Proof of Stake Velocity: Building the Social Currency of the Digital Age - time to iterate ethereum?

Gerd_HG

Reddcoin implements a new algorithm Proof of Stake Velocity (PoSV) as an alternative to Proof of Work (PoW). Is this an useful algorithm for an iteration of Ethereum? What is your opinion?


*AProof of Stake Velocity*: *Building the Social Currency of the Digital Age*

by
Larry Ren

"Proof of Stake Velocity (PoSV) is proposed as an alternative to Proof of Work (PoW) and Proof of Stake (PoS) to secure the peer-to-peer network and confirm transactions of Reddcoin, a cryptocurrency created specifi- cally to facilitate social interactions in the digital age. PoSV is designed to encourage both ownership (Stake) and activity (Velocity) which directly correspond to the two main functions of Reddcoin as a real currency: store of value and medium of exchange. Reddcoin can also function as the unit of account in heterogeneous social context. The technological aspects of PoSV are presented after a detailed review of existing designs. The eco- nomic aspects of Reddcoin are then analysed. Finally the unique position of Reddcoin as a digital social currency in the competitive landscape of cryptocurrencies is discussed."

http://bit.ly/VSlfG6


http://bit.ly/VUZy82

StephanTual

Since we haven't selected a PoW/PoS or Hybrid yet, it's too early to 'iterate' on anything.

Gerd_HG

This was more meant as a provocative question to start a discussion about what people think about the Proof of Stake Velocity (PoSV) concept

Jasper

Afaik PoSV is basicaly PoS, but you dont let the coin age count for to much. Sublinear increase of 'mining power' with coin age. And you let any transaction also increase 'mining power'. Both sound good to me. Note that coin age is kindah handy so you can turn your computer off now and then and not equally lose mining ability. Hmm, you have 'simulated mining power' P(Amount, t, Coinage)∝probability and reward R(Coinage), where Coinage=∫A(t)dt You ultimately always get the same if Reward∝Coinage. The probabilty can then be linear with the amount, and so you cant really attack by aging your coins. Might want to actually promote mining though, for instance by ModifiedCoinage=∫f(t_now-t)⋅A(t)dt with f some decreasing function, so that if you wait longer, the 'older coinage' devaluates. For most f this will be a pita to calculate. Something like ModifiedCoinage(t+1) = g⋅ModifiedCoinage(t) + Amount(t), only requires one additional parameter to keep track of coinage, and depreciates older coinage. That corresponds to ModifiedCoinage(t) =∑Amount(t')⋅g^^(t-t').

Issues with this though, might still want to increase the probability with coinage aswel as the amount, otherwise actually getting it might take too long, and the falloff makes mining sublinear. Any factor over the amount that the coinage adds to the 'mining power' is the same factor less that you need for an attack, just by waiting.

How to use the presence of transactions gives yet more freedom. But the transaction costs are there for a reason, might not be wise to in-effect re-zero them retroactively. So i dont get that, should re-read that bit of PoSV.. (dont think my memory is serving well)

And you also dont want to need to decrypt your wallet to mine. Allowing you to delegate the PoS might help there. Delegating to yourself(i.e. a second wallet you have) to not have to decrypt the wallet, but if you trust them, you could delegate to others to not need to run your computer all the time, and possibly people could in effect mine as a group for better probabilities. Which is a nice option, but the issue with that is that if people do it wrong, could centralize. The group mining thing might be doable with each party signing.(sidestepping the probability and centralization issue, but not the decrypting issue)

PoS has the thing where if 51% delegate to pubkeys owned by that person at one point in time, that person could rebuild any history from that point. I'd like some PoW with that PoS..

This post has gotten longer than i thought, also excuse that it is a bit.. brainstormy.(.. and i havent read the bitshares stuff enough)

Gerd_HG

great comments @Jasper‌ - thanks