Hacker News Reader

Comment by eatitraw | Hacker News Reader

> It also assumes rational players.

All models are wrong, but some are useful. What is the chance that there is an irrational player that spends a lot of money only to destroy bitcoin? This kind of player has to be politically motivated, not just simply irrational.

> Besides, there are plenty of profitable avenues if you control 51%

Like what? Basically you can do a few things:

1) Mine all the current coins

2) Double-spend

3) Not confirm any transactions

All of them are harmful to the network. However, given that it is not an actual single entity that controls 51%, I wouldn't be worried much about anything except 1)

View on HN

Game theory is not even a useful model here. There are plenty of rational players who would desire the downfall of Bitcoin.

The game theorists assume at the outset that everyone involved's primary financial interest is in a valid working Bitcoin network. However, banks and Western Union would not desire competitors for international currency transfer. Governments may not want currencies they can't control. The Russian mafia may want to exploit it for short-term gain or launder money. And now that there's a single pool with >51% of the CPU power, the difficulty of a hostile takeover just went way, way down.

Before, a group would have to set up a massive amount of computing power in order to take it over, but now, taking over or disrupting Bitcoin is within the reach of non-technical entities like criminal organizations and the CIA. E.g., all the CIA has to do is threaten a few people and say "Insert this code so we can freeze the transactions of $VILLAIN_OF_THE_MONTH at any time." There's many more possibilities now.

> There are plenty of rational players who would desire the downfall of Bitcoin.

Then game theory is pretty useful model. Players are rational, after all. Thats what's game theory is about.

> Governments may not want currencies they can't control.

Makes perfect sense. But we're discussing 51% attack. Would a miner with 51% of hashrate willingly help a government(and why would he)?

Or, would a government simply try to legislate bitcoin out of existence? I would say that the latter is more likely than the former.

> And now that there's a single pool with >51% of the CPU power, the difficulty of a hostile takeover just went way, way down.

Remember, it is not a single entity. It is a pool. And, in fact, ghash.io share is down to 39%.

> all the CIA has to do is threaten a few people and say

I do agree with this point. One of the strongest points of bitcoin is decentralization. We should keep bitcoin as decentralized as possible. But I would say ghash.io having 51% is a minor obstacle.

Players are rational, after all. Thats what's game theory is about.

Bitcoin players aren't rational. Karpeles stood to earn way more by not manipulating the market. He did so anyway and bankrupted his own company.

Game theory assumes all players are smart.

> Then game theory is pretty useful model. Players are rational, after all. Thats what's game theory is about.

Sorry, I wasn't clear enough. My argument is that the people citing game theory as a reason Bitcoin players had an incentive to avoid a 51%+ scenario kept assuming that all players wanted Bitcoin to succeed.

(However, as a former cognitive neuroscientist, I think it's mistaken to assume that people always act rationally. Check out the literature on the ultimatum game or anchoring effects in prospect theory (for which Kahneman got the Nobel prize) to see examples of people acting irrationally in a sytematic, biased way.)

> Would a miner with 51% of hashrate willingly help a government(and why would he)?

Who said anything about being willing? My example suggested threats, which is way more likely than cooperation.

> Remember, it is not a single entity. It is a pool. And, in fact, ghash.io share is down to 39%.

Yes, but the people running the pool have the keys to the kingdom as long the pool members don't know. So, more subtle perversion has a good chance of lasting a while, while gross manipulation is more likely to cause pool members to switch.

>What is the chance that there is an irrational player that spends a lot of money only to destroy bitcoin? This kind of player has to be politically motivated, not just simply irrational.

Irrational in this context (game theory) means self-defeating, but only in the confines of Bitcoin. It doesn't mean insane in the common sense.

So you can very well be "politically motivated" and irrational in this regard.

4) Only confirm transactions that give you large transaction fees, and disallow all others

Etc, etc.

Reply on news.ycombinator.com