Which Proof of Stake System is the Most Viable

[DeathAndTaxes]

Say the fyookball-1000 mining rig (a fictional machine) gives you 1TH and runs at 1000w.  Now the fyookball-2000 comes out and gives you 2TH but runs at 100w.  Well, now you doubled the hash but cut the electric bill. By 90%.  

If every miner owned 1 FB1000 and suddenly replaced with the FB2000 , each miner would be making the same when the difficulty changes anyway.....but the power consumption is down.

Except why buy 2 why not buy 10 and make even more money ?  If it is profitable to deploy 2 it is profitable to deploy 10 and people will.  They will keep deploying more and more and more hashing power and until the ROI is right back down to that low level it was before the more efficient hardware came along.

This is the same as ANY commodity business.

Yeah but the thing is the rig is the main cost not the electricity...at least for now.

Well "main" cost is kinda subjective.  At $1 per GH/s and 1 J/GH energy is pretty close to half the cost.  Margins are still high on hardware so expect hardware prices to fall much faster than energy costs and that will shift more and more towards energy.  Still it is a pointless distinction.   Miner total cost is amortized hardware + energy costs.   If the network gives miners $500M annually then miners are going to spend pretty close to $500M annually as they collectively drive the margin very close to zero (possibly negative).   It doesn't matter how efficient the tech is, miners will simply buy and deploy more gear until margins are so low (or negative) that deploying more hardware doesn't make sense.

Once again we have been seeing this continually for five years so it isn't some academic theory.   It is a classic prisoner dilema.  Sure if there is new technology and miners agree to limit the hashrate of the network then they can increase their margins but they can't and if they tried someone would cheat to make more while the margins are higher.  Economics will drive hashrate up and margins down when more efficient tech is possible.

[jonald_fyookball]

Just updated my post above....man you are hard to argue with (mostly cause you are usually right     )
Got your points.  Thx for the discussion.

[Zer0Sum]

Gold Medal:  Peercoin

I like this solution the most because
it is the simplest.  Simple solutions
are usually the best, which is why
proof-of-work and longest chain has
stood the test of time.

It would be nice to see if Peercoin
would work without checkpointing.

Geniuses simplify things...
While merely highly intelligent people complicate the shit out of everything...
(See Ripple, most 2nd Gen Crypto Assets, etc).

Twitter is genius... because it's dead simple.

This is why PeerShares could sweep away the Pretenders...
Sunny knows that "good enough" and "dead simple" wins.

[DeathAndTaxes]

The important thing is the efficiency is getting better and better, and we are getting more security for the dollar...and bitcoin is in good shape.

Agreed.  It just important to understand efficiency can't reduce cost only improve security (by eliminating the potential for an attacker to reduce cost by using more efficient tech).

[jonald_fyookball]

The important thing is the efficiency is getting better and better, and we are getting more security for the dollar...and bitcoin is in good shape.

Agreed.  It just important to understand efficiency can't reduce cost only improve security (by eliminating the potential for an attacker to reduce cost by using more efficient tech).

What you called a pointless distinction earlier is not pointless, because of the ratios of hardware to electricity. I think I almost had a good point.

[DeathAndTaxes]

The important thing is the efficiency is getting better and better, and we are getting more security for the dollar...and bitcoin is in good shape.

Agreed.  It just important to understand efficiency can't reduce cost only improve security (by eliminating the potential for an attacker to reduce cost by using more efficient tech).

What you called a pointless distinction earlier is not pointless, because of the ratios of hardware to electricity. I think I almost had a good point.

It doesn't matter if it is paid for in hardware or electricity.  Also improvements in efficiency come from Moore's law.  The cost per transistor also falls at the same rate (actually power has been lagging behind transistor cost over the last couple generations).

[clout]

At the end of the day, distributing crypto-equity through fixed algorithms is fundamentally flawed because it does not consider market forces.

That is nonsense.   If the exchange rate rises the market value of the money supply also rises.  If miners receive 1% of the money supply annually it doesn't matter if a BTC is worth $1 or $100,000.  The network isn't distributing "equity" it is providing compensation for securing the network.  The subsidy is merely a bootstrapping mechanism.  In the future users will pay for security and if they pay 1% to PoW miners or 1% PoS stakeholders they are still paying for security.

You didn't refute my point at all.... If miners receive 1% (which is an underestimate, it is more along the lines of 9% at its current rate) when the price of bitcoin goes up then there compensation goes up regardless of added efficiency or security of the network. So if you have a spike from $1 to $100,000 the cost of securing the network went from 10 cents to $1,000. Except we are not longer in that price range which is why the cost of mining is now more than $500 million.

[clout]

Sunny knows that "good enough" and "dead simple" wins.

Thats probably one of the worst mottoes I've ever. You are right to suggest that we shouldn't over complicate things, but "good enough" never wins.

[JoeyD]

Sunny knows that "good enough" and "dead simple" wins.

Thats probably one of the worst mottoes I've ever. You are right to suggest that we shouldn't over complicate things, but "good enough" never wins.

Also if Sunnys involvement in PeerShares is the clincher, you might want to reevaluate your position. In a recent chat discussion between Sunny King and Bytemaster on peercointalk it seems he was not as involved with peershares as you suggested.

Peercoin is not complete in my view and has some remaining issues, before it can be considered a viable pos-solution, most glaring one being their checkpointing requirement.

I really want to like NXT and even though it is not strictly a technical issue with their particular solution, but I do not like their cloak and dagger behavior with the source-code and only releasing it to peer review after they "feel" the world is ready.

Clout do you think adding cpos to the discussion and/or poll would be advisable or is that proposal still too theoretical?

One little addendum regarding the whitepapers by Daniel Larimer, part of the "security model" in his theories is that of many competing chains and free market choice between them, instead of the other seemingly isolated central blockchain models. While I understand the philosophical ideal of wanting security solely through algorithms, I'm not convinced that this mathematical Elysium is achievable in practice, besides it all hinges on market evaluation by humans anyway. So while not being limited to tapos/dpos I do think the free choice between chains implied in those whitepapers (ByteMaster always mentioned the competing chains when I've heard him explain his concepts) is more versatile than other models.

PS
What's up with all the references to the scores on IQ-tests and drawing conclusions from them? To me those things were more about how long can you fight the boredom and keep focus. Most of the ones I had to fill in even used multiple choice, talk about ridiculous. Multiple choice makes it just about impossible to give the wrong answer.

[Brangdon]

Of course PoS isn't free either.  If Peercoin was worth as much as BTC (value of money supply) then its 1% inflation rate would be about $70M a year.

PoS does not intrinsically need inflation. Nxt is a PoS coin that has no inflation. There's no block reward, just fees, so the number of coins is constant.

The cost of mining from the perspective of the bitcoin network is the block reward provided to miners. None of the costs or efficiencies within the process of mining matter because the network is incapable of taking those variables into account. The cost of security increases with value of bitcoin. If the value of bitcoin increases proportionately more than given decreases in block reward then there is a net increase in the cost of securing the network.

That's true while the block reward dominates. As the block reward reduces, and transaction volume increases, transaction fees will come to dominate in Bitcoin. The cost of security will be equal to the sum of fees, and the community will get greater or lesser security by paying higher or lower fees.

This is why I care about Proof of Stake. It seems to me that either Bitcoin fees will end up very high, to pay the miners for the work they do securing the network; or else the fees will end up comparatively low, and the network will be insecure. If PoS works (a big "if", I concede), then it's inherently more efficient and will require lower fees for the same security, and that could lead users to shift over to it in the long run. (Much as lower fees is often cited as a reason for merchants to prefer Bitcoin over credit cards.)

[Brangdon]

Agreed.  It just important to understand efficiency can't reduce cost only improve security (by eliminating the potential for an attacker to reduce cost by using more efficient tech).

Although there's no improvement in security in the long run. When someone designs an improved ASIC, the network becomes vulnerable to the danger that an attacker will deploy the new technology first. As the new ASIC is adopted by miners, we return to the status quo. The miners need to keep upgrading, just to stay in the same place both with regard to their revenues, and with regard to network security. As long as technology keeps improving, miners will need to pay for new hardware, and the community will have to pay for that hardware, as well as their electricity costs, either through block-reward inflation or through fees.

[sumantso]

Isn't another issue is you need your wallet open to "forge/mint"...
That seems bad for security.  Not sure if that's an issue
with just NXT, or applies to Peercoin and DPOS as well.

It is an issue with peercoin not dpos.

In DPoS, you vote when sending in a transaction. You don't have to keep your wallet open.

You should keep your client online, though as it monitors the network and alerts you in case any delegate is misbehaving.

[bytemaster]

The new DPOS from Bitshares sounds very carefully thought out, and promises to be faster - its not released yet though. I hope its fairer than PPC.

To me it sounds almost ideal, and I am very excited to see it work. I hope there is no huge overlooked flaw in it.

The huge overlooked flaw is that it's not really distributed consensus.

how is it not?

Well, it could be argued that it is. 

But to me its not a good system because you have to trust the
supernode you voted for, etc.  You get into the whole trust
thing, almost like voting for politicians to represent your interests.
I would much prefer the whole thing be trustless and governed
by math and protocols like Bitcoin.

Bitcoin uses vote by hashpower and thus everyone is trusting less than 10 super nodes (mining pools).  Try to look at the true nature of what these systems are doing rather than at the marketing spin the BTC community has put on things.

With bitcoin you cannot vote a miner or pool out.  The coin holders have no say at all and only a super minority actually mine bitcoin these days.   

The system is still governed by math, it just operates as 99 small mining pools where the users can redirect their hash power at will and where all mining pools have equal say and take turns.   

[jonald_fyookball]

Of course PoS isn't free either.  If Peercoin was worth as much as BTC (value of money supply) then its 1% inflation rate would be about $70M a year.

PoS does not intrinsically need inflation. Nxt is a PoS coin that has no inflation. There's no block reward, just fees, so the number of coins is constant.

The cost of mining from the perspective of the bitcoin network is the block reward provided to miners. None of the costs or efficiencies within the process of mining matter because the network is incapable of taking those variables into account. The cost of security increases with value of bitcoin. If the value of bitcoin increases proportionately more than given decreases in block reward then there is a net increase in the cost of securing the network.

That's true while the block reward dominates. As the block reward reduces, and transaction volume increases, transaction fees will come to dominate in Bitcoin. The cost of security will be equal to the sum of fees, and the community will get greater or lesser security by paying higher or lower fees.

This is why I care about Proof of Stake. It seems to me that either Bitcoin fees will end up very high, to pay the miners for the work they do securing the network; or else the fees will end up comparatively low, and the network will be insecure. If PoS works (a big "if", I concede), then it's inherently more efficient and will require lower fees for the same security, and that could lead users to shift over to it in the long run. (Much as lower fees is often cited as a reason for merchants to prefer Bitcoin over credit cards.)

20 years from now, daily reward is still 112 blocks...should be more than enough to keep fees low.

[bytemaster]

Isn't another issue is you need your wallet open to "forge/mint"...
That seems bad for security.  Not sure if that's an issue
with just NXT, or applies to Peercoin and DPOS as well.

Nxt doesn't require you have to your wallet open to forge.

Website says this

What do I need to start forging?
A client and some NXT. Once you have unlocked your wallet with your passphrase you are able to forge.

That's one way to forge, but there are others. Specifically, you can lease your forging power to a pool, then it will forge on your behalf and you don't need to keep your wallet open.

After talking with the Nxt developers it is fairly clear that Nxt and DPOS are very similar.  They both deterministically select the next block producer, operate on a fixed absolute block production schedule and support DELEGATING your stake to others.  

The difference lies in how the delegated stake is managed.  In DPOS everyone is delegating at all times and all clients observe behavior.  In DPOS it is possible to vote AGAINST a bad actor and thus nullify their own votes.   This prevents someone from buying up 10% or more of the network and then abusing their 10% in some way.  

Where Nxt has block production proportional to balance, DPOS is 1 block per delegate with a changing slate of delegates that are voted on proportional to stake.   There are limits that prevent concentration in one delegate and the wallets automatically rebalance votes from delegates that get too much.  

Lastly the difference is that all shareholders are paid from dividends without having to do anything.  The delegates take a small cut of the transaction fees.  

I think that Nxt / DPOS will be the future as Peercoin has too many issues.

[valarmg]

Blackcoin have announced that they are about to release an improved PoS. Details are still unclear, so not sure if it will solve the issues that the Peercoin model currently has.

[DeathAndTaxes]

At the end of the day, distributing crypto-equity through fixed algorithms is fundamentally flawed because it does not consider market forces.

That is nonsense.   If the exchange rate rises the market value of the money supply also rises.  If miners receive 1% of the money supply annually it doesn't matter if a BTC is worth $1 or $100,000.  The network isn't distributing "equity" it is providing compensation for securing the network.  The subsidy is merely a bootstrapping mechanism.  In the future users will pay for security and if they pay 1% to PoW miners or 1% PoS stakeholders they are still paying for security.

You didn't refute my point at all.... If miners receive 1% (which is an underestimate, it is more along the lines of 9% at its current rate) when the price of bitcoin goes up then there compensation goes up regardless of added efficiency or security of the network. So if you have a spike from $1 to $100,000 the cost of securing the network went from 10 cents to $1,000. Except we are not longer in that price range which is why the cost of mining is now more than $500 million.

The same is true with PoS.  Yes you have proven that 1% of a large number is larger than 1% of a small number.

[toast]

DPOS and NXT are practically identical, the only difference is which mechanism is used to pick block producers (both are essentially by stake-vote)

[Brangdon]

20 years from now, daily reward is still 112 blocks...should be more than enough to keep fees low.

It'll be 3% of what it is now. We can hope that 1 BTC doubles in real-world value at least every 4 years to compensate, and it might for a while, but that rate of increase can't continue forever. I am thinking in longer timescales than 20 years here.

[achimsmile]

I do not like their cloak and dagger behavior with the source-code and only releasing it to peer review after they "feel" the world is ready.

https://bitbucket.org/JeanLucPicard/nxt/src

Care to elaborate?