put btcpool on info.. they are broken.
They've also started IP banning and tax 50% shares if you're there for less than half the round - even if you are a 2% donator.
Ok they are my new most hated pool.
If you are having difficulty determining when deepbit or BTCGuild have completed a block then you could try adopting an alternative pool-hopping approach of fixing one such pool and mining with them for the 5 minutes following every new block on the network. For each block found by the pool in question you are getting 5 minutes at the start of the round, very profitable.
HUh?.. Hop on deepbit for every single block found on the network? yeah thats pretty much the hopper not too long ago(not on purpose but in effect).. we are trying to fix that. It isnt very effecient. And you know using this method you describe we would be 60% of the time, wrong? The methods we are working on have a better success rate than that.
Blocks not found by the pool are essentially 5 minute chunks selected at random and, in the long term, won't be better or worse than 24-7 mining.
so why do it?
mining at another proportional pool will nullify the advantages of this method.
how so? most of our other prop pools not set to "mine_deepbit" WE DO KNOW WHEN THEY FOUND A BLOCK.. plus we have chart porn.. have you seen our chart porn?
.
If BTCGuild and deepbit are effective in hiding their block-discovery information however then this might be the best method.
Thanks for the thought, but I dont agree. There are many ways and yours is probably the first we considered.
You can also do the guess thing, but first see if any of your pools that give you proper info goes to zero, this reduces our error rate from 60% to about 20%
then their is pident, which has it;s own method for guessing who got the block, there is a fork of c00ws trying just to use that.
then there is the lp guess which works great for some pools for some people.. but not great for everyone.(and there are at least two different methods of this)
then there is the voting method.. which every client reports it;s best guess.. this should work better than the previous method for everyone.. and we are working on this method as well.
Sorry but I just dont think hopping on deepbit on every block announce will be as good as the things we have tried and it is the first thing people think to do, as deepbit finds 40% of all the blocks on the network.
I must admit I am not fully aware of the current pool-hopping capabilities and am speaking in general. I thought that one standard method that big proportional pools used to try to reduce pool hopping is to try and hide when new blocks were found by the pool. I only meant to suggest that even if they are perfectly successful and it is impossible to tell when a certain pool has found a block you could still use the idea behind pool hopping to make some gains using only the information in the bitcoin network. If you have any information on when a pool finds a block beyond the bitcoin network new-block announcements then you will certainly do better with the standard method.
A quick look at pident suggests that it is quite easy to tell when a certain pool has finished a block so I assume that the remaining problem is finding a pool which doesn't detect and ban pool-hopping behaviour. Again, I know little about the existing practical situation and am speaking purely in terms of theory and mathematics. I'm sure if I mined with a proportional pool I'd know more but I'm with simplecoin.us (which uses PPLNS) specifically so I don't have to worry about implementing pool hopping.
The "guess thing" you describe is a good idea when you have partial information.
All I meant to point out is that if one would like to pool hop with a big proportional pool and is having difficulty with the LP aspect then they can at least make some gains by using the information on the bitcoin network (if the pool is sufficiently large).