| Satoshi planned for this 51% mining attack problem, and so far we're ok. His premise was that if a miner ever grew so large as to be capable of performing a 51% attack, that they would have so much wealth invested in the Bitcoin network that they would never dare try anything. The Bitcoin network is GHash's golden egg laying goose. The Bitcoin network is entirely safe in GHash's hands imho. The only real worry comes in when the 51% miner is a bad actor. >at 35% hashpower and 3 confirmations, this means that GHash can currently steal altcoins from an altcoin exchange with 15.6% success probability – once in every six tries. Yes, but come on. They never would. Would Bill Gates jump into an active volcano to pick up a penny? Even if he had a 15.6% of picking up that penny without dying a fiery death? >So what if, for the sake of example, GHash gets over 51% again and starts launching 51% attacks (or, perhaps, even starts launching attacks against altcoin exchanges at 40%)? What happens then? I'll bite. The ~75% of GHash's mining power that is being pointed at GHash by regular Bitcoin ASIC miners who are enjoying their 0% fees and merged mining drops off overnight. Their domain names are DDOSd at a record breaking level for weeks on end, and every other pool adopts a modified bitcoind that hard-forks (erases) any damange GHash might have caused. GHash is left with a few thousand nearly worthless Bitcoin (because the USD/BTC value would tank on their 51% attack news) and they are also left without the possibility to basically print themselves free money until the end of time. |
All it takes is one bad actor at GHash, as supposedly happened with the gambling incident, and suddenly you've got a problem again.
Just because GHash is successful now does not mean they won't change. Let's imagine a future where CEX.io decides to ditch GHash. GHash no longer has any form of profit (0-fee pool) and thus has no incentive to remain honest.
Satoshi's logic, that a large minor would have a selfish interest to be honest, actually completely falls down when it comes to mining pools. Mining pools have massive power, and yet (in this case) get 0 profit. GHash could easily have nothing invested in bitcoin while still having control of a large gashing power.
Now on to the other things you said. "Would Bill Gates ... for a penny?" is a terrible analogy. For one, this isn't a penny. 51% attacks could allow GHash to steal millions and extort further. They would also allow someone with a vendetta against bitcoin do do real damage... And the risk isn't nearly as severe as jumping in a volcano. A more relevant analogy might be "Would a company commit tax fraud on the order of a million dollars and risk being caught" (yes many have) or "would an executive steal from the company he works at and, when they notice, cut and run."
Your theoretical for what happens on a 51% attack is also silly. There would be signs of some attacks (like consistent double spending on high-confirmation things), but other attacks like extortion-for-inclusion could be silent for as long as the victims don't speak up. Furthermore, people wouldn't leave overnight.. not that many did when GHash performed attacks the last time. In addition, people are highly unlikely to change. Things won't drop off over night. They weren't DDOSed last time. Your entire speculation is optimistically assuming there won't be significant doubt and inertia... and betting against human lazyness is rarely a good idea.
Finally, you assume GHash is left with nearly worthless bitcoins. Well, sort of. The extortion could lead to them being paid in anything. If the entire company decided to do an attack they'd obviously sell all bitcoins first so any new bitcoins gained through any attack would be pure profit (and wouldn't be worthless for at least several hours until someone notices and news spreads). Furthermore, each employee of GHash that is capable of causing such an attack to occur might not have any bitcoins, even if GHash does.
I think that your defense of GHash is flawed, especially considering they've behaved badly in the past. and your arguments in general aren't strong.
I see it as fully plausible that if they maintain 51% hashing power for an extended period of time, a bad employee might modify their code for their own profit and then GHash will plead ignorance when it's discovered and carry on. I also see it as fully plausible that if GHash decides to close shop at over 51% they'll choose to go out with a bang and do what thievery and damage they can on the way out.
I fully agree with the article that this is a problem.