Economically Unspendable Outputs: A Problem On The Radar

[evoorhees]

How do you not see that this is a problem with the Bitcoin protocol and not SatoshiDice?...The devs need to fix the protocol issue

Yes, I agree that this is currently a vulnerability of Bitcoin and it most definitely needs to be resolved. A simple solution is not so apparent. You're quite well versed with Bitcoin intricacies, can you propose a solution? But in the meanwhile, wouldn't it be helpful for SatoshiDICE to cut down on the spam while we get this thing fixed?

Just because Bitcoin has a weakness doesn't make it right to exploit it for personal gain.

There's an ocean of difference between doing what's right for you, versus doing what's right.

Unfortunately I am not well-versed in the technical intricacies of Bitcoin, and I'm not a programmer. I have huge respect for those who are, and thus I'm happy to contribute to their work by donating some money to assist them. If developers come to me and say they have a good solution, I'll fund it.

Next, I'm not convinced that there even is a weakness. Bitcoin seems to be working beautifully, and already many are thinking about how to scale it even more. These are the growing pains of such a system, and is a sign of surging strength, not weakness. So I do not see that SD is "exploiting" anything... rather it's simply using Bitcoin, following the rules, paying more transactions fees than everyone else combined, and demonstrating to the world how awesome Bitcoin is for gambling.

Yes, SD is profiting from this, but so is the Bitcoin system itself as adoption grows and these growing pains are worked out sooner rather than later.

SatoshiDice is a spark which is driving interest from the entire gambling industry into Bitcoin. Certainly it puts stress on the system, but a system that's used is a system that's stressed, and the Bitcoin economy is far healthier for it.

We cannot advance Bitcoin by asking anyone who is putting stress on the system to "please refrain." I do not think Bitcoin is so delicate and poorly constructed as you. It can handle SatoshiDice, it can handle far more than SatoshiDice, and I congratulate and encourage everyone else out there who is building infrastructure and putting stress on this system.

[1713456535]

1713456535

[1713456535]

1713456535

[alexkravets]

So far there's been no comments in this thread by actual developers, which is not unexpected because such threads are not very likely to converge on any kind of rough consensus for a technical solution. 

This is because actual developers hang out this this forum instead https://github.com/bitcoin/bitcoin/pulls simply because that forum is almost all "signal" and this forum is almost all "noise" from their perspective.

MisterBigg, being a competent C++ programmer that I know you to be, simply submitting a sub-10 line pull request to alter the logic to prioritize transactions unlikely to ever be economically spendable with clear explanation of its logic, test scenario, etc would actually have about a 1000 times more chance of actually making a difference out there in the "real world" of code ... rather than in this talking shop of bitcointalk.org.

Cheers ... :-)

[conv3rsion]

SatoshiDice is a spark which is driving interest from the entire gambling industry into Bitcoin. Certainly it puts stress on the system, but a system that's used is a system that's stressed, and the Bitcoin economy is far healthier for it.

I'm 100% positive that the value of my bitcoins has gone up because of SD's success and interest from the gambling industry that has come from that. So you've made myself, and everyone else who has bitcoins, money. And you've helped us to understand issues that need to be dealt with in the protocol. Thank you for that.

[camolist]

If a small website that 99.9999999% of the world population don't even know about is causing a problem for Bitcoin, then what hope do we have. You don't fix a problem in a p2p payment network by politely (or impolitely in this case) asking them to please stop. That's, of course, assuming that this even IS a problem. I agree with Mike. They're the only site that's really testing the real world (as in wide adoption) stability of the network.

If SDice is considered an attack on Bitcoin, Then Bitcoin itself is flawed.

EDIT: I will add, I cannot for the life of me understand why people are willing to play a negative EV game, with a currency that was worth $13 2 months ago and $42 today... It defies all logic. 

this.

isn't micro transactions a feature in bitcoin? send small amounts low fees unlike credit cards?

using bitcoin to make a provably fair casino? seems to be a perfect match

if 1 in 100 million people in the world using satoshi dice kills bitcoin there is something wrong with bitcoin

(not a player or investor in satoshi dice but this just seems like common sense)

[mobile4ever]

This is because actual developers hang out this this forum instead https://github.com/bitcoin/bitcoin/pulls simply because that forum is almost all "signal" and this forum is almost all "noise" from their perspective.

You are right. Programmers rule in the world of software money (bitcoin).  Thanks for the link and more impetus to learn programming.

[alexkravets]

The also hang out in #bitcoin-dev IRC channel on Freenode

[misterbigg]

isn't micro transactions a feature in bitcoin? send small amounts low fees unlike credit cards?

Yeah when we talk small amounts we're talking about pennies, or maybe a fraction of a penny like 1/10 of a cent. But 1 satoshi = $0.00000045. Is this really useful?

using bitcoin to make a provably fair casino? seems to be a perfect match

Yep, I agree and there is no question that SatoshiDICE has done some great things. But they are also pushing a burden of disproportionate costs onto all full nodes that exist or will ever exist in the future. No other Bitcoin enterprise behaves this way.

if 1 in 100 million people in the world using satoshi dice kills bitcoin there is something wrong with bitcoin

I wouldn't say that it is going to "kill" Bitcoin but it is definitely make running a full node more expensive, and in a way that doesn't grow the Bitcoin user base in proportion to its cost. It has also been driving up transaction fees, again in a way that doesn't really help the network.

All that SatoshiDICE needs to do is pick one of several options for changing the way it sends unspendable outputs to notify gamblers of losing bets. There isn't even a problem with the winning bets, since those are spendable. It's just the losing bets which have a miner's fee and a microscopic output (1 satoshi). Here are a couple of options:

1) Require gamblers to include a tiny additional deposit of 0.01 BTC which is always refunded regardless of a win or loss

2) Send an amount equal to the transaction fee back to losing bets, instead of sending 1 satoshi.

3) Stop sending betting loss confirmations

If you pay any attention to what the developers have to say, you realize that Bitcoin is still not even full release software, it is in beta. The current version is 0.8. It was said that we will probably get up to 0.20 (zero point twenty) before we have a 1.0 release. There are still many to-do items before they consider Bitcoin to be at a level where it is ready for the masses.

It does Bitcoin a disservice for SatoshiDICE to place the network under the constant load of it's unprunable transaction spam in it's current beta state. Sure, if you want to think of it is as doing us a "favor", then okay they flooded the network and now we have some hard data. So, can SatoshiDICE please turn the spam off for a few weeks while the developers get things sorted out? After they make the necessary fixes then SatoshiDICE can turn it back on. That's being a good citizen during the beta stage.

[K1773R]

"SatoshiDICE" notifies gambling addicts of losing bets by sending them a transaction containing 1 satoshi back to their original address. A regular transaction fee is also included in this transaction. The problem is that this produces an unspendable and unprunable output.

u can spend them, wait some time (so they arent "new coins" anymore, dunno how many blocks that would be) and send them even with 0 fees trough the raw tx API.

[solex]

misterbigg, a good level-headed and polite summary

I certainly am not against SatoshiDice as a business or in principle, there are just a few simple house-keeping changes which would be a MAJOR help to bitcoin. Erik loves Bitcoin so I hope he takes these comments on-board.

...Here are a couple of options:

1) Require gamblers to include a tiny additional deposit of 0.01 BTC which is always refunded regardless of a win or loss

2) Send an amount equal to the transaction fee back to losing bets, instead of sending 1 satoshi.

3) Stop sending betting loss confirmations

There IS also lesson for the devs, because SD shows a potential attack vector which a malicious Central Bank (for example) could use to choke the Bitcoin network for long periods of time.

[Elwar]

Maybe everyone should stop using Bitcoin to help out the miners at this early point of Bitcoin's evolution.

We can start using it again once we have enough transactions where this problem will not exist.

[Peter Lambert]

I think calling the amounts "unspendable" is not technically correct - provided an unspent output is old enough even one that contains a single Satoshi can be spent (at least that is the case currently AFAIA).

I don't see why these are unspendable? Couldn't you just have more incoming bitcoins land on that address, and then spend them as a group?

[K1773R]

I think calling the amounts "unspendable" is not technically correct - provided an unspent output is old enough even one that contains a single Satoshi can be spent (at least that is the case currently AFAIA).

I don't see why these are unspendable? Couldn't you just have more incoming bitcoins land on that address, and then spend them as a group?

u could, but the TX is getting big and you would have to pay a higher fee for it, altough u still can do it for free. scroll up or click here: https://bitcointalk.org/index.php?topic=150493.msg1601755#msg1601755

[Peter Lambert]

I think calling the amounts "unspendable" is not technically correct - provided an unspent output is old enough even one that contains a single Satoshi can be spent (at least that is the case currently AFAIA).

I don't see why these are unspendable? Couldn't you just have more incoming bitcoins land on that address, and then spend them as a group?

u could, but the TX is getting big and you would have to pay a higher fee for it, altough u still can do it for free. scroll up or click here: https://bitcointalk.org/index.php?topic=150493.msg1601755#msg1601755

If you can send transactions for free, is there any reason you would not be able to send a big transactions with a reduced fee?

Just wondering: Is the size of the bitcoin transactionss cumulative, or does it reset after each transaction? What I mean is if I take a bunch of small amounts and send them to one address, that would be a big transaction, but if I then send the coins from the one address to somewhere else, does that also make a big transaction or does it become a small transaction because it is from only one address now?

[misterbigg]

I don't see why these are unspendable? Couldn't you just have more incoming bitcoins land on that address, and then spend them as a group?

Okay, you'll have to forgive me if my technical analysis is incorrect but I believe that when you spend an aggregation of coins, there needs to be a separate input coming from the chain of each separate coinbase. That means that when you combine the single satoshi with some other coins there will be at least two txin (I think). The problem is that the extra txin will incur additional per-byte transaction fees which can only be greater than the size of the input (1 satoshi in this case).

It seems that no matter what you do, the added fees required to process the single satoshi will always be larger than one satoshi. Considerably larger, even a 10 satoshi output cannot be spent.

In fact I believe that SD changed the amount it sends from 1 to a small random number between 1 and 5 because some of the mining pools dropped transactions with just one satoshi output.

If you can send transactions for free, is there any reason you would not be able to send a big transactions with a reduced fee?

Free transactions are only relayed or included in a block if the coins are of sufficient age, to prevent attackers from flooding the network with messages.

[K1773R]

I think calling the amounts "unspendable" is not technically correct - provided an unspent output is old enough even one that contains a single Satoshi can be spent (at least that is the case currently AFAIA).

I don't see why these are unspendable? Couldn't you just have more incoming bitcoins land on that address, and then spend them as a group?

u could, but the TX is getting big and you would have to pay a higher fee for it, altough u still can do it for free. scroll up or click here: https://bitcointalk.org/index.php?topic=150493.msg1601755#msg1601755

If you can send transactions for free, is there any reason you would not be able to send a big transactions with a reduced fee?

Just wondering: Is the size of the bitcoin transactionss cumulative, or does it reset after each transaction? What I mean is if I take a bunch of small amounts and send them to one address, that would be a big transaction, but if I then send the coins from the one address to somewhere else, does that also make a big transaction or does it become a small transaction because it is from only one address now?

i once did send a TX which was around 53KB if remember this correct with 0 satoshis as feed, i could have set the fee to 0.0005 and it would still work. altough u would need to use the rawtx API, otherwise u cant do that.

[alexkravets]

+1 for MisterBigg's last post

[Mosper]

So for the currency who prides itself on being "cheaper than credit/debit" you foresee and look forward to when transaction fees are so high it is actually prohibitive to high volume usage? Bitcoin is a neat concept and I'm enjoying learning about it/getting into the whole scene but reading some of the stuff you guys post makes my head hurt. I'm not sure half of you understand the concepts you hold in such high esteem when talking about the benefits of bitcoin in the first place. Seems like most of you are having trouble reconciling the fact that basically everything which helps bitcoin reach its worldwide goals actually hurts you as a single user of low financial power. You will be the same ineffective actor that you are in the fiat world because people who are more skilled and have grasped more power and made more money will push you to the side just like they do now. The vision you all seem to have of how this Satoshi "problem" should be handled is a not-so-free market once it negatively impacts you or you fail to be the one benefiting from a position of advantage over others.

Just keep hitting the reset button until you land on top right? I mean statistically it has to happen sometime huh? No, probably not.

[Elwar]

Free transactions are only relayed or included in a block if the coins are of sufficient age, to prevent attackers from flooding the network with messages.

So if someone wants their 1 Satoshi they can just send it as a free transaction. Once it is of a sufficient age they will be sent.

Problem solved.

[misterbigg]

So if someone wants their 1 Satoshi they can just send it as a free transaction. Once it is of a sufficient age they will be sent.

Yeah for a 1-satoshi amount it would have to age for one or two thousand years. During that time, every miner that ever lives has to store a copy of the transaction even though it can't be spent.

Plus, by that time nodes will no longer relay free tx because miners wont include them in blocks (since there's no fee). This will happen well before the subsidy drops to zero.

[stdset]

misterbigg +1 for discussing this important issue in peaceful, patient and respectful way.

Title is a bit misleading though, but attracts much attention 

And btw, it's hard not to notice, that HDDs capacity growth rate significantly decreased last years. As far as I remember, the last HDD I bought, I bought in the first quarter of 2010. It was a 2Tb HDD for about $100. Now, 3 years later, price of a 2Tb HDD is about the same. And, well, we have 4Tb ones for $300. So, I wouldn't count now on incredibly cheap future storage space.

Those, who don't even know, that low priority transactions pay fees proportional to their size (in kilobytes) should rather read more and post less.