Bitcoin Forum

Hello,
anyone knows if it its possible to cancel an unconfirmed transaction using electrum that has replace by fee enabled?

since you are a hero member i guess you should know by now that bitcoin transaction cannot be cancelled once they enter the network. i assume that you didn't pay enough fee and your transaction got stuck now right ? have you tried doing the child pay for parents way ? google it it may help solve your problem

You are wrong. Transaction can be cancelled if it doesn't have any confirmations.
I know it can be done using bitcoin core. I am asking if it can be done using electrum.

what really ?!! where the heck did you heard that ? so you mean that if a transaction is broadcasted to the network it can be cancelled ? are you serious right now

I am answering in the same (very unfriendly) way you did:
Go google double spend. Since you are a "Full member" you should know that.

double spending  is something.... and cancelling a transaction is something else !
and for your record when you double spend a transaction it wont get cancelled, it's actually the opposite as the transaction will go though  ;)

"For your record": Double spend to yourself == tx cancelled.

well that's a new thing, i never heard of it before to be honest. so if you are right than i apologize  for this

hope your problem will be solved, peace :)

anyway check this link it may help:
https://www.reddit.com/r/BitcoinBeginners/comments/6fki5j/eli5_how_to_doublespend_with_electrum/

This is not entirely accurate. You CANNOT cancel a transaction. What you can ATTEMPT to do, is send another transaction that uses the same inputs but spends them in a different way (aka "Double Spend") and hope that the new transaction gets confirmed before the first transaction... but you can't simply make the first transaction "disappear".

There is NO guarantee that a "double spend" will be successful or even accepted by the network. Most nodes will refuse to relay a "double spend", you'll get "mempool conflict"-type errors because they'll already have your original transaction spending the inputs and will either reject or refuse to relay your second transaction.

As for being able to "cancel" a transaction in Bitcoin Core... are you referring to "Abandon Transaction"?  That does NOT cancel a transaction. All it does it remove it from your local wallet and local mempool and stops (re)broadcasting the transaction. However it does not remove the transaction from the mempools of other nodes, so there is still a chance that your transaction will get confirmed.


ONLY if the double spend gets confirmed before the original transaction. Again, there is NO guarantee that this will happen (or that the network will even accept or propagate your "double spend" transaction).

That's exactly what i was referring to

You can double-spend your input and replace your unconfirmed tx with another one with the same input. Read more here https://eprint.iacr.org/2017/394.pdf

I'm not disputing that you can ATTEMPT to make a double-spend... what I'm saying is that it is NOT a trivial task... and is NOT guaranteed to work.

It isn't like you simply create a new transaction, broadcast it and your old transaction simply disappears. It just doesn't work that way.

You have NO control over the mempools of other Nodes... and which transaction they choose to keep, reject or relay to other nodes... or which nodes they choose to relay to. If ALL the other nodes on the Bitcoin network choose to ignore your new transaction, then it doesn't matter what you do unless you are mining and can include (or you have a way to get a miner to include) your new transaction in a block.

Likewise, even if say half of the nodes accept your new transaction, there is NO guarantee that a miner who receives transaction information from these nodes will include your transaction in a block before a miner that has your original transaction.

I didn't say that it's trivial, but it's still possible. I agree with you that the nodes' behavior regarding such a move is unpredictable.
Personally I haven't attempted to execute such a thing, but I'm interested to see the results of someone who did.
After all, replace by fee is essentially a double-spend, you only replace (decrease) the output amount (thus increasing the fee).
I'm a bit busy these days but I can set up an experiment later next week using bitcoin core and see what happens.
I will post tx data here so we can all watch the show :)
In the meanwhile, if someone has a good answer to my original question, please post (although I think that only bitcoin core can do these stuff)

The behavior is not really "unpredictable". In the end its just simply coded. And with enough observation you can estimate how different clients
react to different kind of unusual transactions.


Replace-by-fee isn't really a double spend. A double spending TX is using already used inputs, creating a conflicting transaction.
Where to properly execute a replace-by-fee you need to have a flag set in the first transaction, which indicates that the fee might be changed in future.

Depends on the individual node's feature set.. Not all nodes support RBF and/or recognise "RBF" as valid... so these nodes will flag the "2nd" RBF transaction as an attempted double spend.

Nodes that support RBF, are designed to drop the first transaction when they receive the 2nd RBF transaction that meets the RBF conditions.

I have gone through the topic and Even I don't know how to do that also google will not help you much about this telling this from my past exp. on google you will find many misleading posts related to double spend.

I think someone should explain everthing about it in a topic.

That's why I intend to actually try this in bitcoin core and see the results.