Bitcoin Forum

And if so, is there a list of people that have gained from double spending and amounts that they were successful at double spending?

I am just wondering because lots of people talk about double spending being a major problem, yet I don't really see any cases of it.  Just wondering if I missed them.

No, the circumstances under which double-spends are theoretically possible are well-understood; a successful double-spend under such a circumstance wouldn't be particularly meaningful.

For example, a $10,000-equivalent double-spend was successfully issued during the March 2013 blockchain fork event: https://bitcointalk.org/index.php?topic=152348.0
The price rallied significantly in the subsequent month.

Sure. Go ahead and accept BTC for payment and accept zero-conf transactions. Sell digital goods like Steam codes where the exchange is instant.

It doesn't happen very often because the market regularly prunes stupid people out.

ETA: No cheating by having the service check against the tx queue for double-spends. :P

double spend by a 51% attacker doing a multi block reorg would be confidence shattering.

double spend by a hacker on a 0 confirmation transaction, not so much.

No. I'm not an overpowered guy that controls and/or can fool trusted bitcoin nodes.

There was a fork in the blockchain back in March 2013 I think, due to incompatibility between new and old versions of Bitcoin QT. During that period, I seem to remember there was a successful double spend against an exchange. The situation was dealt with very quickly and I think the double spend was reversed.

As far as I'm aware, there haven't been any major successful double spends of Bitcoin since then.

I remember reading it when I first started with bitcoin. Didn't understand much of it at the time.


The bitcoin nodes are so well linked. I think double spending on a healthy bitcoin network is extremely difficult. There is a page at blockchain.info on double spend attempts. I think none of them are successful:
https://blockchain.info/double-spends

That has to do with 51% attacks right? I haven't heard of any cases either. I did some research and there was a successful double-spend on 2013 (of course people already told you that lol) other than that nothing.
I don't really get the page you linked though, on some of them it says "spent" which means the transaction was successful? sorry I'm newb :P

there have been no double spends of bitcoin.. EVER

there have been instances that services have given away DIFFERENT currencies multiple times, but there is no actual double spend of bitcoin. the fault is of service providers not interpreting transactions appropriately.

I don't think there have ever been double spending on Bitcoin. There has been case with altcoins which are easy to attack due to their low hashrate.

If Bitcoin is going to be attacked its going to be a coalition of big pools for whatever reasons. In that case everyone will get disrupted and it will be a huge event, and possibilities like this exist in all spheres of life.

There have been several well-documented attempts but never successfully as far as I am aware

Question is: how could a Joe like me that has  0 hash power be able to double spend even with 0 confirmation transactions that sooner or later will be identified?

i was thinking about that too and found on google how to do it though i quit because i couldnt find any gambling sites that accept 0 confirms deposits and payouts instantly

Well you can't technically make a double spent. You could however 'double spend' funds from a merchant in case of a bug in their software or something.

i don't think anyone will tell this even if he has done it successfully
many people do this, or scam others with fake transactions

Define "double spent", please. There are different definitions of this term and depending on what you are asking about the answer may be YES or NO.

If you are talking about 0 confirmation transactions then it's rather easy to double spend. Send a transaction with no fee, then roll back your transactions and send the same transaction with a higher fee so it gets included into the Blockchain first, rejecting the first transaction. This only applies to accepting 0 confirmation transactions though, you would need a lot of mining power to try and get transactions with one or more confirmation through.

It just so happened that my first transaction I sent with Bitcoin was with no fee and was stuck for 4-5 days trying to get into BTC-e with no confirmations and so had to do something. The first transaction was left in perpetual limbo as 0 confirmations so technically it wasn't a double spend.

I think in the old times yes. Blockchain actually has a list of all double spents that have occured till now.

^^ This.  Very much this. ^^

A true double spend (where a previous transaction output is used as an input in two different transactions that are both permanently recorded in the blockchain) is impossible.  It never has happened. It never will happen.

However, people that don't understand the technical detais of how bitcoin works like to use the words "double spend" to describe a variety of scams that allow an attacker to fool their target into believing that they've received bitcoins that they never actually received.

So, if you are asking about a true double spend, then no, it hasn't ever happened.  It never will.

If you are asking about a scam where someone is tricked into believing that they've received bitcoins that they haven't due to the target's own misunderstanding about how bitcoin works, then yes.  That has happened many times.

double spant will cost so much money , no borning people will do this thing for bitcoin
but i think this will happen in altcoin

I think most wallets won't let you try , so would need a slightly modified version of a wallet that would allow you to broadcast a (non valid ) transaction that has already spent funds. 

What would happen if you and your wife had cloned wallets and visited two separate merchants who accepted bitcoin and made purchases using synchronized watches?

If:

• You both were connected to different peers
• You created the transaction so that both wallets spent the same unspent output
• You could time it so that both of your wallets transmitted your transaction at the exact same time

Then each merchant would have a 50% chance of seeing the unconfirmed transaction at all.  In other words, each merchant would have a 50% chance of seeing a transaction with 0 confirmations that sent bitcoins to their address, and a 50% chance that they'd be sitting there waiting for you to pay them, because they don't see the transaction at all.

Eventually 1 of the transactions would be confirmed.  At that time, the competing transaction would cease to exist.  It would be dropped from the memory pools of all peers.  The transaction that did get confirmed would be seen by the merchant that the payment was made to with 1 confirmation.

So, if you got lucky and both merchants saw their own unconfirmed transaction, and if the merchant allowed you to leave with the merchandise while there were 0 confirmations, then you would have committed "shoplifting" (since you left with merchandise that you didn't pay for) and they would call the police and report the theft.  Then the police would review their security cameras, get a description of you from any witnesses, and the shop would be watching for you to return.  If the shop is accepting transactions with 0 confirmations, and they aren't completely stupid, then they will require that you provide identifying information (such as a government issued photo ID, or a "frequent customer" card issued by the shop) along with the purchase that will have been recorded with the transaction.  This identifying information will also be turned over to the police.

Additionally, any merchant that is willing to accept transactions with 0 confirmations really should be running custom software that will maintain connections to a very large number of peers.  This custom software should alert them if it received conflicting transactions from different peers.  This alert should occur within a second or so of you both sending your transactions.  As such, both merchants will immediately know that you are attempting a scam, and that they should not allow you to leave with the merchandise until they see at least 1 confirmation on the transaction.

Essentially, a transaction with 0 confirmations is a lot like a personal check.  If you have $20 in your bank account, and you and your wife in separate shops each write a check for $18, what will happen?  Assuming that the bank doesn't allow you to overdraw your account, one check will "bounce".  You'll have committed fraud, or theft, and the merchant can press charges in court if necessary. A merchant should treat a transaction with 0 confirmations with the same scrutiny that they treat a personal check.  Of course, since all transactions in bitcoin are completely public, the merchant has extra tools available to them to reduce their risk with bitcoin transactions that they don't have with personal checks (such as monitoring peers for conflicting transactions, and the ability to know for certain that the funds being spent actually exist).

Yes i agree. Double spend is a crime and crime never pays off.

+1.  We don't hear much of this idea and it is a great answer to complaints about block times.

May be relevant - GHash.IO and double-spending against BetCoin Dice (https://bitcointalk.org/index.php?topic=327767.0)

I have happily accepted transactions in excess of $5000 with 0 confirmations when it met the following conditions:

• I had reliable identifying information about the sender
• I was able to see that it was a standard transaction
• The transaction included a reasonable transaction fee
• I was able to determine to my satisfaction that there weren't any competing transactions with identifiable propagation

I had no concern at all.  Any merchant can do the same.

Blockchain.info has a flag on one of my addresses accusing it of double spending. However, I think that is based on blockchain.info's flakey timestamping, registering an output several minutes before an input or something.

I see you like gambling... :)

You'd be gambling, Mr Hamilton here is 2nd Dan in Blockchainfu.

Life is a gamble.  Every time we step out our front door, we are gambling and taking a chance that we might die.

All we can do is mitigate the risks to a tolerable level.

The four conditions that I place I accepting a transaction with 0 confirmations are sufficient for bringing the risks to my own personal tolerable level.

The increased reduction in risk by getting confirmations is negligible if those four conditions are already met.

I'm assuming that English isn't your first language?  I'm having difficulty understanding what you are trying to say in this comment.  Perhaps if you post it in your native language, I can decipher it a bit better?

I sp34ks interweb, anything-fu = comparison to a martial art, masters of martial arts have black belts of rank 1st dan, 2nd dan, 3rd dan etc...

Not to nitpick but dan typically applies to Japanese martial arts. Kung Fu uses a different term for black belt ranks.

Traditional Shaolin Chinese Kung-Fu uses sashes (not belts).

Many people including Mt Gox have double spent bitcoins :o

Its not easy. You will have to set up two transaction points spatially separated and do transaction from them quickly so that the transaction from one is not yet propagated to the nodes close by the other. As soon as the transaction gets to most nodes, they will reject any new transactions.

Its not that hard if you have time because you dont want to screw anyone. Nodes tend to remove transactions from their mempool. Not sure if there is a timeout, but nodes get restarted and rebooted all the time removing your 1st unconfirmed transactions from their memory. They will now accept your 2nd transactions and broadcast it. I had a similar "problem" where I tried to combine 4 low value inputs into a single input without paying a fee. The TX had an overall BTC value of 0.0022 and the inputs had been several hundred blocks old, but of course this was not enough to get a high enough priority. Thus I had to remove the transaction from my bitcoin core with pywallet and create a new one - a double spend - which paid a fee. The 2nd transaction propagated fast and was confirmed within the next few blocks.

It is that easy with 0 confirmations and no fee, multi-bit will let you set your date back in order to have balance again. I had no real knowledge of the Blockchain when I did it, just that I needed to resend the original transaction with a higher fee so it would be included into miner's transactions before my first transaction.

The first condition is the most critical and makes it much less of a gamble:


As long as you can reliably identify who you're transacting with, you can sue them if they defraud you. When deciding whether to accept 0-confirmation transactions, the question is not, "Can they double-spend?" (of course they can), but rather, "Can I do anything to recover my losses if they do?", and in many cases the answer is "yes, probably".

Double spend is closely linked to the 51% attack issue.
As long as you wait 6 confirmations for your bitcoins, you should be safe from somebody double spending those bitcoins.  :)

so there are cases where a transaction with 1 confirmation was undone? 

from this thread I have learnt that people have double spent on accident, or on purpose trying to scam from sites not taking secondary measures, but these were cases with low or no transaction fees so the network hadn't processed them. 

It means intentionally send a transaction of an low priority and hope it does not get confirmed. This kind of attack can be easily prevented by accepting 0 confirm transactions only for a sufficient high priority. I think thats what the merchants do.

That won't work, the nodes will reject it if you do soon after the first transaction. You will have to wait 24-48 hours, try rolling back and then sending it again.

There have been attacks on buyers who accepted zero-confirmation transactions. This happened most notoriously with SatoshiDICE. At one time they paid out immediately on wins of zero-conf bets. Someone realized they could undo a losing bet by attempting a double-spend.  This didn't generate extra Bitcoins, but it invalidated the payment to the betting firm, effectively canceling the previous transaction. With everything anonymous, this could be repeated endlessly.

So, nobody does zero-conf transactions with instant return of anything of non-trivial value any more.

A true double-spend would generate new Bitcoins from nothing. So far, nobody has been able to do that. If they had, it would be recorded in the block chain.

it doesnt make sense

This is not true, nor is your statement about not being able to generate bitcoin out of nothing true.

A double spend is exactly what it's name implies, spending the same input twice. If someone were to double spend an input they would broadcast a transaction to the network, receive something of value in exchange for that transaction and then get a 2nd transaction that spends at least one of the same inputs to get confirmed in the longest blockchain, invalidating the 1st transaction.

When there was still very little transaction volume someone was able to manipulate some part of of a transaction so that the output would be an unlimited amount of bitcoin; they were able to generate billions of bitcoin to an address they choose. Luckily transaction volume was very low at the time so an announcement on bitcointalk was enough to get people to not send any additional transactions until the issue was fixed, the protocol was quickly forked to prevent this "bug" from being exploited while reverting back to the block prior to when the bug was first exploited   

That's the way to look at this. The only possible real-world double-spend is a transaction lanced in order to trick your victim into believing they received their Bitcoins, when the real transaction that is being accepted in the Blockchain goes to an address you control yourself.

I think double spending bitcoins would not be an easy task. however double spending other coins would probably be easier. I tried it a couple times just to see if I could do it, but was never successful :( haha. I just wanted to see if it could happen and since i had the hardware to have over 51% of the network hashrate I thought I would try it. But I never never mined the block I sent so :( pity for me.

Incorrect. If you have a transaction that is already confirmed by the network then you will need to get a miner to mine enough blocks so that they have a longer chain from the time the TX was originally confirmed and to include a conflicting transaction with at least one of the same input

I've never heard of double spending, how often does this actually happen?

Nah, I guess he's right, man! You could issue a transaction to the network to send your BTC to another address of your's and at the same time issue the payment-transaction to your 'victim's' BitcoinQT. They'd now think your transaction paid for your goods.
Another way would be to issue the payment transaction to the network, but then mine the next block with the BTC going to a wallet of your own!
But those things only can happen if you accept 0-confirmation transactions.

You'd first need to figure out how to isolate the victim's Bitcoin-Qt from the network without them realizing it.  Otherwise, they will see the transaction that you issued to another address of yours and they will refuse to accept the transaction that you try to send to their Bitcoin-Qt.

If you send to their Bitcoin-Qt first, then they will relay the transaction to the network, and the network will refuse to relay your transaction that pays to your own address.


Sure, if you have enough hash power. You've still got to find someone that will accept your unconfirmed transaction and that doesn't have identifying information about you.  Even then, you won't be able to reliably pull off the attack unless you have enough hash power to pull off much more profitable attacks against the network (in which case it's silly to be using all that hash power to try and steal $1 hear and there).


And if your victim can't identify you.

Blockchain Info thinks this address has double spends...


" You should be extremely careful when trusting any transactions "

How does one 'be careful'?