Quote from: kano on November 29, 2013, 01:40:18 AM

Quote from: stridergt on November 29, 2013, 01:11:20 AM

https://dl.dropboxusercontent.com/u/109686709/minepeonstats.jpg

The percentage is DAcc / Diff1 as seen in the screenshot, so Diff1 are all shares counted in difficulty1.

Maybe Disc(arded) isn't displayed in difficulty1 so DAcc + DRej + (sum of Disc * difficulty) = Diff1

Am I getting this right? Which means that I had much lower stale percentage depending on the pool.

No. You've pretty much got everything wrong.

"Discarded" counts for nothing in the amount of work you are doing.
It's simply work you could have used but didn't use due to an LP.
Ignore it.

"DAcc / Diff1" is random and pretty much meaningless.

As per the FAQ:

Code:

Q: Why don't the statistics add up: Accepted, Rejected, Stale, Hardware Errors,
Diff1 Work, etc. when mining greater than 1 difficulty shares?
A: As an example, if you look at 'Difficulty Accepted' in the RPC API, the number
of difficulty shares accepted does not usually exactly equal the amount of work
done to find them. If you are mining at 8 difficulty, then you would expect on
average to find one 8 difficulty share, per 8 single difficulty shares found.
However, the number is actually random and converges over time, it is an average,
not an exact value, thus you may find more or less than the expected average.

Multiplying ANYTHING by "Difficulty" is wrong.
Work difficulty can (and usually does) change.

LP=longpoll?
"This request is not answered by server until it wishes to expire current block data, and new data is ready. The answer is the same as getwork on the main connection. Upon receiving this answer, miner should drop current calculation in progress, discard its result, and start working on received data and make a new request to a long polling URI."

I think you misunderstood me DAcc + DRej + (sum of Disc * difficulty) = Diff1
With the parenthesis I meant a quantity "DDisc", like DAcc and DRej , so the sum of these three converging parameters would eventually converge to Diff1
In this light would DAcc / Diff1, as a ratio between the convergent quantities, give after some period of time an indication of the quality of the pool?

PS Maybe the luck of the pool could skew results but if there is some consistency in detecting a difference between the ratios of two pools -all else equal- for a long time period, it could mean a better pool setup/efficiency/uptime, but I am beginning to get lost in all these.. :-)