Bitcoin Forum

We know that Bitcoin's resistance to attack is based on the cost for the attacker to overwhelm the network, or at least to control a large fraction of its computing power. We assume this cost will continue to rise in the future so attacks will be infeasible even for well-funded attackers.

But the computational power of the network is proportional to difficulty; and it appears that difficulty is proportional to bitcoin price (http://bitcointalk.org/index.php?topic=4339.0). It follows that unless bitcoins become substantially more valuable than they are today, the Bitcoin network will never be substantially more resistant to attack than it is today.

For Bitcoin to succeed and become secure, bitcoins must become vastly more expensive.

Eventually the block reward will be insignificant, and then the value per BTC won't matter.

Wouldn't the incentive for attack rise as well if bitcoins were worth significantly more?

Difficulty is proportional to the value of the expected reward for solving a block.  As long as the reward is dominated by the newly minted 50, 25, 12.5, etc. BTC, difficulty will be proportional to price.

Eventually the reward will be dominated by transaction fees.  It is reasonable to expect BTC-denominated transaction fees to be adjusted as the value of a bitcoin changes.  For example, maybe a typical transaction will cost the BTC equivalent of 0.01 USD.  If bitcoins increase in value by a factor of 10, BTC-denominated transaction fees will likely decrease by a factor of 10.

Therefore in the long term, computational power and hence security will be proportional to the volume of transactions, or more precisely, the total value of transaction fees per block.

It's probably a bit early to try and draw hard and fast rules for the BitCoin markets behavior. In the long run this makes sense but for now there's still quite a bit of non-rational behavior going on. And people are taking into account not just the current price but the potential future price of BitCoins as well.

The recent runup in prices doesn't have any obvious explanation - certainly the economy didn't grow that fast. If anything perhaps the network is "too strong" right now, that is, the resources being put into securing it are out of proportion to the value of what is being secured.

The problem is we've backed ourselves into a corner where any malicious transaction reversal would be seen as a huge blow to the entire project. But in reality even if 10 transactions were reversed by botnets or malicious actors with chip farms this year, it'd still be significantly more trustworthy than credit card transactions (from the sellers POV).

I think there's too little data to make a meaningful statistical analysis. We've only had 55 changes in difficulty, and the first dozen or so were pegged at the minimum. There are a lot of people joining up, mining for a week or so, and discovering it takes a lot longer to generate bitcoins than they thought, and dropping out. Network processing power is quite volatile right now, and therefore difficulty will be volatile. We're looking at a minor decrease after a 37% increase on the next difficulty change.

I expect things will level off at some point, but I don't know when that point will be.

I'm not arguing any of your other points, I'm just saying it may be premature to try to correlate difficulty with price. But on the other hand, my investment history will show that I'm not a good predictor of the market (basically, if you do the opposite of what I do, you'd probably do very well :-D