The Red Baloons paper

[zby]

Was the recent drop a result of that?  http://coderrr.wordpress.com/2011/11/13/simplified-summary-of-microsoft-researchs-bitcoin-paper-on-incentivizing-transaction-propagation/

Looks like it is at slashdot (http://science.slashdot.org/story/11/11/15/0456206/researchers-locate-flaw-in-bitcoin-protocol) and hackernews (http://news.ycombinator.com/item?id=3237810)

Kudos to the trander that reacted to it so quickly.  I think it will draw the price down even more as more people read these alarmist titles, but actually it is not as severe as they suggest - so it should rebound.

[1714141365]

1714141365

[1714141365]

1714141365

[1714141365]

1714141365

[1714141365]

1714141365

[1714141365]

1714141365

[1714141365]

1714141365

[bittenbob]

I read most of that paper and it sounds like an attempt to try and scare people out of bitcoins by microsoft. The likelihood of someone actually being able to explain said flaw is near impossible and the headline is rather misleading. I especially liked the 2 page long calculus calculation. Someone has put a lot of time and effort into this and one has to ask themselves why. Bitcoins will survive as you said because this is not near as severe as they are making it out to be.

[ineededausername]

I read most of that paper and it sounds like an attempt to try and scare people out of bitcoins by microsoft. The likelihood of someone actually being able to explain said flaw is near impossible and the headline is rather misleading. I especially liked the 2 page long calculus calculation. Someone has put a lot of time and effort into this and one has to ask themselves why. Bitcoins will survive as you said because this is not near as severe as they are making it out to be.

Sir, I don't see any calculus calculation
The entire thing is graph theory.

[bittenbob]

Its about halfway through the 18 page document I was looking at. Didnt spend too much time interpreting it because it didnt seem like it warranted my time or effort

[Crypt_Current]

Was the recent drop a result of that?

no

[bittenbob]

Was the recent drop a result of that?

no

I would have to agree. The paper was published on the 10th so if it was because of that they werent very quick at all. This is profit taking and market manipulation. That is all.

[Steve]

I read the paper and came to the conclusion that the disincentive for miners to forward fee bearing transactions is actually a good thing.  The reason is that it creates a condition where real markets for transaction processing can develop (where fees get negotiated with miners).  All non-mining nodes have a very strong incentive to relay transactions and blocks (it helps ensure they stay in sync with the rest of the network…which is essential if you need to gauge whether a transaction is valid or not).  I don't think there is any reason to believe transactions will ever cease to propagate very rapidly through the network even if every single miner refused to relay transactions.

[eldentyrell]

Was the recent drop a result of that?

If so it was an irrational response.

The Microsoft paper solves a problem that doesn't exist.  The vast majority of the nodes on the bitcoin transaction-propagation P2P network are not miners (or are mining only because they forgot to use the -gen=0 switch to turn off useless CPU mining).

Most mining is motivated by self-interest; most transaction-propagation is not. Participation in the transaction-propagation network is essentially a gift to the network which costs the giver almost nothing.  For example, I run a bitcoind -gen=0 on one of my machines in a data center, just to "be a good citizen".  If it cost me more than $0.01/month I'd reconsider.

However, the problem they point out is a legitimate concern for so-called "CPU friendly" cryptocurrencies like tenebrix.  If those cryptocurrencies achieve their goal (already highly doubtful) there will be few or no "dedicated miners" and you can expect the majority of the P2P nodes to be mining.

[notme]

Indeed... Microsoft's one and only response to open source has always been embrace and extend.

[evoorhees]

I read the paper and came to the conclusion that the disincentive for miners to forward fee bearing transactions is actually a good thing.  The reason is that it creates a condition where real markets for transaction processing can develop (where fees get negotiated with miners).  All non-mining nodes have a very strong incentive to relay transactions and blocks (it helps ensure they stay in sync with the rest of the network…which is essential if you need to gauge whether a transaction is valid or not).  I don't think there is any reason to believe transactions will ever cease to propagate very rapidly through the network even if every single miner refused to relay transactions.

+1

[cypherdoc]

i won't add anything here that has not already been said but i agree i think their conclusions are wrong.  once a buyer signs coins over to the seller the tx is broadcast to the entire network, not just miners.

the seller then has every incentive to get that tx confirmed into his wallet asap so will gladly propagate the tx.  don't forget miners are just pools of self interested and usually Bitcoin advocates so if any mining pool started some shenanigans that could jeopardize the network i'm sure the individuals would pull out immediately.

the glaring thing about the paper is that it totally ignores human behavior and mass psychology.  the authors don't realize that most of the networks participants have an allegiance to the Bitcoin philosophy and will do anything to protect it.

[julz]

the glaring thing about the paper is that it totally ignores human behavior and mass psychology.  the authors don't realize that most of the networks participants have an allegiance to the Bitcoin philosophy and will do anything to protect it.

That was my first thought on the issue too.  I do wonder however if this holds true for the long-term possibility of a vastly bigger and more complex node space.

Could it one day be profitable to create vast armies of selectively-relaying nodes, with the hope of 'surrounding' some proportion of transaction-generating participants? 

Transaction-funnelling aside...  there is already an incentive to create such a system of nodes which 'instrument' the bitcoin network in order to de-anonymize transactions and analyse flows. (as mentioned by Dan Kaminsky)    For an organisation already set up to sell bitcoin econometric data from their massive node network - it may be tempting to earn some kickbacks from large miners to do preferential-relaying.

What could the Bitcoin supporters do to counteract this?  Run their own clouds of honest nodes? Even if the cost per node is pretty low - they're competing with bad node operators who are earning something from their nodes.  I guess the miners themselves would then have an incentive to run as many honest (or self-directing!) nodes as possible too. Do we end up with not just the hash-power arms-race - but also a node-count competition?

[Nagle]

That attack isn't a problem. Bitcoin will tank long before all possible Bitcoins have been issued and mining is only for keeping the block chain alive.

[RodeoX]

The slashdot article says microsoft researchers found a "flaw" in bitcoin. A flaw? well, that is what they called it maybe. But I don't consider it a flaw.
Windows Vista is a much better example of a software flaw.

[BadBear]

If you said Windows ME my eyes would just gloss over and I would pound +1 in rage, but Vista is awesome cause it paved the way for and ensured driver support for Windows 7. 

[RodeoX]

If you said Windows ME my eyes would just gloss over and I would pound +1 in rage, but Vista is awesome cause it paved the way for and ensured driver support for Windows 7. 

Roger that, ME is the worst distro ever. But I have a copy of vista 64 that has to be second place. Man it's $hitty.

[Vandroiy]

I kind of feel slapped in the face with all of this.

Yeah, I was that guy holding a month-long discussion about dynamics in Bitcoin. It was about the dynamics of transaction fees after minting becomes small, the natural equilibrium of fees converging to one Satoshi in the "ideal" case. Know what? Nobody gave a shit. Everybody was like "lol Bitcoin after minting, that's in how many years?", "ah we know already" or "Herp I no believe, here is my personal theory why" which turned out different each time.

And now, somebody points out a problem that has the same conditions, but is tiny in comparison, and everybody talks about it? I don't usually swear. *Clears throat*. THE FUCK, SERIOUSLY? This shit is extremely inexpensive to counter, so most likely people will solve it for free, and if not, the miners will pay someone to relay transactions. Yes, any non-idiotic client will broadcast to those who relay. Problem solved! Duh!

Now I'm not saying the paper is wrong, but it's a ridiculously unimportant issue. Unlike the one concerning possible attacks on the network, it can be solved when -- if -- it ever comes up.

Please just stop bumping this thread, just as a favor to me. Talk about the dynamics if you want, but not as "speculation", seriously.

PS: No offense to MS research in general. I believe they're usually better than this.

[julz]

Yeah, I was that guy holding a month-long discussion about dynamics in Bitcoin.
...
Please just stop bumping this thread, just as a favor to me.

I don't recall even seeing your username before let alone a month-long thread on some topic related to this.
Thread-reference - or suffer the bumps more humbly please.

[Vandroiy]

I don't recall even seeing your username before let alone a month-long thread on some topic related to this.
Thread-reference - or suffer the bumps more humbly please.

No wonder, since you joined in June, when I was already occupied modeling the bubble.

The thread that started it: https://bitcointalk.org/index.php?topic=6284.0 -- however, that is pretty chaotic since nobody seemed to understand the whole situation when I started the thread, including myself.

Instead of telling people to read through a partially derailed discussion, I ask this: try to find an equilibrium state of transaction fees and the size of the miner network, noting that neither are locked. Take into account that there is no consensus on keeping current block size limit and no well-supported idea for an alternative.

Anybody trying to solve this, using the most basic assumptions from game theory, is in for a surprise. It's not very clear what all of this means in the end. But it is certain that Bitcoin mining will have radically different dynamics after minting.