"Game theory" free bitcoin question

[MrFreeDragon]

It is the theoretical question (no real bitcoin award) from the game theory.

Suppose, that somebody grants valuable amount of BTC (let's say 10BTC). The private key is published online: for example, a separate thread is created (here on bitcointlalk) with the announcement to publish private key on certain date and time. While the private key is published, everybody (who read the topic) could release the funds. However, we know that replace-by-fee (RBF) transactions are possible. So, if the user 1 makes a spending 9.99BTC with 0.01BTC comission, so another one could make RBF transaction with 9.5BTC and 0.5BTC, and miners will confirm the transaaction with higher fee. The 3d user could make 9BTC spending with 1BTC transaction and so on... The users also could negotiate online making pools, promissing equal shares to each other. However they also could break such negotiations to maximize personal income.

The questions are:
[1] How much will be spent from such 10BTC address and how much will be the transaction fee to miner?
[2] Suppose, that only 4 online anonymous to each other users have the access to disclosed private key. These users also could negotiate online. How much will be spent from 10BTC address and how much will be the transaction fee in this case? 4 users can easily negotiate (let's say to receive 2.49BTC each, transaction fee 0.04BTC), but everybody will still have temptation to replace the transaction with higher fee and higher personal reward (9BTC spending with 1BTC transaction fee for example, where 9BTC is much more than 2.49BTC)

[BigBoom3599]

I think the only solution in this case is that the transacted BTC amount goes to 0 and the mining fee goes to 10BTC right? Every individual party will take a smaller reward with a higher fee over another party getting a larger reward, thus the reward will eventually go to 0 and the fee will keep growing. IMO it doesn't make sense for different parties to come together and form an agreement as there is no way to guarantee an outsider will not outbid them.
Even in the case of [2] there is no trustless way to do this (unless you could form some sort of smart contract like on Ethereum? not sure) and it would not make (economical) sense to not betray another individual.
Interesting thought experiment though. I'd love to see it play out in the real world, anyone have a spare 10 BTC? 

[Red-Apple]

if the private key that could unlock 10 bitcoins worth nearly $100k were known to more than one person there will not be any negotiations between them. there will instead be an arms race to create a final transaction with highest amount of fee and propagate it throughout the entire network before the other(s) do.

you don't really need to theory-craft about this either! there has been incidents where someone accidentally sent funds to an address which has its private key as public knowledge and the arms race began!
like the brainwallets with easy password. the private key on bitcoin wiki page,...

[NeuroticFish]

[1] How much will be spent from such 10BTC address and how much will be the transaction fee to miner?

I think that the miner fee will be the highest miner fee one of those "spenders" will set before the "spender" block  is mined.
Everybody who knows the private key will try to send it to his own wallet.

[2] Suppose, that only 4 online anonymous to each other users have the access to disclosed private key. These users also could negotiate online.

As long as different people have the Private Key, any negotiation attempt has a high chance to be fake (one trying to "scam" the others). And we are back to point [1] 

[MrFreeDragon]

-snip-

[2] Suppose, that only 4 online anonymous to each other users have the access to disclosed private key. These users also could negotiate online.

As long as different people have the Private Key, any negotiation attempt has a high chance to be fake (one trying to "scam" the others). And we are back to point [1]  

Actually all the first answers predict 0 spending with the highest 10BTC miners fee ) However let's look at the 2nd sitation ehere only 4 "players" are involved. Every player understands that it is better to negotiate. In case of negotiation with other 3 players, he could receive 2-2.49BTC. But in case of violation his personal reward tends to 0 BTC. Of course to scam others and receive 3.5-4 times more (9-0.5BTC instead of 2-2.49BTC) is better, but such scam could lead to actually 0BTC reward (because every other could also scam, and in transaction fee race the spending amount become 0). Considering these facts, it is better to make the negotiation between 4 players and receive some agreed reward. Why not?

[NeuroticFish]

Actually all the first answers predict 0 spending with the highest 10BTC miners fee ) However let's look at the 2nd sitation ehere only 4 "players" are involved. Every player understands that it is better to negotiate. In case of negotiation with other 3 players, he could receive 2-2.49BTC. But in case of violation his personal reward tends to 0 BTC. Of course to scam others and receive 3.5-4 times more (9-0.5BTC instead of 2-2.49BTC) is better, but such scam could lead to actually 0BTC reward (because every other could also scam, and in transaction fee race the spending amount become 0). Considering these facts, it is better to make the negotiation between 4 players and receive some agreed reward. Why not?

Would you trust the others into negotiating? Would you have any guarantee that while pretending to accept the negotiation one of the four would try to send with higher fee to a (new) wallet the others don't know?
Imho a fair negotiation is not a valid option.
I expect that if proposed, everybody will agree to split. And then break the agreement. No one would know who broke it.

[ABCbits]

If miner/pool owner know about existence of the thread, then it'll be competition between miners and pool owners.
Whoever managed to mine next block is the winner and the miner get everything.

If the amount of "free Bitcoin" is ridiculous high (e.g. 6x higher than block mining reward), it's possible 51% attack, block withholding attack, re-org attack will happen which make game theory more complex

[squatter]

The private key is published online: for example, a separate thread is created (here on bitcointlalk) with the announcement to publish private key on certain date and time.

[...]

How much will be spent from such 10BTC address and how much will be the transaction fee to miner?

The most likely outcome is that miners will take it for themselves and pay no transaction fee. Which miner/pool gets the additional reward is a matter of hash rate.

Otherwise, this is fundamentally a market mechanism, so it can't be predicted before the fact. It really depends on when the next block is found, because miners will be incentivized to confirm the transaction as long as it pays above market rates. Until that point, people will keep bidding higher and higher fees.

[hello_good_sir]

It really depends on whoever is willing to get the less amount of profit. All of these games are based on greed, and whoever is the least greediest ends up winning.

Imaging someone who only opts for 1BTC and puts 9 BTC as a transaction fee, you could do an RBF there and go for 0.1 BTC and a 9.9BTC fee which would be nuts.

People who know miners and can get them to include their transaction in the pool would also be at an advantage, they might get away with 9.9 BTC.

Interesting post mate.

[hatshepsut93]

However, we know that replace-by-fee (RBF) transactions are possible.

Replace-by-fee transactions are opt-in, you have to specifically make transaction with this flag for it to work. Regular transactions can't be easily replaced, today's nodes are configured to ignore new transactions that conflict the one's that are already in their mempool. So, the answer is, either a miner would mine it for themselves, or the fastest sender will take it. Tx fee will be higher than average, but nothing too crazy.

The most likely outcome is that miners will take it for themselves and pay no transaction fee. Which miner/pool gets the additional reward is a matter of hash rate.

I doubt it's likely, there's only ~10 minutes timeframe to do it, and the person who does it can't be just a regular miner, they have to be the administrator of a mining pool, and they have to quickly reassemble the block they want to mine, and then win the race against the network. And for miners this transaction looks like any other transaction, so they can't automatically detect it on their side.

[franky1]

ok firstly most pools do 'first seen' and then ignore other transactions. this is why CPFP was invented to have a child transaction that spends it. thus making the linked parent be required,

what then happens is that you get a chain of transactions
which instead of RBF
1abc (10btc) -> 1bcd (9.99999900) (on person)
1abc (10btc) -> 1ghi (9.99999800) (competing person)

becomes CPFP
1abc (10btc) -> 1bcd (9.99999900) (one person)
                        1bcd(9.99999900) -> 1cde (9.99999700) (same person)
                                                         1cde (9.99999700) -> 1def (9.99999600) (same person)

where they try as fast as they can to make spending transactions hoping to get one tx more than a competitor to then get the confirm

ofcourse this can go on forever or until the guy sees the first transaction confirms then he know the funds are safe in his posession

[hello_good_sir]

ok firstly most pools do 'first seen' and then ignore other transactions. this is why CPFP was invented to have a child transaction that spends it. thus making the linked parent be required,

what then happens is that you get a chain of transactions
which instead of RBF
1abc (10btc) -> 1bcd (9.99999900) (on person)
1abc (10btc) -> 1ghi (9.99999800) (competing person)

becomes CPFP
1abc (10btc) -> 1bcd (9.99999900) (one person)
                        1bcd(9.99999900) -> 1cde (9.99999700) (same person)
                                                         1cde (9.99999700) -> 1def (9.99999600) (same person)

where they try as fast as they can to make spending transactions hoping to get one tx more than a competitor to then get the confirm

ofcourse this can go on forever or until the guy sees the first transaction confirms then he know the funds are safe in his posession

This does make sense, great explanation.

Another thought popped into my mind though. Would bitcoin transactions get all simulations processed when a block gets mined (eg, block XXXX gets mined, and at the same time 50,000 transactions get processed since they are included in the block), or would there be a delay in processing transactions with the block (a block processes but some transactions get processed faster than the other ones).

If it's simultaneous, wouldn't the first person who sent the transaction, as long as it's involved in the block, it'll be processed first, yeah?

[franky1]

1abc (10btc) -> 1bcd (9.99999900) (one person)
                        1bcd(9.99999900) -> 1cde (9.99999700) (same person)
                                                         1cde (9.99999700) -> 1def (9.99999600) (same person)

This does make sense, great explanation.

Another thought popped into my mind though. Would bitcoin transactions get all simulations processed when a block gets mined (eg, block XXXX gets mined, and at the same time 50,000 transactions get processed since they are included in the block), or would there be a delay in processing transactions with the block (a block processes but some transactions get processed faster than the other ones).

If it's simultaneous, wouldn't the first person who sent the transaction, as long as it's involved in the block, it'll be processed first, yeah?

no the block creators mempool just decides to keep the 1abc->1bcd->1cde->1def as active viable transactions and just puts the
1abc (10btc) -> 1bcd (9.99999900) into a block
then the next block(s) can then contain the other transaction decendants
they do not all need to be in the same block.

plus a block can only fit on average 2500tx so again dont worry about a chain of 50,000 decendant transactions needing to all fit into one block. as long as the first viable transaction thats has an unspent input gets confirmed the funds then are in that persons possession as he is then the only owner of the private key for 1bcd

[squatter]

The most likely outcome is that miners will take it for themselves and pay no transaction fee. Which miner/pool gets the additional reward is a matter of hash rate.

I doubt it's likely, there's only ~10 minutes timeframe to do it, and the person who does it can't be just a regular miner, they have to be the administrator of a mining pool, and they have to quickly reassemble the block they want to mine, and then win the race against the network.

I thought it was obvious I was referring to pools.

The event will have been widely publicized, so including a sweep transaction in the mempool is trivial. At that point, it's effectively the same as 10 BTC being added to the block reward. I think offering 80% of a block's mining subsidy is enough to get the attention of most pool administrators. Maybe not, but it seems to me that miners would take it > 50% of the time. They have a significantly better chance than any non-miner, so it's rationally worth their time.

[pooya87]

ok firstly most pools do 'first seen' and then ignore other transactions. this is why CPFP was invented to have a child transaction that spends it. thus making the linked parent be required,

what then happens is that you get a chain of transactions
which instead of RBF
1abc (10btc) -> 1bcd (9.99999900) (on person)
1abc (10btc) -> 1ghi (9.99999800) (competing person)

becomes CPFP
1abc (10btc) -> 1bcd (9.99999900) (one person)
                        1bcd(9.99999900) -> 1cde (9.99999700) (same person)
                                                         1cde (9.99999700) -> 1def (9.99999600) (same person)

where they try as fast as they can to make spending transactions hoping to get one tx more than a competitor to then get the confirm

ofcourse this can go on forever or until the guy sees the first transaction confirms then he know the funds are safe in his posession

i don't see any point in doing something like this. all the person who is taking the 10 bitcoin has to do is to
- be fastest in getting his transaction to the pool
- pay a high fee compared to the highest fee in the mempool (like 1000 sat/byte when highest is 1 s/b)
- mark the tx as final (max sequences)
when this tx is in the pool's node mempool anything else that is received will be rejected as double spend. CPFP doesn't do anything here. if the pool were to drop such a tx, they don't care whether it has 10 tx after it (in a chain) or 0.

[hello_good_sir]

1abc (10btc) -> 1bcd (9.99999900) (one person)
                        1bcd(9.99999900) -> 1cde (9.99999700) (same person)
                                                         1cde (9.99999700) -> 1def (9.99999600) (same person)

This does make sense, great explanation.

Another thought popped into my mind though. Would bitcoin transactions get all simulations processed when a block gets mined (eg, block XXXX gets mined, and at the same time 50,000 transactions get processed since they are included in the block), or would there be a delay in processing transactions with the block (a block processes but some transactions get processed faster than the other ones).

If it's simultaneous, wouldn't the first person who sent the transaction, as long as it's involved in the block, it'll be processed first, yeah?

no the block creators mempool just decides to keep the 1abc->1bcd->1cde->1def as active viable transactions and just puts the
1abc (10btc) -> 1bcd (9.99999900) into a block
then the next block(s) can then contain the other transaction decendants
they do not all need to be in the same block.

plus a block can only fit on average 2500tx so again dont worry about a chain of 50,000 decendant transactions needing to all fit into one block. as long as the first viable transaction thats has an unspent input gets confirmed the funds then are in that persons possession as he is then the only owner of the private key for 1bcd

Thanks for the explanation Franky! I get it a lot better now - was a bit confused at the start since I've never been much of a technical person.

The most likely outcome is that miners will take it for themselves and pay no transaction fee. Which miner/pool gets the additional reward is a matter of hash rate.

I doubt it's likely, there's only ~10 minutes timeframe to do it, and the person who does it can't be just a regular miner, they have to be the administrator of a mining pool, and they have to quickly reassemble the block they want to mine, and then win the race against the network.

I thought it was obvious I was referring to pools.

The event will have been widely publicized, so including a sweep transaction in the mempool is trivial. At that point, it's effectively the same as 10 BTC being added to the block reward. I think offering 80% of a block's mining subsidy is enough to get the attention of most pool administrators. Maybe not, but it seems to me that miners would take it > 50% of the time. They have a significantly better chance than any non-miner, so it's rationally worth their time.

There are certain apps (there's one that works when you search up confirm BTC transaction) that do this sorta stuff, and the one I'm talking about likely works like how your saying, they charge around 50 percent of the TX but most times it's confirmed in the next block.

Sorry guys. Another question: what is marking a TX as final? Aren't all transactions able to be doublespent/RBF as long as you have the right software?

[BigBoom3599]

--snip--

Sorry guys. Another question: what is marking a TX as final? Aren't all transactions able to be doublespent/RBF as long as you have the right software?

You could say that a TX is final once it is included in a block, and most of the time that is also the case.

But that is not completely true, it could be that the block which included your TX is orphaned, this means that it will not be part of the main chain. This can happen when 2 blocks are found at (roughly) the same time and most of the miners continue to mine on the block in which your TX is not included, and thus it will abandoned. This is also what would happen during a 51% attack. So theoretically a TX is never final (as long as an attacker has enough hashpower) but after a certain amount of confirmations (# of blocks after the block which included your TX) it is infeasible/unrealistic to reverse. This is also the reason why exchanges ask you to wait for a certain number of confirmations.

So in practice a TX is marked final after a number of confirmations.

[Dabs]

At least in the reference client, or Bitcoin Core, you can uncheck the box that says RBF, so it becomes a final transaction. Set a higher than normal fee, broadcast the transaction, and wait ... It will get included in the next block. Any double spend transactions sent more than 10 seconds AFTER the first one will probably not work, or will be ignored.

There are giveaways, and timelock "test" transactions, albeit of little value (0.001? BTC) and people simply forgot about it, so the creator of the puzzle just claimed it back to himself a full day AFTER the transaction was good to be published.

[MrFreeDragon]

In 2018 one girl with good nails posted her private key to twitter: https://twitter.com/_pronto_/status/821072274442829824

So the key became public, and several users tried to make the spending tx. However miners confirmed the transaction with 152.47 sat/B, and not the older transaction with 446.43 sat/B https://bitcointalk.org/index.php?topic=4345245.msg39982208#msg39982208

In practice it is not a race up to the highest transation fee.

[MrFreeDragon]

It looks that one guy wants to test this game theory. He put 1BTC to segwit wallet and said that it was generated from 12 seed words which will be releasing within the next 30 days:

https://twitter.com/alistairmilne/status/1266037520715915267

As soon as the last words are released, tens or hundreds of people will try to release funds increasing the transaction fee. Let's see if it is the case