Hi
Before I start I'd like to say that aTriz & nearmiss deserve a lot of credit for the work they have put in. I am also impressed with their responsiveness to questions and suggestions.
As a lazy miner I am really pleased with both the automatic selection of coins to mine and the automatic conversion to
BTC.
With that in mind I hope you take this posting in the spirit it is meant
I carried out a very unscientific review of the performance yesterday by saving and analysing the Past Round stats from the website.
I recorded rounds 4758 - 4803. In 18 hours the total recorded mining time was 12h 30m 50s which, if the stats are accurate, equates to only 69.5% productive mining time. If this is the case then I would suggest that concentrating on improving the transition between coins should be high on the development list.
I have also noticed a massive change in the performance of my miners when working with the different coins, up to 40% difference in hashrates on a stable rig. When setting the weighting for each coin is this taken into account?
As you can tell I do like to research and crunch the numbers. With that in mind my own personal wish list has better stats at the top. I would love to be able to look at more detailed historic data, beyond the Last 30 Rounds currently visible on the website. I would like to see the actual
BTC paid out as well as keep My Estimated Earnings, for comparison. I would also love to see the profitability as calculated at coin selection time and the actual profitability once paid out.
We always appreciate the feedback!
Regarding the switching of coins, making this transition smoother is already in the works and we are hoping to start some tests on it sometime in the next few days. We are holding off on a second stratum server until we get this one sorted out. Though to the point of time lost....I'm not entirely convinced its as bad as the client reports at times. This is just my own opinion based on my own superficial evidence, but while testing things awhile ago, I'd notice during switches the client would be lagging in response, still showing disconnected, despite the servers stratum logs logging shares being submitted. I'm not sure if that lag is also reflected in mining time stats when you close the client or not.
Either way, we recognize its an area to address, and its in the works.
Regarding the stale/reject rates of certain coins (especially FST as of late, which we've been hitting a fair bit), each coin has a penalty applied to it based on average stale/reject/orphan rates encountered while mining. FST specifically, for example, is hit with a 35% penalty. This means it has to be 35% (plus a bit more for standard difference thresholds to trigger a switch) more profitable than the current coin before we'll hit it. Each coin has its own penalty. We review the actual btc results of these coins to ensure our numbers are still accurate from time to time. So far, FST has done ok for us. Other coins have been removed from the rotation, as we found the rejects/stales could not be properly counter-balanced by penalties, and btc profits suffered.
As for stats, more stats are always ideal, and we've got a few ideas on the backburner. As you can imagine, the TODO list is ever growing. With the holiday season approaching this is probably not realistically something we'll be able to tackle in the short term,but we recognize the desire for it, and hope to be able to deliver more stats at some point.
