Hashrate and 50% attack

It’s just the reasoning factor on a risk avert basis. If you maintain stability in the earning you maintain a certain security.
I don’t affirm that it isn’t a biased reasoning. And i’m clearly open to any correction.
To be a little more precise : since a round earning is more or less the same, having more round by x time might provide a better earning.
I hope i’m clear enough to made my point.

1 Like

The principle that pools need to be above a given size to ensure reasonable variance of rewards is clear. The question is, what size? As far as I can see, there is no evidence that the answer is anywhere near as big (in terms of proportion of network hash rate) as Flypool, although I’d be open to being contradicted with concrete figures.

Indeed, if this were really the case --that you needed a pool with around 50% hash rate in order to achieve acceptable reward variance-- then PoW would be seriously broken and inherently unable to achieve decentralisation (regardless of the algorithm used), which I don’t believe has been demonstrated.

Ah! I wasn’t saying that 50%+ was needed, i was saying that the bigger the better. From my experience, the size of a pool like coinotron seems good enough. (Coinotron having a high diff setting it’s different)
Coinmine, slush and such aren’t big enough for example.

I don’t have enough knowledge and experience to be sure of anything.

1 Like

If you have a lot of different GPU’s to mine, maybe splitting them between several pools could give you better variance than just using one?

1 Like

I would say that spreading the power over different pools is not only a good thing to reduce potential variance in your income (depending on the set of pools you pick), but also better for redundancy.

1 Like

Regarding the 50%+ network hashrate centralization concerns. Why not amend ZEC protocol in a way appropriate for solo mining rather than collective mining? Say, dynamic block size & block issue frequency depending not only on current network hashrate \difficulty, but also taking into account current average miner hashpower (sorry, If propositions are silly, I’m new to crypto)?

Because that would require a complete redesign of that layer relative to Bitcoin, and we were already redesigning a different layer to add zero-knowledge privacy.

Nothing personal but L.Carroll truism “here we must run as fast as we can, just to stay in place” is relevant, bearing in mind the number & pace of competitors