4 ms·

So, in practice, what is the probability that a transaction would really be cancelled if the issuer of the payment uses this service approximately two minutes a

by stoev 13y ago

So, in practice, what is the probability that a transaction would really be cancelled if the issuer of the payment uses this service approximately two minutes after initiating the transaction? (I assume that two minutes are a realistic estimate of the time it takes to realise that one has made a mistake, remember this service, search for it online, and fill out the necessary info.)

espringe 13y ago
There's too many variables to give a simple answer. It depends on the current state of the network (the miners mempools), the priority of your transaction, the fee/bytes of your transaction, and what % of the hashing power bitundo has. Some transactions take a few blocks to get confirmed, these are the ones that have a much, much higher likelyhood of being undone.
jere 13y ago
>Q: How likely is it to undo my transaction A: The likelihood is approximately the same as the percentage of mines using bitundo. Right now, this is pretty low. If I understand correctly, the time left doesn't really matter. It should be totally dependent on their share of hashing. If undoing requires them to pick the winning nonce, it should be much like a lottery (your chance at any given time depends more on your share of purchased tickets more than anything else). And if that's the case, this seems like a pretty pointless service. Until they get a large share, it won't work the vast majority of the time. And even when they do get a large share, undoing still won't work most times.
gus_massa 13y ago
Easy back of the envelop calculation: If a block is generated every two minutes in average, you have a 80% probability that the new block is not generated in that two minutes. So if everyone is in this service then the probability of success is 80%. The same idea can be used with a rogue client. Just generate a conflicting transaction. I’m not sure, but if there are conflicting transactions, then most of the miners just pick the first transaction they received. But let’s be optimistic and suppose that they pick at random, so you have a 80%/2=40% that the new conflicting transaction is picked and the original dropped. IIRC some miners are filtering the transactions to process the transactions with bigger fees, even in the case of not conflicting transactions. So perhaps a bigger fee can be used as an incentive to choose the “right” transaction.

by stoev 13y ago

espringe 13y ago

There's too many variables to give a simple answer. It depends on the current state of the network (the miners mempools), the priority of your transaction, the fee/bytes of your transaction, and what % of the hashing power bitundo has. Some transactions take a few blocks to get confirmed, these are the ones that have a much, much higher likelyhood of being undone.

jere 13y ago

>Q: How likely is it to undo my transaction A: The likelihood is approximately the same as the percentage of mines using bitundo. Right now, this is pretty low. If I understand correctly, the time left doesn't really matter. It should be totally dependent on their share of hashing. If undoing requires them to pick the winning nonce, it should be much like a lottery (your chance at any given time depends more on your share of purchased tickets more than anything else). And if that's the case, this seems like a pretty pointless service. Until they get a large share, it won't work the vast majority of the time. And even when they do get a large share, undoing still won't work most times.

gus_massa 13y ago

Easy back of the envelop calculation: If a block is generated every two minutes in average, you have a 80% probability that the new block is not generated in that two minutes. So if everyone is in this service then the probability of success is 80%. The same idea can be used with a rogue client. Just generate a conflicting transaction. I’m not sure, but if there are conflicting transactions, then most of the miners just pick the first transaction they received. But let’s be optimistic and suppose that they pick at random, so you have a 80%/2=40% that the new conflicting transaction is picked and the original dropped. IIRC some miners are filtering the transactions to process the transactions with bigger fees, even in the case of not conflicting transactions. So perhaps a bigger fee can be used as an incentive to choose the “right” transaction.