I read all these discussions about the constant block reward and the protocol changes. This made me think about a new PoS algorithm proposal. I would like to call it “PoS with minting club”.
New protocol features:
- 90x more secure than the current system (an attacker will not be able to collect coindays)
- the same degree of decentralisation as the current system.
- the minting incentives will be changed such that big and small stackholder are encouraged to mint and keep their “minting power online”
- I think this algorithm is compatible with sigmike’s cold minting proposal
The new protocol finds blocks pretty much the same as in the current model. There is this PoS-difficulity, which has to be met by a stack holder in order to release a new block and the longest chain is the main chain. But there are two big differences. At first, stackholder can not collect coindays in order to get a better minting power and secondly, a blockfinder, lets call him bob, does not get its reward instantly. The idea behind the new algorithm is to try to keep bob minting. This could be archive by the following steps:
-After bob has found his new block, he joints the “minting club”. The minting club is made of all the stackholders, who found one of the last x blocks, and some lucky stackholders. I would propose that everyone, who found a block in the last 90 days, should be in the minting club, thus we would choose x=90246.
-bob will receive its annual 1% reward, if and only if the members of the minting club find enough new blocks. I.e. after x blocks, the algorithm (or finder of the block x) looks up, how much minting power was allocated by the minting club during these x blocks, then calculates a specific threshold -a number y- that depends on this minting power and will honor bob with its annual 1% reward, if the members of the minting club found more blocks than this threshold y.
This protocol motivates bob to keep on minting, since he wants the minting club to find more blocks.
Now, let me explain these “lucky stackholder”. If only finders of new blocks would be part of the minting club, the system would be unfair since small stackholder are unlikely to find a block. But in this new protocoll even small stackholder could be treated as if they are a blockfinder:
-Each time a member of the minting club finds a new block, he can choose randomly some stacksholders, the lucky stackholders, which he allows to join the minting club. These lucky stackholder can now get their annual 1% reward by the same process as new blockfinders. Let me specify this:
We need to define some kind of metric d( , ) between a block hash and a peercoin address. Now assume the block “block1” has be found. Then any node, which has a peercoin address “peeradress” such that d(hashof(block1), peeradress)< “constant c”, can send a signed message to the network, which states that it wants to get lucky stackholder. Now assume a member of the minting club finds a new block “block2”. Then this finder is allowed to choose some of these signed messages and put them into his new block “block2”. Now all the stackholder, who find their signed message in the new block, are lucky stackholders. I would propose that the protocol allows the finder of block “block2” to reward himself for each lucky stackholder, which is added by him, by a very small amount. For security reasons, we would only allow small stackholder with many coindays to get lucky stackholders.
Notice, since a stackholder can earn only its annual 1% reward, there is pretty much no incentive to mint on different chains at the same time and there is no incentive to split its own stack in order to get a lucky stackholder.
Problems with this protocol.
- How do we have to choose all these parameters in the protocol? I guess only experiments will tell us.
- The minting club could suffer by some the “tragedy of the common effect”. But I think that the minting club provides an additional incentive that is strong enough. If we provide an easy way to everyone to mint with their own Raspberry Pi in a cold lock minting modus, I am convinced that they will be minting.
- Does not resolves the “nothing at stack problem”.
In the current peercoin protocol only 365246 accounts can get their annual 1% reward. This number can be much higher by the new proposal.
I like this protocol, since there is strong incentive to mint, it is relatively fair and it keeps this high degree of decentralisation of Peercoin, we love so much!
I looking forward to read your criticism, thoughts and proposals for even better protocols :).