Current Operators....Actual Cost to Operate vs Actual Returns thoughts?

Hi @Chrisomi,

Welcome to the club!

A few facts first:

  • you can’t start a pool with 50ada, since you need 500ada to even register your pool. you will get it back when you retire your pool
  • a registered pool with 50ada stake will have a statistical chance to find a block every 270 years. check this to roughly calculate your chances: Mnemonic device to help estimate the odds of your stake pool finding blocks (ada4good.com)
  • you need at least 2 servers, 3 i recommended. Quality cloud servers that meet the resouce requirements will costs about 150 dollars a month to rent.
  • finally, a small pool (under 1m stake) will give their delegators about 1,5% less annual returns than a big pool (5%).

@Chrisomi sorry I misread your message. you said 50k not 50 :slight_smile:

the chances of finding a block with 50k are every 100 days on average.

This calculator will tell you exactly:
Cardano Reward Calculator - Dynamic Strategies

I’m also running a pool primarily for research/learning. The operational cost is minimal, because I run on hosted RaspberryPi4 which is also CO2 neutral. The 1% pool margin goes to charity, which is then used to plant trees. The pool rewards go to loyal delegators on top of what they get anyway (i.e. 850 ADA to 14 delegators) - we started to do this 3 months ago (i.e. since it became apparent that 360 p.e. is much more than actually needed) and here is how it works.

If you own the pool, the math is simple …

73 * YourStake / TotalStake * SlotsPerEpoch => AnualBlocks
AnualBlocks * BlockReward / YourStake => ROA

73 * 50k / 22.75bn * 21600 => 3.62
3.62 * 750 / 50k => 5.4%

From this you subtract your actual running cost to get your profit. For example …

3.62 * 750 - 1200 => 1515 ADA p.a.
1200 / (3.62 * 750) = 44.2% cost

Now, lets compare this to simply delegating your 50k …

The cost for delegating to a pool depends on pool size of the pool more than anything else, so you would want to delegate to an almost saturated pool. We can also ignore pools that have a > 0% margin and here is why.

With a saturated pool you would have a cost of 0.7% …

3.62 * 750 * 0.07 => 19 ADA

which is considerable less than the 1200 ADA we assumed above. Attracting delegators would work in your favor, but who would want to delegate to a your pool and have a cost of > 35% ? The story of your pool would have to be convincing enough to compensate for that.

There is a point of break even, which for this example sits at …

1200 / 0.7% / 750 / 73 => 3.13 b.e.
3.13 * 22.75bn / 21600 => 3.3m

So if you have > 3.3m, it will be more profitable running your own pool. It also doesn’t matter whether that pool attracts delegators or not, your ROA will be the same.

Bottom line, the cost of running a pool is considerable both in terms of ADA and effort. It may still be worth it, depending on your long term objectives. When delegating, choose a pool with 0% margin - otherwise the rich will simply get richer. Also, a good pool would not simply pocket all of its pool reward - i.e. surely not all of the (6 x 340 => 2040 p.m.) are needed to run a pool.

Still, there often is good reason to charge > 0% margin for example to use some/all of that to do some good in the world. When deciding which pool to delegate to, this can be a decisive factor in line with - “Making the world work better for all”.

9 Likes

I am sorry but are you saying here that it is ok to not pay people who put considerable time and effort on the table ?!

1 Like

What will happen with your fees when you are done researching and learning? And the operational costs are little compared to the actual time operators spend to operate their pool, add value to the community and to attract and keep delegators.

1 Like

I’m saying the math works like this when you apply it. There used to be a time when it was perfectly ok to say “The operator needs those 340 p.e. and a little extra to run a pool”. Economic realities have changed since and now 340 are worth a lot more, which of course we are all very happy about. It also means, that no one can honestly still claim to “need” those 340 p.e. to run a pool.

Aside from (subjective) fairness, the math also says that eventually all stake will cluster around k saturated pools. The 340 is the deciding factor in that - it essentially makes large pools more attractive than small ones to achieve that convergence toward k saturated pools.

What I do with the payback scheme is essentially working for the benefit of the delegator. It essentially makes a small pool as attractive as a much larger pool because the effective cost is similar. For me its currently not a problem that I do this non-profit.

When I’m done researching/learning the running cost will drop to what is actually needed to run two simple servers. The payback mechanism will then be implemented by a Plutus smart contract, which is also the beneficiary of all pool rewards. Because I then don’t receive any pool rewards anymore, I won’t have to hold back the tax escrow anymore. The payback will converge towards a theoretical maximum of …

6 x 340 - running cost => 2040 ADA - 60 USD (or so)

In case the pool should attract more delegators, the 1% charity margin might become unreasonably high for planting trees. Which would then be a good time to review the charity model and either support a wider range of good causes, add some to the payback pool or do something else with it completely. In any case, we can cross that bridge when we come to it.

2 Likes

OK, but you didn’t address my main point.
You assume that the costs of running a pool are only in the out of the pocket costs of the the servers. I assume you expect all pools operators to invest their time and expertise for free. Lucky for the network, most pool operators are incentivized in a more economically sane way so we can be sure they are professional and will continue taking care of the network, knowing their expertise and time is paid for accordingly.

An as you mentioned above, the minimum fees where set when ADA was a few cents and that is why this is being lowered soon. But still, this is a minimum not covering what I am pointing out above.

So to summarize, sure it’s great that lots of people are willing to start a small pool and want to be competitive and therefore it would be nice if the fixed fees would be lower, which they will be soon, but in the long run, we should expect to pay our operators also for their time, if we want a resilient and flourishing network.

3 Likes

I guess, we’ll see what actually happens with the fix, a0, k and other changes to the reward formula and how that effects how SPOs want to run their pools.

1 Like

I am afraid your calculations are way too optimistic.

  • Pools (serious that is) will not get away with 60$/month of infrastructure.
  • Long term commitment also involve way more than just server rental. This means taxes for professional revenue, legal framework and accounting.
  • Time invested by one or more (preferable) people.

Nonetheless, I understand what a free market is and best of luck to anyone out there !

3 Likes

No problem, you can plugin the numbers that work for you. I’m not claiming that running a pool should cost this or that and because I haven’t yet seen convincing evidence that $2040 are actually needed to run two simple servers (especially not on RaspbarryPi4) we decided to give our excess back to loyal delegators.

1 Like

Raspberry won’t cut it if you plan on staying around for the long run. My 2 ada.

2 Likes

Agreed. “staying around” is the main point. To stay around for the long run almost always means, good and consistent investment of time and material. The rest is marketing in my opinion to attract delegators so that pools can stay around. And when they get the delegators, they need to increase fees and stop giving back, since that is needed to stay around while providing good quality service.

For the time that ASTOR has been around, RaspbarryPi4 did cut it, then it didn’t for a while, then it did again and I’m quite happy to adjust the setup to wherever the mood of the network takes it. In all case, we can adjust the payback amount accordingly. Currently, it’s quite a bit, with Alonzo it’ll likely be lots more and perhaps some day the scheme may become obsolete. In any case, we are committed to only charge what we actually need. Again, I have not yet seen convincing evidence that $2040 are needed.

I don’t care much about the $2040. That is mainly due to the high minimum fixed fees which will change soon. What is more important to me is that I don’t believe in sustainable pool operation for a couple of dollars a month. We see that in normal business all the time. New businesses starting and giving everything away for a discount to attract customers. Which is fine as long as they don’t go bankrupt or deliver the quality that matches their prices. Disappointing their customers along the way.

I am against charity pools. Charity doesn’t work. It is a means for the rich to massage it’s gilt but continue to screw the poor. I have seen many a country ruined by charity.

In a decentralized system the goal should be for each individual to be able to run once own staking pool.

Every work done to enhance the network security should be compensated but not in perpetuity. That is should not be the purpose of an open decentralized system.

The argument that SPO is just like any other business and one must work hard to succeed is the same argument used by the present economic system, where some are born privileged and destined to stay ahead and others are not.

If we let the pre-existing advantages such as geography and money in the bank decides the outcome of the SPO competition then Cardano will quickly evolve in to a centralized platform and enabler of centralization no better than banks increasing the gap between the poor and the rich, which already is happening.

Moreover increasing difficulty for the disadvantaged doesn’t help network security which should be the only consideration in designing a distributed system.

In this regard Alogorand may have better design than Cardano.

ETH2 has a system with currently 138000 validators and still some pending

Of course, they also have a limited number of blocks to mint per epoch and a maximum of ETH they can distribute among all validators. As a result, a validator can propose a block about every three weeks and in between receive rewards for attesting blocks from other validators. You currently need to have 32 ETH (i.e. $128k) to take part in this. I quite like this approach, because it does not put validators into a competitive position against each other. If a validator fails to propose it’s assigned block it gets slashed. So, you still needs to know what you’re doing on top of providing that initial stake.

The Cardano POS system is designed such that all stake will (eventually) cluster in k saturated pools. With k at 500, it means that > 2000 pools will be driven out of business in a more or less painful way. These are the fundamentals that folks sign up for when starting a pool. One could call for an increase of k, but (because there is no notion of attesting) there is an upper limit beyond which it’d be nonsensical to increase even more (i.e. there are 21600 blocks to mint / epoch). It would however not solve the “problem” for folks that are not among the “privileged” k - not matter what k actually is.

It is easy to say “It should work like this or that” without having done a fair bit of academic research and game theory. I suspect, that over time POS will evolve such that it does not only work for the network, but is also socially sound in a way that SPOs get compensated for what they actually do (but not more).

In this regard Alogorand may have better design than Cardano.

Lets hear, in what way that would be so.

What do you propose then, socialism? There is no such thing as a free lunch.

1 Like

No! fair playing field. The barrier for participation has to be removed. At the moment the Cardano consensus encourages centralization, by allowing a pre-existing inequalities to determine allocation of future rewards. Right now not only once salesmanship, or how many rich friends one has, but the geographical location of staking pools determines success. That is the system we are trying to leave behind.
We may try not acknowledgement but a royal class is emerging in Cardano.

That said I do acknowledge the dilemma the designers have in creating the necessary network effect for the platform to succeed, which needs the privileged, preventing Sybil attack and making it fair for all to be able to participate.

And I hope the “royalties” who are fairly compensated by now and holding large stakes will not be greedy to let the present barrier to continue. I am grateful for the work they did.

“…The Algorand platform requires minimal processing power and modest IT resources to join. All online users who possess algos are automatically eligible to participate in block consensus…”
As I said on another post this may have difficulty reaching the network effect, but Cardanos goal should be to make running a staking pool automatic with owning a stake.
The hardware or geography or the ability to control code should not be a pre-condition. And we do that not for charity or even to be fair. Decentralization is required to have a healthy network. Monopoly in any form weakens the network.