A cryptocurrency with many consensus methods to avoid centralization

[d5000]

One of the challenges for cryptocurrencies is centralization, which makes possible 51% and other attacks.

Most consensus methods (Proof of Work, Proof of Stake etc.) tend to some form of centralization. For example, in Bitcoin and other Proof of work-based currencies, large mining pools have advantages over smaller miners. In Proof of stake currencies, the same is true for large "stakers" or holders.

Now you can say that the reward mechanism of every consensus algorithm tends to benefit certain groups of people:

- Proof of work: Benefits owners of large mining farms.
- Proof of stake: Benefits participants with large amounts of the currency.
- Proof of importance (e.g. NEM): Benefits participants which hold and actively transfer large amounts, mostly services like online wallets and exchanges.
- Proof of burn (e.g. Slimcoin): Like Proof of Stake, it tends to benefit people with large amounts of currency, but involves more risk for the participants and the amount of time to recover investments is larger than in most PoS variants.
- Proof of disk capacity (e.g. Burst): Like Proof of work, it benefits owners of large amounts of hardware, but a different kind of hardware than PoW itself.
- "Proof of time connected" (e.g. Timekoin): Benefits people who maintain a stable node with few interruptions.

Maybe there are more, but these are the ones I know.

Now - couldn't it be a way to lower the centralization degree if you combine many of these algorithms in one single cryptocurrency? It would not stop centralization totally. But there would be not a single group, but various groups benefitting from the block rewards. So the game theory could be more complex and attacks could be more difficult.

Sure, there may be users that could try to participate in various mechanisms (e.g. holding large amounts for Proof of Stake and mining (PoW) at the same time) but that should be more challenging than simply buy large amounts of mining hardware or simply staking large amounts of coins.

There are some hybrid cryptocurrencies out there which use more than one method (the best-known being probably Peercoin and Decred) but to my knowledge no altcoin uses more than two or three of these methods. Obviously, combining more of these methods in one single cryptocurrency, without security holes, could be a complex coding challenge.

What do you think, could a "many-consensus-methods"-cryptocurrency be an interesting experiment or would it be an innecessary effort?

[Snail2]

If you can encourage spending - therefore wider distribution - instead of hoarding then yes, that would be interesting and a valuable feature. Perhaps demurrage (or negative PoS "earnings") on long held coins (like freicoin), combined with "Proof of time connected" to reward stable nodes and PoW for transaction processing. I'd also put a Proof of Service into it, what would reward real word services according to customer feedback and also could be a sort of escrow service.

[d5000]

If you can encourage spending - therefore wider distribution - instead of hoarding then yes, that would be interesting and a valuable feature. Perhaps demurrage (or negative PoS "earnings") on long held coins (like freicoin), combined with "Proof of time connected" to reward stable nodes and PoW for transaction processing.

Yes, one of the goals of such a hybrid design would be to encourage actual usage instead of pure speculation, but also better security, and even possibly price stability because with a more even distribution there could be less manipulation by large whales.

Demurrage is interesting and has already been discussed in some old threads about "stable coins", but the problem is that a "demurrage coin" would probably not have the level of acceptance a non-demurrage coin can have. Almost the same thing you can achieve with a slightly inflationary design (5% inflation instead of 5% demurrage), because people tend to ignore supply inflation (as in BTC and most other coins) but fear demurrage.

Nevertheless some years ago I proposed  a combination of proof of stake with "locked deposits" (as in Decred) and demurrage: in this design you would get rewards when you lock your coins for some time and "stake" them, but be punished by demurrage if you hold coins without having your client participating in the network. That also could be an element of such an hybrid design.

Timekoin's "proof of time connected" is really interesting because it rewards stable nodes and punishes short-time involvement. It's difficult to implement in a Bitcoin-based coin, I know there was a discussion in Peercoin but it would have led to blockchain bloat because in each block, the connected peers would have been recorded, leading to very large blocks. In Timekoin the whole block processing process is different, that's probably why there this was not a problem, but I don't know the details of the implementation.

I'd also put a Proof of Service into it, what would reward real word services according to customer feedback and also could be a sort of escrow service.

I have already thought about a "Proof of service", but I doubt it would be possible, it almost certainly can be gamed by fake businesses and fake clients. NEM's Proof of Importance is the closest try I know, it calculates an "importance score" based on transactions and stake. So "hodling-only" accounts would achieve a lower score than typical accounts of crypto businesses (like exchanges, casinos or online wallets) that not only hold large amounts of coins but also move them.

[h3rlihy]

Certainly would be interesting to see something a little more 'hybrid' that discouraged centralisation. I like the idea of actually transacting the coin being rewarding.

[TPTB_need_war]

- Proof of work: Benefits owners of large mining farms.
- Proof of stake: Benefits participants with large amounts of the currency.
- Proof of importance (e.g. NEM): Benefits participants which hold and actively transfer large amounts, mostly services like online wallets and exchanges.
- Proof of burn (e.g. Slimcoin): Like Proof of Stake, it tends to benefit people with large amounts of currency, but involves more risk for the participants and the amount of time to recover investments is larger than in most PoS variants.
- Proof of disk capacity (e.g. Burst): Like Proof of work, it benefits owners of large amounts of hardware, but a different kind of hardware than PoW itself.
- "Proof of time connected" (e.g. Timekoin): Benefits people who maintain a stable node with few interruptions.

Maybe there are more, but these are the ones I know.

Now - couldn't it be a way to lower the centralization degree if you combine many of these algorithms in one single cryptocurrency? It would not stop centralization totally. But there would be not a single group, but various groups benefitting from the block rewards. So the game theory could be more complex and attacks could be more difficult.

There is only one solution and I already outlined the design in my decentralization thread and in the thread where I explained Satoshi didn't solve the Byzantine Generals Problem:

UNPROFITABLE Proof of Work: Benefits the users of the currency who mine at a loss because they are forced to attach a PoW to every transaction. Note there are numerous details to making this design work, and I am not going to explain them all now. Some of the details were explained already in the linked thread.

[d5000]

@TPTB_need_war: I had already read some of your posts about this topic before starting this thread and perhaps this proposal is a bit influenced by it. You basically say that all of the current algorithms or "consensus methods" tend to centralization.

If we can't avoid centralization, at least it may be a significant improvement to have a cryptocurrency that tends not to a monopoly, but to an oligopoly - centralization on different groups.

A cryptocurrency featuring multiple different algorithms (not only PoW  mining algorithms!) would probably do the trick: Each algorithm benefits disproportionately large players of a different type: holders, miners, services etc. like I'm outlining above.

That is based on the assumption that it is most profitable for a cryptocurrency business to specialize in one type of service (or at least, in related services that need the same type of knowledge). A mining business (benefitted by PoW) and a cryptocurrency exchange (benefitted by PoS and PoI) would probably not be run by the same people as they require a different kind of knowledge. The actual economy relies on exactly that assumption that specialization is better than generalization. But if the assumption is false, obviously a hybrid cryptocurrency like I'm proposing would not solve the problem.

I will read a bit about your "unprofitable PoW" proposal, as it sounds interesting.