to saturated pools. An automated system would reallocate those ADA to other pools.
Every one has done this math, but if such a system would split all the current ADA available for staking among all pools, at current figures you could have 8M ADA per pool (considering 70% of 31B ADA and 2500 pools). All of them would mint blocks every epoch, the delegators would get their 5% and the system would be robust, sustainable, sensible and fair. And there would be room for many other pools.
But this doesnāt make anyone rich.
We criticize the inefficiency of bitcoin by wasting a lot of electricity, but we waste a lot of human resources doing marketing in twitter or youtube that should not be needed at all. We are automating transactions and removing human interaction to add validation and trust and yet we have to ātrustā our delegation based on better marketing or SEO?
I donāt think that a system where marketing people win over technical savvy people is a sign of robustness or sustainability. This should be all about technical performance and high availability.
If you were to sign a contract to run your countryās identification system, what would you like to hear:
- the system does not depend on individual operators, stake is evenly distributed, there are no single points of failure
- it depends on the people that do better marketing, we hope they know what they are doing but we really donāt know.
Many people have already pointed out that those pool farms are a danger to the stability and sustainability of the blockchain. Some have up to 5% of all ADA staked behind a single pair of relays. Those farms keep growing and growing, and if those go down for whatever reason, what happens?
If they keep growing at this pace, they could even perform a 51% attack. Is this a good thing long term?
The fact that some of them are huge even without marketing, just by being ranked top of the Daedalus app proofs my point that people just want a āstakeā button and donāt want to think about it.
I understand people that came early have a very good position they do not want to lose, and they have now a lot of power and interest in not changing the rules. They are getting rich by the current state of things.
But I donāt think this current situation helps with the sustainability, image or security of a blockchain with such high goals as Cardano has.
Cardano should think longer term.
My 2 cents.
12 Likes
Excellent - Well Informed & Considerate
4 Likes
Hello javix,
Undoubtedly the small pools are disadvantaged by the current protocol parameters, but the main principle of assignment of slots based on stake is sound and much fairer than the random as you propose.
Another stumbling point would be approval of the pools, for which youād need either a central authority (not good) or a consensus such as catalyst vote - then you are back to marketing.
If an approved pool misses their blocks, why should everybody single delegator in the entire system be punished and not just the delegator of that pool.
The big players could set up 100ths of pools with lower costs per pool then a regular guys, so they would easily accumulate more power over the ecosystem.
Iād love to see the minimum fixed fee of 340A reduced to basically close or equal to 0 and implementation of a parameter reflecting pledge/active stake ratio.
Then for any delegator the comparison of pools would be very simple a small pool with 1 estimated block charging 0 fixed and 2% will become the same attractive as a big pool with 50 estimated blocks charging 0 fixed and 2% or 500 fixed and 1%. So its a win win situation.
Nevertheless its good to see you or anyone think about and propose other ideas.
3 Likes
Very nicely said. I think since this network has recently become decentralized, this should totally be an initiative to makes those changes in the parameters. If the network end goal is exactly that sustainability you are talking about it would be adopted through the governance system Cardano has created. I am in support of what you stated earlier and it has been stated elegantly and efficient. We should make a proposal to see what the community really wants.
4 Likes
Hello Triton-pool, thank you for your reply.
but the main principle of assignment of slots based on stake is sound and much fairer than the random as you propose.
Sorry if I explained myself incorrectly. My suggestion is to randomize the assignment of stake to pools, not the blocks. The block assignment / lottery mechanism based on the amount of stake makes perfect sense.
Another stumbling point would be approval of the pools, for which youād need either a central authority (not good) or a consensus such as catalyst vote - then you are back to marketing.
Well, yes and no. You could have IOHK/CF or you could do peer review by the existing pool operators. And need a reasonable amount of peer reviews to get approved. IOHK or the Cardano Foundation could be just witnesses of the process and help in case of litigation. Catalyst vote has the same bias as stake. The more you have, the more power you have, less democratic, more centralized and corrupt (self-sustainable for the monarchy) in the long run.
You could restrict approval or pool certification to 1 per identifiable person, and there are plenty of ways to verify this. You could also check that they donāt share same infrastructure, etc. A peer-review approval process could include a personal interview, so you donāt add pools using your motherās passport number, etc. This is a blockchain project, isnāt it? There are many things that could be done to make it trust worthy.
I doubt you could spin up 100s of pools with such a system - of course it would need further thought.
This would of course request from the SPOs to participate in peer reviews, but I think that time is way more valuable than filling up the SEO space with the same content over and over.
IMHO, if you want to be an SPO you should collaborate in useful, efficient ways.
Iād love to see the minimum fixed fee of 340A reduced to basically close or equal to 0 and implementation of a parameter reflecting pledge/active stake ratio.
Then for any delegator the comparison of pools would be very simple a small pool with 1 estimated block charging 0 fixed and 2% will become the same attractive as a big pool with 50 estimated blocks charging 0 fixed and 2% or 500 fixed and 1%. So its a win win situation.
From my point of view this is still following the approach of wanting the delegator to be informed and decide among the pools. I think that is a nice idea but a bit naive, and as proven, susceptible to abuse (or they just click on the first one they see, regardless if they are saturated or not - most probably donāt know what that means).
If all pools have more or less the same amount of stake, rewards are the same, no need to choose. I donāt say you should not modify the current rewards system, but in my opinion competition between pools makes no sense and no good to the sustainability of the project. We should collaborate for the greater good, not fight between peers for individual gains. If some feel this is not going to make them rich, maybe this is not the right project for them.
If a pool is not properly managed and misses a block, the system can behave as it does now and reassign the block to another pool, so nothing is lost, only the delegators of that pool are affected, not the entire system. Then the SPOās community can have a system of strikes with the rogue or mismanaged pool. If you are not wanting to manage and maintain the pool at the professional level that is required, then you lose your certification, and the stake gets assigned to other pools. Just a thought.
From my point of view, such an approach would professionalize the SPO role and that would have positive implications for the robustness, image and future of the project.
2 Likes
Hello javix,
If you randomize stake distribution, then you randomize the slot assignment with it.
IOHK/CF has no intention to be a centralised authority, as expressed by Mr. Hoskinson, the role IOHK will be fading as community governance get promoted.
Cardano is based on proof of stake. The stake is what protects the system from being kidnapped. The stake is your skin in game, the higher your stake, the more you are interested in a success of the project.
Catalyst is the execution of POS. What bias are you talking about? Anything else than your stake would be biased.
This is completely wrong. I am sure you cannot present even single one way to do it without creatin a central authority = you create coruption potential. Why would anyone not be allowed to operate more than 1 pool? It is a blockchain, You can generate as many addresses and run as many pools as you wish as long as you pay the deposit.
This one is right. You would not be able to start even one pool without a permission of the privileded actor, whose you are supposed to persuade by demonstrating then past capability of a pool that never existed.
By the way someone would have to pay the privileged actor for their āserviceā.
Here again. A group of the privileged SPOās (regardless of their pledge I guess) will decide if they let competitors to enter the market and take a portion of their profit. Are you even serious? This is a textbook example of a conflict of interests.
So you suggest that for the good of the minority who do not care about their own money enough to do little research and choose a decent staking pool, we should strip everyone else of their right to choose?
Seriously?
Then there no reason to keep more than one pool.
Competition is what drives improvement in all aspects of human activity, this proves right again and again
Didnāt you say one paragraph above that all pool will yield the same rewards so delegators should not have any right of choice?
I had to read your post several times, and the more I think about it the more I hope none of it ever gets adapted, because this would be a disaster. A huge betrayal of the Cardano mission and ideas and all the people who believed in them with their investment both by their money or effort.
2 Likes
I had this error on testnet. After the KES rotation, I only updated kes.skey on the producer, and not node.cert.
I was also surprised that it didnāt validate this at startup - the metrics all looked good until it was time to produce a block. Anyway, I copied over the node.cert and then it produced blocks.
Iād definitely recommend using leaderlog each epoch and checking you produced the expected blocks, so you can fix it more quickly ir something is wrong (the RNG makes it hard to tell by just monitoring the creation of blocks).
1 Like