Bitcoin Forum

This address has, from November 2nd 2012 till now, produced more than 12000
transactions betting on satoshidice and immediately re-betting the return.

Some of these obviously take a while to confirm (despite 0.5mBTC-fee) and
are binding other funds into never-confirming transactions and double-spends.

How could the bitcoin-ecosystem defend against such? Or is it normal business?
One thing that comes to my mind is, that returning a satoshi for lost dice-game
bets seems wrongheaded in hindsight.

https://blockchain.info/address/1PEDJAibfNetJzM289oXsW1qLAgjYDjLgN
https://blockchain.info/address/1PEDJAibfNetJzM289oXsW1qLAgjYDjLgN?offset=12000

I don't see the problem. He is not tying up anyone's else money than his/her own. That guy is the only loser, in addition to playing losing game.

But the real thing to note is that transaction costs will go up.

Maybe the problem is limited to the gambling corner of BTC-world, where
waiting for confirmations before re-spend is rarely seen.

If that's the case, then the owner of the address merely trolls the gambling-scene.

I'm not yet long enough in the scene to judge how that might affect the rest in
medium or longer terms.

+1

It is normal business. That's what transaction fees are for. Miners are free to increase transaction fees if it's a problem.

A single satoshi is completely useless as money unless you don't mind paying a five million percent transaction fee to spend it. What it is useful for is proving that a losing bet was processed correctly, otherwise there's no easy way to tell the difference between a losing bet and a failure to pay out a winning bet due to technical problems.

Ok.  I was also thinking, that if a substantial part of his txs gets confirmed, then that one
guy alone probably added several megabytes to the blockchain within merely a month.


Still it feels like spamming the blockchain in the long run.  There have been discussions about
embedding messages in transactions and all responsible service-providers (fortunately) refuse
to do it, but single satoshis are still being spammed into the blockchain in large numbers
essentially for "message passing purposes".

That isn't exactly accurate, as hardly anyone re-spends a received satoshi in a tx all
on its own. Typically, a small number of single satoshis are added to other transactions
without further raising the fee for that tx.

Some more knowledgeable people might create statistics of how many of the one-satoshi
outpus are spent versus still ignored after a certain time.  Most wallets probably don't even
offer a feature to "forget" certain outputs, so likely they'll all be used sooner or later.

So what? Who cares what anyone does to the blockchain as long as they pay the transaction fee?

A transaction with any outputs smaller than BTC0.01 requires a BTC0.0005 fee. A transaction with low-value inputs does not incur this fee, but an extremely large number of such inputs are required to create a large-value transaction, and that does incur a larger fee. Note also that when a large number of outputs are spent, they no longer need to be remembered by clients that implement pruning, so consolidating single satoshis into a large transaction actually benefits the network rather than hurting it. Sending single satoshis individually does hurt, but that's what the fees are for. There's no problem, and if it ever becomes a problem it can and will be solved by increasing the transaction fees.

Unproductive growth of the blockchain appears to me like a similar type of problem
as environmental concerns in real life. As in: "Who cares, if I use ozone-slayer-sprays?
I paid for them, when I bought them."
I do not realistically expect *everyone* to give a sh*t for "environmental" matters, but
I hope that certain central role-players do, where the effect multiplies by large numbers.


That's irrelevant to the problem that the majority of single satoshis *WILL* be re-spent,
and each re-spent satoshi (just like any other value) costs about ¼ of a kilobyte in the
blockchain.

It doesn't really matter at all, that spent satoshis no longer fill up the owner's wallet.
That's like: once the garbage collectors empty my waste bin with nuclear wastes, the
disposal problem is solved.

All I'd like to see is betting addresses that just don't waste a response on loss, and I'll
just have a look at the website to know if I lost, or the bet went wrong. In exchange for
not wasting typically 0.00050001 BTC for lost bets, the betting sites could drop the edge
a bit ;-)

(Edit: ¼ rather than ½ of a kb for each re-spent output)

If you expect to need to increase fees to ration space later, the efficient thing to do is begin increasing fees now.

Blockchain space is not priced efficiently at all. To get an efficient price, you need full centralization of the blockchain. Without full centralization, you need some form of central planning (e.g. coding rules capping block size). However we can see from the above example that a cap on block size is not an efficient policy. An efficient policy would be considerably more complex.

Good thinking. We could use more people with such good intuition around here.

What did you smoke? :-)  I want some of that sh*t, too!

Please tell me, that its only my sarcasm detector that needs recalibration.

You said it yourself.


If you want to price access to a common pool resource like blockchain storage efficiently, you need to have a single owner controlling access. If multiple owners can sell access, then they will set inefficiently low prices. The blockchain will get cluttered with crap.

Since having a central owner is off the table, you need to create rules that mimick the decisions made by a central owner. A central owner would conserve space now to save room for later. There is no attempt to do this in the blockchain, so it is filled with spam.

I understand why. It is a difficult problem. We don't know how much space will be worth later. Nevertheless, it never makes sense to allow complete spam.

How? The blockchain isn't an ecosystem, it's essentially a special-purpose cloud-storage service. Storing information (normally financial transactions, though any kind of information can be stored in the blockchain if you want) in exchange for a fee is the whole purpose of the blockchain. That's what it's designed for and that's what it does. Storing excessive amount of data in the blockchain isn't "pollution", it's ordinary use of a service in the way it is purposefully designed to be used.

Nobody other than miners even needs the full blockchain, and miners charge a fee to add to the blockchain to cover their expenses. A pruned blockchain is good enough for non-miners, though of course they're free to use a full blockchain at their own expense if they want. And that expense is very minimal. Hard drives are cheap and getting cheaper all the time.

You're using the wrong tool for the job. Sarcasm detectors don't function as expected when your interlocutor truly believes what they say. For these people, you need a BS meter. When dealing with people like cunicula, be sure to set it to high-range mode to avoid overloads. :)

I used to do this, but then I discovered the forum's "ignore" feature.

Do either of you dullards have any professional training in economics at all? Didn't think so. I am just explaining how collective storage works using neoclassical economic theory.


The OP had the correct intuition, which is unusual. Most people find these situations confusing, like yourself, for example.

If you think I am incorrect, then you should explain how. This allows me to pick apart your reasoning and identify your errors.

If you simply dismiss what I have to say, then you will remain ignorant.

Maybe a bit off-topic, but I have always wandered what prevents satoshidice or similar service to include your bet in a losing (for you) round instead of a winning one?

As I understand it there are several dice rounds per minute and if the house waits just a bit it will have a great chance to encounter a loosing round for you within that minute. It can always claim that it didn't insert your transaction earlier because it didn't see it.

Any thoughts? Did I get the whole thing wrong?

+1 
coinicula also made me activate this highly useful feature.

You're merely unlucky in making your point.  You're like someone going into
a cat-lovers forum and answer questions by first suggesting adopting a dog,
and only later continuing to point out some cat-inherent facts. That way you
collected ignores by people not bothering to read past the initial nonsense.

Actually, I think the fees are ok as they are, and all I'd like to see are
dice-games allowing me *not* to waste essentially *my* money on
some extra  ¼kb blockchain storage that I do not really need nor want.

They could offer alternative no-response-on-loss bet-addresses.
If the house-edge is then just a bit lower it will attract the customers
and the service-providers will still even earn more as they no longer
waste 0.5mBTC per lost bet. Customers would instead just look on the
website if they don't see a win-payout within a minute.

I had better make that point in a dice-game forum, though. (did so, in
the btc-dice forum once, but it shut down shortly afterwards - without
having implemented the suggestion, yet, btw.)

You are right. But I have no patience with people who cannot support their views with a logical argument.
I'm certainly not willing to reduce myself to their level.

On a related note: https://blockchain.info/address/1SEXYws5t1C8E4ziQTQbR9fMRAZbFxVkk

Seems like an attempt to spam the network as well, without paying any fees.

The associated blockchain info page is a symphony of beeping.

Wonder if it is a restless miner trying to force fees to go from zero to 1 satoshi?

must be a bot.

cunicula is right, the pricing is inefficient because the block size limits are arbitrary, and without a central entity controlling them it's unlikely that prices will settle at an optimal value. Actually, just read his posts again, they provide an accurate explanation of the problem.

The Bitcoin devs just had no better idea than an arbitrary limit, and since it's not an easy problem, it's hard to blame them. We can can probably deal with high future fees by introducing extra systems that use Bitcoin as backing. Thus we have limits that are only somewhat high now but probably low on the long run.



It's amazing how people just use some sort of mob-truth-finding and wander on. It seems that "Haha, I ignored you" remains the best BS-indicator.

Cunicula (I didn't ignore him) raises his point, in that blockchain-space is currently "too cheap" and thus wasted.
But he talks utter nonsense, when his solution goes about centralizing control over the blockchain.
If anyone wanted a centralized infrastructure in digital money world, they could just as well use
paypal, web money, e-money, Linden dollars... and whatnot.

There already is a price for getting satoshi-outputs into the blockchain.  The problem is,
that the dice-games (which are probably #1 generators of satoshi-outputs) give the players
no choice for opting out, yet. (Short of stopping bitcoin dice-gambling at all, of course.)

It's like as if all groceries gave you the shopping cart as a present (for you to keep), each time
you go shopping there (and effectively you have to buy it with your money, as the prices of the
goods would then have to be set such as to include the price of the cart).