Topic: [OLD] Eligius: ASIC, no registration, no fee CPPSRB BTC + 105% PPS NMC, 877 # - page 94. (Read 458370 times)

Great new feature !!!

no reply on the idle notification....

I apologise in advance for sticking my nose in, but a reliable way to estimate long term luck is mean(diff 1 equivalent shares per round/network difficulty at which block was solved) and then calculating the Erlang distribution CDF. Over the last 2123 rounds, the mean shares per round /D is high at 1.057, and the CDF is 0.977 - so pretty unlucky. However for the last 1000 rounds the mean shares per round / D is only 1.023 with CDF 0.769 - not very unlikely (or unlucky). I also took a quick look at  some boxplots and they didn't appear to be unusual.

Without completing a proper analysis I can't be sure, but Eligius' luck doesn't look unlikely.

Thanks for the reply.. I think its time to release those backup coins to the masses.

And just wanted to say.. Thank You for the great work - stats, merged mining and everything

I admittedly was being a bit of a smart ass with that opening line, no offense intended.

I'm glad the remainder of the post was informative.

-wk

"It's just normal variance" is a perfectly fine explanation, as long as you can show that falls in the normal range. The 40-day/50% explanation is a plausible one.

Not even sure what explanation you're looking for.  Asking me to explain why we've had some bad luck is like asking me when exactly we'll find our next block... neither of which anyone is really capable of answering definitively.

We can speculate, however, and ff you're trying to say there is some underlying issue with the pool which "causes" bad luck, then you're welcome to go through the data yourself and try to find a cause.  I've checked and double checked everything about the pool from top to bottom to make sure everything works as intended.  I also regularly scan the shares database for any duplicate work submissions (which would indicate the pool giving useless duplicate work to miners) and find none.  I also check the top miners for block withholding, statistically, and have never found anything outside of normal variance.  Trust me, I'm pretty paranoid and I've gone over everything as much as possible.

The only issue the pool has ever had which would contribute to bad luck is the overwhelming amount of orphaned blocks that occurred back when the pool was running SMPPS (over a year ago).  With our new server our orphan % has been very low. (1.5% over the past 3 months compared to 4% all time).

But, I think your math's result is mainly due to variance recently introduced by the recent leaps in network difficulty and the increase in the pool's overall hash rate over the past year, both add substantial variance to your broad all time difficulty_sum/accepted_sum stat.  Those sums include 19 months of counts, but, over 50% of the shares accepted over all time by the pool have been accepted in just the past 40 days! 40 days is not enough time for variance to level out.

Here's some interesting data I came up with fiddling with the stats database just now:

• 50.51% of all accepted shares were accepted in the past 40 days.  This would make any "long term" variance stats which includes recent data pretty irrelevant.
• From start of the pool to block 240000 overall pool luck was 99%. (5437253099.88402 difficulty / 5490920747 accepted shares). A good glipse at a period of time before the exponential rise in difficulty where the variance of our pool hash rate vs network difficulty would actually have had time sufficient time to level out, and it did as expected.
• The worst luck for a 60 day window was about 88%.  The best was 111%.
• A miner who started mining at the pool on day one of CPPSRB and mined at the exact same hash rate since then will have been paid just over 98% PPS as an all time average.

Given time, the variance of the diffculty/accepted ratio will average out to be 100% minus % of orphans (Law of Large Numbers).  With the network difficulty rising sharply, it will probably take a while for this to happen, and until then we'll swing on either side of that variance, along with every other pool.

Since all pools are subject to variance, and most other pools have fees... you're still probably best off with Eligius.  I'll be the first to say, pool operator or not, if I thought I could get more coins for my hash power elsewhere I wouldn't be mining at Eligius myself.

-wk

Edit: fix some typos, add a couple of links

There's nothing wrong with CPPSRB itself, but if I add up all the shares in the block list http://eligius.st/~wizkid057/newstats/blocks.php , I get a total of 29618979908 shares found, with total difficulty of 27904252293. This means average luck over the last 19 months (all of the stats) is only 94%. That seems awfully low to be just a case of bad luck over such a long period.

Any explanation ?

Greetings Eligius Miners!

Hope everyone is doing well.  Aside from dealing with forum trolls, I've been busy working on code to add some more features to the pool.

The Eligius stats now fully support individual worker names appended to your mining address with an underscore.  For example:

1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF_blade35

You can view my stats for 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF and see my individual worker hash rates in the hash rate graph, along with the aggregated stats for all workers. I'm currently running some FPGAs and USB block erupters on my normal 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF username, and my four ASIC Miner Blades on their own usernames: 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF_blade35, 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF_blade36, 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF_blade37, and 1EXfBqvLTyFbL6Dr5CG1fjxNKEPSezg7yF_blade38. (I'm sure someone is curious: the numbers are the suffix of their static IPs on my LAN).

You can see that each worker is functioning individually.  I started testing with them all mining under the worker "bladestest", as you can see, then switched them to their own individual worker names.

As of right now, each worker's variable difficulty is calculated individually, so, your workers will likely get lower difficulty work when they're split up, for now... Luke-Jr is working on an update to the Eloipool open source pool software that will help me get variable difficulty setup per user across all connections, as well as allow the minimum overrides in the user control panel, to work properly.

Back to worker names.  If you have more than one miner running (not more than one device in one software, but miners with multiple connections to the pool) you can now give them unique worker names, desired.  The normal way of just mining with them all having the same bitcoin address as a username will still work as normal.  However, if you want to be able to see stats for them individually, simply append an underscore _ and a short alphanumeric worker name to your bitcoin address that you use for a username.  See my examples above.  You will then be able to see their individual hash rates on your stats page.

I will be adding some more functionality around this soon, such as a table of worker names, their hash rates, accepted/rejected shares, etc.  I'm also working on a beta of worker-idle email notifications to go with this, and many other features.

Stay tuned!

-wk

You still make no sense...




You obviously are incapable of following the simplest of instructions and still have not read the FAQ on the CPPSRB Reward System.  

Since you're not mining here and are not supporting the pool, please stop trolling this thread with your nonsense.

-wk

It is the moment I've referred in my post about fixing the thing which is not broken. On the live pool. What a smart move, again.

Definitely is. It costs me BTC2.51 to prove it is. I'm happy 

Technically yes, but there are enough pool-wide shelved shares to ensure that the pool is not likely to have any "reserve" and time soon.

-wk

Hey what about the reserve... AFAIK when the pool is lucky the extra coins are stored in cold storage and manually redistributed... is that correct ?!

Probably going to be taking a closer look at this in the near future.  My main concerns are the security implications of allowing the transferring of shares in the first place.




I'll assume this has corrected itself by now.  Generally this happens in the time the payout queue list is cached and the balances list is cached with slightly newer data.




It definitely is!




-wk

Huh?

Default payout, 0.17219640 BTC total, and it says this:

+1000

Something I mentioned on IRC:

It would be nice to offer people some compensation for their shelved shares.  That way, you can get paid immediately by selling your shelved shares to another buyer.  Similar to selling bad debt, in the real world.

Each shelved share is worth more, the closer to the top of the stack you go.  Therefore, you would want to give some indication how close to the top of the stack are the shelved shares for which you are bidding.

CPPSRB is a good system.

I think you are doing a great job. Once i reach roi, i will start paying constantly to you.
I think it is good that you are answering questions here in such detail, it really increases your credibility.
Sad to see though how people with such negligence can burst such accusations.
Chin up. Keep calm.
Is there anywhere where one can read about your infrastructure? I mean, what are your preventions to DDOS attacks and such? As i believe this could be targets for upcoming rogue pools to sink all other pools. Or will that have no effect on their own mining?