Quote from: Kuroth on August 06, 2013, 01:22:11 AM

Quote from: cryptopi on August 06, 2013, 01:07:38 AM

Wanted to provide a bit more information about the problem that I noticed has been going on in this pool lately in the hopes that it can be fixed or explained.

I actually take my mining stats from the pool and record them every few minutes (all day long) to see what is happening in order to evaluate pool performance. In doing this, I get a fairly detailed understanding of the workings and don't get misled by minor trends. My little program factors in all aspects of the mining output (including immature and unexchanged), not just the balance available at the end of the day to transfer out. I generate trend information data every few minutes, and summarize like data together until the trend is broken. I produce summary statistics during every like trend period. As mentioned before, have 2.4 Mhash pointed at the pool.

It generated about $10 a day 3 and 4 days ago (and I was not mining for about an hour each day). That level I defined as good. Then a couple of days ago I noticed that it appeared to start mining some coin with a very low difficulty (the miner just reports a constant string of "new block" messages). While this coin is being mined, statistics go to pot, almost nothing gets posted to the account in any form until the pool switches to another coin. I have had occurrences where my gross balance (including immature and unexchanged) actually go backwards while this type of mining is in progress for 30 minutes or more (some immature coins previously mined get killed, and there is nothing productive to replace it). Typically, you can also see the pool's total reject rate skyrocket (mine as well), moving on multiple occasions to 30% of the total shares submitted by everyone in the pool. It appears that everyone is caught working on obsolete shares or something. Once the pool moves back to a "normal" coin, things go back to normal and the coin mining rate trends back to normal rates (was about .005 bitcoins per hour previously, for last 2 days down to .0025 per hour, which are rates of approx $12 and $6 per day. The pool continues to jump into these coins for various periods, and stats go to pot consistently at these times. At one point, my shares were rejected over 50%. Not sure who is benefiting when these coins are being mined, but it certainly is not me. This has caused my total production to drop from the $10 per day reported earlier down to less than $6 now.

I understand that mining is a random payout and there are multitudes of factors affecting the payout. I have studied this particular effect in enough detail (and it is rather consistent), so I believe that it is not a random effect at this point. I also noticed that the pool had a few hiccups in the last day or so, and I have factored out these time periods.

The premise of this pool is great, and I'm a big fan. Love the idea, and think that if properly implemented the operator(s) could and should make plenty of $$$. However, unless the operator(s) can tackle this problem of autoswitching into dud coins, I'm going to have to mine elsewhere. I'm posting this in the hopes that the operator(s) are not aware and can look into this problem and make adjustments to protect the viability of the pool. Maybe the low difficulty coins are great for some segment of the miners (huge ones?) but certainly not for the 2.4 Mhash crowd like me.

Please advise if there is something that I am wrong about, as I don't want to move elsewhere but cannot ignore the facts...

Interesting Theory. But How can we be "autoswitching into dud coins"? it makes no sense.. I thought the whole point of auto switching was to try and always to be mining the most profitable coin?

Unfortunately it's much more reality than theory.

Basically "fast coins" have fast blocks enabled by lower difficulty ratios. Lower difficulty means easier and faster to solve. Faster transactions, faster block rewards etc.

This is all good and fine until the blocks start getting solved WAY too fast. Distributed computing is fast, but still has its limits.

I like the faster coins, they are great to mine and transfer WHEN the block rate is at or near specification.

Lets say all coins are limited by this distributed network (internet, p2p etc) to a maximum block generation rate of 1 every 5 seconds.

How many times can the designed average be exceeded on a 2.5 minute coin before reaching the 5 second limit VS a .30 minute coin?

So What happens when the blocks are solved faster than every 5 seconds?