Topic: Why you should care about miners (Read 2102 times)

From a game-theoretical standpoint, zero-sum is identical to fixed-sum.

You are correct that the primary effect is that higher valuation increases the incentive to mine, and hence the total hashrate and network security, and that a common myth is that the primary causality is reversed. This doesn't mean the reverse effect doesn't exist too - at least some people, when making decisions about investment in Bitcoin (both buying bitcoins and taking part in the economy) will want to know that the network is secure, thus improving security with independent means will increase the valuation.

I don't know the current status of that; I don't think it counts as an attack; and yes, we should also avoid changes that make it easier for botnets.


First, I consider p2pool to be a type of pool, and it reduces variance and is in no way centralized.

Second, even a centralized pool can allow miners to have control over blocks, a setup sometimes called "split pools" or "smart miners". The miner composes a block himself, or receives it from an independent party, and credits the pool in the generation transaction; he submits to the pool the Merkle branch of the generation transaction of every share to prove that he works for the pool.

Third, it's not necessary that 3-4 pools will be >50% of the hashrate. There are several ways to keep miners' variance low even with many small pools.

Yes, increasing block frequency would reduce variance. But it will also increase effective network latency (point #7). I'm getting ahead of myself, but I'll hint that the reason I'm making this post now is that I want to talk about block frequency next.

Since a lot of changes are going to affect good and bad miners alike he was probably referring to that sort of change.

Well, agreed.

But what about the current prolonged attack, empty block mining. Have they stopped.
It is only a minor incontinence for conformations taking longer, but still it is an attack.

Also client development seems to be centered more about people using Bitcoin to issue payments than to verify them. To date there's still no real easy way I'm aware of to set mx own transaction inclusion rules as a miner. Especially with the rise of P2pool (which relies on a locally hosted bitcoind + it's ruleset) this might become more and more important in the future.

I disagree, or don't fully understand, that miners should join mining pools. To me, it seems that mining pools are the obvious source of the attack. No attacker is likely to build super computers necessary to reach 51%. They're just going to corrupt the three or four major mining pools into doing their bidding.

If I were a criminal organization I would go to the handful of individuals that are in control of the minimum number of mining pools necessary to achieve the 51% attack, and I would threaten them. I would threaten their families in the most horrific ways. Or I would bribe them. I would blackmail them. I would do whatever it takes to simply corrupt the people who run those mines. Or I would seize their equipment on the same day that I fire up my own mining equipment.

The point I'm trying to make is that with so much mining power concentrated in so few hands, successfully attacking Bitcoin seems trivial. The goal should be to decentralize as much as possible. Ideally every miner would mine for themselves and themselves alone. It's inconceivable to achieve a successful attack of several thousands or millions of miners. It's much easier to go after a handful of pools.

But you do make a good point that the variance is too brutal for individual miners. Who would want to set up their own mining rig, not to join a pool, but only hit the lottery once in a blue moon, if ever?

Now, I don't claim to know the answers, but if I had to take a guess I would say increase the frequency of blocks while decreasing the reward for successful block mining. But I'm sure there is something wrong with that. I just don't know it.

More secure network leading to higher bitcoin valuation? I would bet my ass that won't be the case, at least for the foreseeable future. Mining difficulty will follow bitcoin valuation, not the other way around.

Mining is a fixed-sum game for miners, not a zero-sum one, as a set amount of BTC are awarded for mining.  Mining is not even necessarily fixed-sum in terms of USD, as a more secure network could lead to bitcoins becoming more valuable.

That said, I agree with your post.

I was trying to make a case for there being a good reason, in general. The specifics should of course be considered on a case-by-case basis.

There's most of the problem right there.  Those don't occur unless there's a darn good reason.

tl; dr: The easier it is for honest miners to mine compared to an attacker, the more secure is the Bitcoin network.

This is a point which seems completely banal to me, but if even Bitcoin's lead developer can be confused about it, it deserves a post.

Mining is indeed a zero-sum game for miners. The total amount of bitcoins to be gained by mining is fixed by protocol and by the transaction fees users are willing to pay, and if it's easier for one person to mine, what he gains will come at the expense of other miners.

Mining is not a zero-sum game for Bitcoin, and this is evident if we recall the main reason mining exists, which is to secure the network. The more people are able to mine, the greater the total honest hashrate, which makes it that much harder for an attacker to overtake the blockchain. Also, more miners means more decentralization, and more difficulty in commiting attacks by compromising specific mining entities.

Note that this isn't a one-to-one "helping X GH/s worth of miners join increases the total hashrate by X". The new miners will increase the difficulty, causing some old miners to leave; a new equilibrium will be reached, which has a higher total hashrate than before.

Following up on the "zero-sum game" theme, doesn't making mining easier also mean it will be easier for an attacker? Well, that depends on how exactly we make mining easier. For example, if some change is done that doubles everyone's hashrate, it will apply equally to honest miners and attackers and thus have no effect.

But any change that gives honest miners an advantage over attackers, by addressing difficulties that honest miners are more likely to care about, can increase the total honest hashrate without increasing the hashrate that can be amassed by an attacker at a given expense, thus making the network more secure. Some examples of things that can help honest miners (especially at-home miners) more than they help attackers are:

1. Reduce mining variance. An attacker won't have much relative variance, because he is big enough to carry out an effective attack; and he will not care much about the reward or its variance, since the mining reward isn't his main objective.

2. Make mining technically simpler for various use cases. An attacker isn't going to care about technical difficulty; he will hire the best engineers and do it at scale to reduce the amortized cost. For an at-home miner, even a trivial inconvenience can mean the difference between starting to mine or not.

3. Support mining pools. An attacker isn't going to use pools, because he doesn't need them and because he wants control over the blocks. At-home miners must use some form of pool.

4. Make efficient mining hardware available to the masses (or make the hardware available to the masses efficient for mining). An attacker is possibly going to develop his own ASIC solution to be deployed at scale; smaller miners can't afford this luxury, so they should not be left behind in this arms race.

5. Make it possible to leverage existing resources. An ability to mine on the idle CPU of a computer, or a mining device on a PCIe card which can be added to a PC and utilize existing computing, electrical and physical resources will help at-home miners more than an attacker, who will need to make a dedicated purchase of the entire infrastructure.

6. Offer mining assets such as companies and bonds. These will allow people without the means to physically mine to put more funds into honest mining, while being completely useless to attackers. The issuing entity will need some accountability to make sure that it is not itself an attacker.

7. Reduce network latency. The more time it takes to learn about new blocks found, the more hashes wasted on old blocks. Attackers will not suffer from this, because they mostly mine on their own blocks, which they can be instantly aware of.


This also means that some ways to make mining "easier" can arguably actually help attackers more than honest miners, and thus harm network security. Computation/hashing rental services, which give the renter full control of the contents of the computation, if done at scale, can allow an attacker to make a blitz without investing in a long-term deployment.


So if you want the blockchain - the major technical innovation that separates Bitcoin from anything that predated it - to properly do its job of securing Bitcoin transactions, then you should care about mining, you should care about miners, and you should care about changing the Bitcoin ecosystem (which also includes the core protocol) in ways that are beneficial to the typical miner.