Exploring double spends, a thought experiment

[monsterer2]

We all know why bitcoin was created. Double spends are the problem. Without double spends, you don't need a blockchain, miners or any of this extra complexity. Yet, because information is easy to duplicate, we must accept that double spends are inevitable.

So, as a thought experiment, what happens if you ignore double spends completely?

In a system with UTXOs, like bitcoin, if you ignore double spends, and just credit the spend of a given UTXO to each party involved, the 'victim' of the double spend doesn't lose out, but the money supply of system will increase, leading to hyper inflation and everybody using the system pays the price. This is socialised losses.

If you had a system with accounts, instead of UTXOs (like NXT) then a double spend would leave the sender's balance negative. If you require the sender to post collateral in his account the size of the maximum transaction he will send (analogous to the Lightning Network), then for a single double spend, this problem is mitigated, but if he spends the same amount more than twice, the balance goes negative again, and we have the same hyper inflation problem.

The question I've been pondering is whether you could create some kind of nash equilibrium around the hyper inflation caused by just allowing double spends?

[1713863646]

1713863646

[Anti-Cen]

Nice to see you are thinking out loud !

If you had a system with accounts, instead of UTXOs (like NXT)

These accounts are ledgers and you can hash each record to turn them into block-ledgers and hold the
ledger ID in a block-chain and then place the actual ledgers on separate replicated nodes and address them 
to find them in a similar fashion to how Bit-Torrent works with it's hash codes.

Transactions in these ledgers can reference the inputs by pointing back to the hash-ledger where the spend came from
by including the row-id or the hash of the row if you follow my line of thinking so that should stop money being printed for nothing
which becomes a danger in the lightning network and the balances are allowed to become inflated.

[Xynerise]

If you had a system with accounts, instead of UTXOs (like NXT) then a double spend would leave the sender's balance negative. If you require the sender to post collateral in his account the size of the maximum transaction he will send (analogous to the Lightning Network), then for a single double spend, this problem is mitigated, but if he spends the same amount more than twice, the balance goes negative again, and we have the same hyper inflation problem.

How does this work in a decentralised setting?
The reason it works in centralised settings eg banks, is that they're in control of the account and the funds in it.
Bitcoin addresses are free to create, so what stops people from trivially exploiting this?
Also, what do you mean by "collateral"?
And how will the collateral transaction be distinguished from a normal transaction?
If people have to have a "collateral" before making a transaction to prevent double spends so how do they get funds at all, since they have to put some up for collateral?

How does a hyper inflation currency have any value whatsoever apart from the manipulation of centralised entities like banks and governments?

[Anti-Cen]

Bitcoin addresses are free to create, so what stops people from trivially exploiting this?

Good point and this is absolute key but client side (the wallet) is a ledger but server side it maps to specks of data
all over the place.

Make block-ledgers signable by the private key to the wallet and then replicate the ledger 100 times on the network
and you have a system that scales and keeps the user in control plus inputs reference spends and would show up
during audits even if the 100 nodes conspired to steel my money

[DannyHamilton]

Lets try your system out...

Bob sends Alice 1 BTC.
Bob spends the exact same coins and sends himself 1 BTC at a new address.

Bob now has -1 BTC at his old address, and 1 BTC at his new address.

Bob abandons his old address, his new address is now his only address.

Bob sends Charlie 1 BTC
Bob spends the exact same coins and sends himself 1 BTC at a new address.

Bob now has -1 BTC at his old address, and 1 BTC at his new address.

Bob abandons his old address, his new address is now his only address.

Bob sends David 1 BTC
Bob spends the exact same coins and sends himself 1 BTC at a new address.

Bob now has -1 BTC at his old address, and 1 BTC at his new address.

Bob abandons his old address, his new address is now his only address.

Bob sends Edward 1 BTC
Bob spends the exact same coins and sends himself 1 BTC at a new address.

Bob now has -1 BTC at his old address, and 1 BTC at his new address.

Bob abandons his old address, his new address is now his only address.

Bob sends Frank 1 BTC
Bob spends the exact same coins and sends himself 1 BTC at a new address.

Hmm.  I can see where this is going.  It looks like as soon as Bob acquires ANY value at all, he can just use that tiny bit of value to purchase the entire world.  For that matter, lets look at another scenario...

Bob has 1 BTC at address_A

Bob sends 1 BTC to himself at address_B
Bob spends that same coin to send 1 BTC to himself at address_C

Bob abandons address_A with a -1 BTC balance
Bob now has a total of 2 BTC (address_B + address_C)

Bob spends the 1 BTC received at address B to address_D
Bob spends that same coin to send 1 BTC to himself at address_E
Bob spends the 1 BTC received at address_C to address_F
Bob spends that same coin to send 1 BTC to himself at address_G

Bob abandons addresse_B and address_C (each with a -1 BTC balance)
Bob now has a total of 4 BTC (address_D + address_E +address_F + address_G )

Bob spends the 1 BTC received at address D to address_H
Bob spends that same coin to send 1 BTC to himself at address_I
Bob spends the 1 BTC received at address E to address_J
Bob spends that same coin to send 1 BTC to himself at address_K
Bob spends the 1 BTC received at address F to address_L
Bob spends that same coin to send 1 BTC to himself at address_M
Bob spends the 1 BTC received at address G to address_N
Bob spends that same coin to send 1 BTC to himself at address_O

Bob abandons addresse_D , address_E, address_F, and address_G (each with a -1 BTC balance)
Bob now has a total of 8 BTC (address_H + address_I +address_J + address_K + address_L + address_M +address_N + address_O)

Can you see where this is going?  Given any arbitrarily small amount of value, Bob can immediately use inflation to give himself ANY amount of value that he wishes.  Give him a Satoshi, and he can inflate it until he has 21 trillion bitcoins.

[monsterer2]

Can you see where this is going?  Given any arbitrarily small amount of value, Bob can immediately use inflation to give himself ANY amount of value that he wishes.  Give him a Satoshi, and he can inflate it until he has 21 trillion bitcoins.

Yes, that's the hyper inflation I was talking about.

The question I was pondering was whether there is any nash equilibrium you could build around this.

[Anti-Cen]

Lets try your system out...

Lets see if there is any merits to his thinking given that Bitcoin needs fixing instead of being always on the attack.

I like his way of thinking and when shop lifters get away with steeling then the cost gets past on to me and you
anyway and two months ago Bob dare not send Alice a penny due to the $55 transaction fee but I bet you
have forgotten about that one now.

[DannyHamilton]

Lets try your system out...

Lets see if there is any merits to his thinking

That's what we are doing.

We are discussing his thinking, and demonstrating the shortcomings so that he (or anyone else) can consider possible solutions to those shortcomings.

One KNOWN solution is a proof-of-work blockchain.  It has already been demonstrated to work VERY well in a small scale, and larger scale enhancements are being developed.

Another KNOWN solution is a centralized authority. It has already been demonstrated to work VERY well in a large scale, as long as the authority doesn't take advantage of their position or take actions that have a negative long term effect.

OP is suggesting that a Nash Equilibrium might be another solution.  He or someone else is going to need to explain how that would provide a solution.  As far as I'm aware, such a thing hasn't been attempted, and I've not heard anything about how it would solve the problem.  At the moment it sounds a bit like asking: "Could you create some kind of a ham sandwich around hyperinflation caused by just allowing double spends?"

I'll openly admit I'm no expert on Nash Equilibrium, but just using the phrase Nash Equilibrium isn't going to solve a known problem.

given that Bitcoin needs fixing

That's off topic for this discussion.  This isn't a discussion about efforts that are being made to "fix bitcoin".  This is a discussion about an altcoin that would allow anyone to spend value that they don't have.

instead of being always on the attack.

Discussing problems with a concept which have not yet been addressed in a shared attempt to identify an unworkable concept or solve the problems in the concept is NOT "being on attack".

I like his way of thinking and when shop lifters get away with steeling then the cost gets past on to me and you
anyway

Correct. Which is why we have a centralized systems of authority: a shopkeeper, a police force, a lawmaking body, and a judicial body.  If we want to rely on centralized systems of authority, then the existing systems of money are fine and there is no need to replace it with anything new.

Furthermore, this isn't about stealing, this is about counterfeiting.  If "Bob" has 1 coin, and he can reproduce that 1 coin and spend his reproductions, then he has created counterfeit coins.  The existing system of government currencies implements many features to make it difficult to counterfeit the currency.  It also implements significant penalties for counterfeiting. The question being asked by the OP is "What if we just allowed everyone to create as much counterfeit government currency as they want to? Could we build some kind of Nash equilibrium around this?"

and two months ago Bob dare not send Alice a penny due to the $55 transaction fee but I bet you
have forgotten about that one now.

Nope. Haven't forgotten about it, but that's not relevant to this discussion.

[HeRetiK]

Can you see where this is going?  Given any arbitrarily small amount of value, Bob can immediately use inflation to give himself ANY amount of value that he wishes.  Give him a Satoshi, and he can inflate it until he has 21 trillion bitcoins.

Yes, that's the hyper inflation I was talking about.

The question I was pondering was whether there is any nash equilibrium you could build around this.

Let's postulate three axioms of money:

1) The singular purpose of monies is to be a medium of exchange.
2) To be a medium of exchange, monies require value.
3) To attain value, monies need to be scarce.

If you accept these axioms, you must also accept that hyperinflation destroys the singular purpose of a money and thus makes it useless.

So to solve your problem you need to either look at an application other than money or figure out which of these axioms is faulty.

[pebwindkraft]

Let's postulate three axioms of money:

1) The singular purpose of monies is to be a medium of exchange.
2) To be a medium of exchange, monies require value.
3) To attain value, monies need to be scarce.
 

Whereas 1 and 2 are still valid, item 3 has a tough standpoint with quantitative easing around the world.
We don’t see the inflation, that should have been expected. So there must be other axioms valid, that replace item 3?

[HeRetiK]

Whereas 1 and 2 are still valid, item 3 has a tough standpoint with quantitative easing around the world.
We don’t see the inflation, that should have been expected. So there must be other axioms valid, that replace item 3?

Regardless of fiat currencies being inflationary, there is still a certain amount of scarcity to it. Even with quantitative easing and fractional banking there is a limit on how much money can be conjured out of thin air. In cases of actual hyperinflation money, has always been completely devalued, with people resorting to foreign currencies or other forms of trade where possible.

Nonetheless I concur that axiom 3 hinges on the definition of scarcity. However I'd still argue that most monies we see today are scarce enough, while currencies affected by hyperinflation are empirically not scarce enough to be useful. It follows that hyperinflation as suggested by monsterer2 does not satisfy any viable definiton of the term scarce as required by axiom 3.

Point being, unless hyperinflationary money is shown to be potentially viable, any discussion about some sort of game theory based consensus is moot. In a free market there is a Nash equilibrium regarding hyperinflationary money -- using other forms of money instead.

[monsterer2]

OP is suggesting that a Nash Equilibrium might be another solution.  He or someone else is going to need to explain how that would provide a solution.  As far as I'm aware, such a thing hasn't been attempted, and I've not heard anything about how it would solve the problem.  At the moment it sounds a bit like asking: "Could you create some kind of a ham sandwich around hyperinflation caused by just allowing double spends?"

I'll openly admit I'm no expert on Nash Equilibrium, but just using the phrase Nash Equilibrium isn't going to solve a known problem.

A nash equilibrium around hyper inflation would be any process which allowed network participants to profit from the deviation between the standard coin issuance rate and the actual rate due to hyper inflation. So, the devaluation of the currency would be exactly balanced by this other unknown process.

[Anti-Cen]

hyper inflation would be any process which allowed network participants to profit from the deviation between the standard coin issuance rate and the actual rate due to hyper inflation. So, the devaluation of the currency would be exactly balanced by this other unknown process.

We get inflation anyway from the forks and the money supply in circulations get doubled each time so really it's a mute point.

Watch this about USDT https://www.youtube.com/watch?v=MwKYbT9MoPE because this guys views are respected and he's saying
the fake "Out of the air" USDT pegged to the price of USD is being used to inflate the price of Bitcoins and apparently we have
BTC 200,000 coins missing too so nothing is as solid as it seems.

[Xynerise]

A nash equilibrium around hyper inflation would be any process which allowed network participants to profit from the deviation between the standard coin issuance rate and the actual rate due to hyper inflation. So, the devaluation of the currency would be exactly balanced by this other unknown process.

What would be the standard coin issuance rate and how would it work out since there are no (needs for) miners since we accept double spends as kosher, so transaction ordering isn't necessary?

If there would be miners in this system --for whatsoever reason -- what precludes them from taking the cheating to the next level?
They can make Coinbase transactions and double spend far easily more than non-mining users would be able to.

I don't see how there's any way to "profit" from a system that is so trivially easy to game (see Danny's response above on how a user can turn 1 Satoshi into trillions of Bitcoins).
Value (usually) comes from items being scarce; if the system is open and everyone can cheat easily then it wouldn't have any value whatsoever.

Like I said above, systems with hyperinflation are usually centralised in other to work. The centralisation acts as checks and balances, so even if it doesn't actually work, there is an illusion of it working.

[ArsenyP]

Here is a concept: lets name every single satoshi in circulation. Whenever someone doublespends one, we destroy it and also punish the wallet that doublespent. Now you only need to check for two factors in a p2p interaction: 1) the sender of a satoshi is the owner of it according to the ledger. 2) he owns at least some extra money.
This way you do not need to care about sidechains and doublespending because doublespending would be damaging for both parties.

[TheQuin]

Whereas 1 and 2 are still valid, item 3 has a tough standpoint with quantitative easing around the world.
We don’t see the inflation, that should have been expected. So there must be other axioms valid, that replace item 3?

Regardless of fiat currencies being inflationary, there is still a certain amount of scarcity to it. Even with quantitative easing and fractional banking there is a limit on how much money can be conjured out of thin air. In cases of actual hyperinflation money, has always been completely devalued, with people resorting to foreign currencies or other forms of trade where possible.

Nonetheless I concur that axiom 3 hinges on the definition of scarcity. However I'd still argue that most monies we see today are scarce enough, while currencies affected by hyperinflation are empirically not scarce enough to be useful. It follows that hyperinflation as suggested by monsterer2 does not satisfy any viable definiton of the term scarce as required by axiom 3.

Point being, unless hyperinflationary money is shown to be potentially viable, any discussion about some sort of game theory based consensus is moot. In a free market there is a Nash equilibrium regarding hyperinflationary money -- using other forms of money instead.

Quantitative easing although inflationary has not caused hyperinflation. To see the effects of hyperinflation have a look at the social and economic problems currently occurring in Venezuela where inflation is running at over 2300%. Exactly as you suggested other forms of currency are preferred in Venezuela. I would say that there are plenty of existing examples that prove hyperinflationary money is not viable.

As already pointed out above:

Furthermore, this isn't about stealing, this is about counterfeiting.  If "Bob" has 1 coin, and he can reproduce that 1 coin and spend his reproductions, then he has created counterfeit coins.  The existing system of government currencies implements many features to make it difficult to counterfeit the currency.  It also implements significant penalties for counterfeiting. The question being asked by the OP is "What if we just allowed everyone to create as much counterfeit government currency as they want to? Could we build some kind of Nash equilibrium around this?"

Allowing double-spend in a decentralised currency would most likely lead to similar hyperinflation and thus make it an unattractive currency to use.

[monsterer2]

What would be the standard coin issuance rate and how would it work out since there are no (needs for) miners since we accept double spends as kosher, so transaction ordering isn't necessary?

If there would be miners in this system --for whatsoever reason -- what precludes them from taking the cheating to the next level?
They can make Coinbase transactions and double spend far easily more than non-mining users would be able to.

I don't see how there's any way to "profit" from a system that is so trivially easy to game (see Danny's response above on how a user can turn 1 Satoshi into trillions of Bitcoins).
Value (usually) comes from items being scarce; if the system is open and everyone can cheat easily then it wouldn't have any value whatsoever.

These are all excellent questions that would need answering eventually. The whole point of the question is to get myself, and others thinking about whether you can design a system around the game of trying to cheat it; approaching the entire double spend problem from a totally different angle.