Quote from: organofcorti on December 19, 2013, 06:18:23 AM

Quote from: eleuthria on December 19, 2013, 03:20:39 AM

Quote from: organofcorti on December 19, 2013, 01:55:13 AM

How can ghash.io negatively influence another pool's orphan rate? I can see how being well connected can reduce their own orphan rate, but how can they increase someone else's orphan rate?

Most pools are in at least decent datacenters, which happens to mean they have low latencies with the other pools since they're both close to backbone bandwidth providers. This reduces overall network orphan rate because the average time between block notifications between the majority of the network hash rate is significantly lower than what it would be between people just running on random local ISPs. At the same time, GHash.io has abysmal latency and/or absolutely no peers setup between pools, or they're purposely ignoring competing blocks sometimes based on the sheer number of orphan races they compete in. This is based on admittedly more casual observation rather than deep analysis, but I know other pool ops have been seeing more losing orphan races than usual as a result of ghash.io.

So you mean that GHash.IO are could be causing more orphan races than usual, rather than winning more orphan races than usual? Or both?

Am I right in assuming that this is only an issue because they have such a large chunk of the pie?

I'll do an exact count later. They are not winning more orphan races than usual given their hash rate. But for the most part, the seem to more often than not lose an orphan race to anybody else if they didn't solve the next block themselves (because rarely is their block the first seen by the rest of the network when they're in a race), and they definitely are participating in an unusually large number of orphan races for being ~30% of the network.