...
Profitability is the same at most pools.
While we are 'expected' to find multiple blocks per diff change, i.e. multiple blocks per 2 weeks, profitability wont be affected much by changing pools.
You can't predict luck and luck is the biggest factor in PPLNS.
Of course if the pool gets 'too' small then that will noticeable affect Profitability, a fact that certain smaller pools like to hide from their miners, however we've been growing pretty well over the last month ... so that's not currently a concern.
It already got too small as soon as you dropped to less than 20 blocks per month expected. That's pretty much the minimum amount of chances to avoid what is going on right here. You can't ignore time frames when you get small and try to relate a diff% in your current state with when you had a much higher expected block %
This place has turned into a social experiment more than a mining operation
So ... can you back up that comment with maths?
Yeah - here's a real world example that gives a similar way to look at how bad luck effects low datasets in short time frames.
Say you have $200 a month to play blackjack and you play everyday but you only play through the money once per day. All your winnings go into your right pocket and once you finish for the day you leave with what you won.
The next day you take what you won, put it in your left pocket, and play through that once. Etc, each day for the whole month until you get your next $200 to add to whatever current pile you have. If you get through the whole month you can see if you were a bit lucky or unlucky with gain or loss but you got through the whole month without getting busted.
Ok, here is the spin that will show you why your pool is playing with fire:
If you are a larger pool, it is like you are playing at the $1 or $5 tables and you get lots of bets to start out and most likely get through most bad streaks and have money for the next day all the way through the month. You won't feel the effect of bad streaks that much since you have enough bets to get win streaks to counterbalance within the month (relate that to a diff cycle in mining)
If you are at $20 tables it starts to get a bit dicey if you get some bad streaks that can break you before the end of the month since smaller bad streaks can knock you out.
If you go to $50 or $100 tables you start to see how getting a little bad luck early in the month can knock you out till next month... or you can play with fire like this pool did for the previous three months and skate through and feel normal by not getting enough bad luck to really see that you are playing with LOW MARGIN OF ERROR
This example isn't exactly what is happening with mining, the math is different but it illustrates my point about small mining pools sputterring in short time frames (diff cycles) when they get too low of an expected blocks per month... you will experience the same feeling as getting busted early in a month and a long wait. Maybe things go back to normal the next month or whenever but you just don't get enough tries to have luck work itself out during each diff cycle.
Ignore this if you want, but you've been playing with low margin of error since Nov and this last diff cycle you crapped out. And with mining, you need consistent luck since mining returns diminish every diff cycle increase
I said maths, not some gambling story you experienced, that has nothing to do with mining
Long term luck is expected to be 100% ... guess what ... it is here also ...
If you are attempting to say (but are failing to) that variance is higher here than at larger pools, well guess what, I keep saying that also.
I even said in the first post of this reply that there is a risk if the pool gets too small - but it's diff related, nothing to do with having a gambling addiction.
The larger the pool, the smaller the variance.
My preferred size for a pool is actually 5% of the network, not based on some random story or some inconclusive maths, but based on the fact that 5% means that you can handle a 720% block in a single day.
... and at 5% you'd expect a 720% or worse block to happen, about twice a year - so not very often.
720% CDF = 0.99925341419162 = 1 in 1339.4 and since 5% means on average 7.2 blocks a day, 1339.4/7.2 = 186.03 and 186.03 / 365.25 = 0.5093 i.e. about twice a year.
A pool is unlikely to ever get a 1200% block, and that would be 1.6666 days.
So although you might on rare occasions get the worst possible luck (> 1000% would be about once every 8.4 years at 5% of the network) even then it would only take about 1.4 days and you'd be over it.
Yes this is all to do with something that I repeat regularly - variance.
... but why is this 5% number good?
Because people who don't understand variance and mining don't get worried about the pool luck for more than a day or so.
A lot? of the people on the pool get that variance is a fact of life, but also that in the long run it's OK as long as we aren't too small.
While we are still big enough and not disappearing off the hash rate graph, but still growing, we're OK.
