If someone sends me 51% of all the peercoins in the world, I would wait more then 1 confirmation before handing over my briefcase with USD. Would I wait 6 confirmations?
What I would do prior to the transaction, is to create a program that finds out how many coin days were consumed in the transaction. Then based on this information, the program would extrapolate into the future how many confirmations I would have to wait before the risk of the sender of the coins would be able to pull of a double spend, approaches zero.
Since we don’t know how many coin days will be consumed in the future, the program would update this information continuously and eventually it would tell me that it is safe to hand over the cash.
Perhaps this program could be patched into Peercoin? Perhaps make it into the API so that exchanges could query it? Perhaps it would then be possible to even integrate this feature into trust-less p2p decentralized exchanges?
This is an interesting idea which I hope gets more discussion; however does it only apply in the highly unusual case of purchasing 51% of total Peercoins? It seems that for any significantly smaller amounts, one never has any certainty that the seller of the Peercoins does not also control some other hidden address with a much greater coin-age capable of accomplishing a viable alternative chain.
Right. One should probably not track ones own transaction if its small and instead focus on the minters transaction and the difficulty here is that we can’t know if the attacker has split his wallet into more then one.
However, the real danger for the network in total, is if a whole lot of coin days are being consumed (in order to later overtake the network using a hidden chain). In order for the double spend to be profitable, there has to be a buyer. It will be very much in the interest for the buyer of the coins to look for these things, since the buyer knows how many coins are being purchased (even if they arrive from different wallets).
Then again… we could potentially also have an attacker shorting peercoins on a derivatives market using leveraged futures, options and that kind of stuff. In this case the attacker could potentially spend the coins to himself, do the double spend, take the financial loss in the exchange rate and be compensated for the asymmetrical profit in the short position. Even if the attacker choose to not go public and complain that his own coins has been double spent - the other transactors certainly will.
My proposal is far from full proof, but I think its one step forward away from the 6 block confirmation. In my opinion, exchanges should at least have 48 confirmations. It would take 8 hours to confirm, would be totally improbable to double spend against that I think. Peercoin was never design for small quick transactions, it was designed to be durable storage of value. And how often does one have to deposit and withdraw from exchanges really? Usually there’s a steep fee associated with it too.