[ExI] Kevin Dowd on Bitcoin

[romyen]

On 01/18/2015 07:20 PM, Mirco Romanato wrote:
> If two blocks are created at the same time at the same level of the
> blockchain and they are distributed to the network nodes, there is no
> voting on anything. The nodes receiving block(n.a) first will add it
> to their blockchain, the nodes receiving block(n.b) first will add it
> to their blockchain. This is called a fork. The miners accepting
> block(n.a) will start mining the successive block(n.a+1) and the
> miners accepting block(n.b) will start mining the block(n.b+1) The
> first to find a new block will start distributing it to the network;
> if block(n.a+1) is found first, every block with block(n.b) will be
> presented with a valid blockchain longer then their. They will discard
> the old blockchain (in this case only block(n.b) and accept the longer
> blockchain as valid (block(n.a) and block(n.a+1). Given the latency of
> the network, the speed of transmission and the average probability to
> find a block every ten minutes, the chance to have an orphaned chain
> longer than 6 blocks are so small it is improbable it will ever happen
> before the universe die.
An attacker could reverse his own transactions if he controls most of
the hashing power, by mining six consecutive blocks. A few big mining
pools could conspire to do this, but that fact would become known
immediately, and they would immediately put themselves out of
business.The integrity of the blockchain would remain intact, except for
those fraudulent transactions,

Kevin Dowd doesn't understand how mining works. He claims that mining is
a natural monopoly, and uses GHash as an example. GHash is a mining
pool, not a miner, because they don't own the hardware supporting their
hash rate. An individual miner uses a pool such as GHash so as to even
out his return. Mining is a Poisson process, with the probability of an
event proportional to one's own hash rate divided by the total hash rate
of the network. A small miner would likely die before he succeeds in
mining even a single block. Therefore, he joins a pool with the reward
split among the participants.

There is a natural tendency for mining pools to get very large because,
first one pool being slightly better leads miners to choose it over the
others (i.e. a small improvement is leveraged into a huge return), and
second, the larger the pool grows, the smaller becomes the variance in
an individual miner's return. That is because the variance of a Poisson
process equals the mean and the total return is split among the
participating miners. That explains why GHash got so big.

At one point, GHash briefly gained control over more than half the hash
rate. This caused some alarm in the bitcoin community, and as a result
GHash stopped accepting new miners. They didn't even need to do this
because enough miners jumped ship on their own initiative. Among the
pools, there exists the opposite of a Tragedy of the Commons situation,
where it is not in any pool's best interest to gain control of more than
half the network.

Kevin Dowd is describing a well known problem, but he misunderstands
it's nature. He even claims that the U.S. government could destroy
bitcoin by gaining control of the hash rate. That's pretty silly. He
doesn't realize how big the hashing network is, with servers located
throughout the world. The U.S. couldn't do that logistically,
politically, or legally.

I predict bitcoin will outlive Mr. Dowd.