Quote from: phzi on February 24, 2014, 10:31:30 PM

@PoolWaffle: is there a reason you have chosen to hide sfire from the miners page?

Quote

I'm doing my best to not treat him/(them?) any differently.

Yet, 7a8678b8 definitely isn't showing on the miners page anymore...

Overall, I couldn't care less about this whale - big hashing power like this causes way less server load then tons of little miners. And I suppose I understand why you might hide him from the miner's page (to stop questions), but it still smells off when you seem to have been very upfront about the pool's operations until now.

He politely requested to be removed from that list via email, was nothing to do with what I wanted, just that he requested it. Didn't seem like a problem (also is literally 1 line of code), if anyone else would specifically like to be removed from the list, feel free to email me, I'd be happy to make it happen. Same goes, if people think that list is annoying/useless (I'm not sure there actually is much use to it), I'm perfectly OK with removing it entirely, so we just have per-address stat pages.

As for all of the talk of 512 difficulty for CPU miners. Again, it really doesn't change anything. You won't get any higher reject rate (or at least you shouldn't unless something is wrong on your miner) than with vardiff. You will see many different blocks causing work restarts, but these work restarts take a few microseconds or so (especially on CPU - where they take much less time than GPU) and you won't get a share in during each block, but your stats will be (over a long enough time period) exactly the same if the difficulty is hardlocked at 1, or hardlocked at 512 (or vardiff).

The only logical point I've heard here is the issue with PPLNS timeframes and hardlocked 512. But again, even this, overall you should end up with the exact same amount of payout over a long enough time. There might be PPLNS rounds where you have 0 shares in the full 10 rounds that are paid on a block, and you'll earn less than you would have on vardiff. But there will also be sets of 10 rounds where you have more than your normal amount of shares, and will be paid significantly more than you should have. There is some variance on the side of which blocks were found at which times, and you might have different per-round earnings than another person with a slow hashrate, but again, over a long enough time, you will have the exact same average earnings.

I keep using the words "over a long enough time", and I haven't explained what that means, partially because there really isn't a good answer (without me doing a bunch of math). It essentially means that over infinite time, things will be equal. Obviously we don't have infinite time. The longer the time frame, the closer your averages should be, the shorter the timeframe, the more the variance could show.

For example, if I ask you to flip a coin 2 times and tell you I'll only pay you if you flip a tails, a lot of people are going to have 2 tails, a lot of people would have 2 heads. This would be 2 extremely different results (in terms here, some people would be paid 0, some people would be paid 200% of what they expected). Now, if I tell you to flip it 100 times each, you'll notice some people still have lower than average numbers (maybe 30 tails), some have higher than average (70 tails), some have just about correct (45-55). Over a long enough time (number of flips here), everyone will come very close to 50/50 (on average). People with faster miners just flip coins faster, people on CPUs flip coins slower.