Topic: [0 GH/s 0% fee SMPPS] ArsBitcoin mining pool! - page 30. (Read 123750 times)

Yeah. Quicker blocks = good.

Yes, excellent news.  Three recent blocks <1,000,000 shares each.  They should confirm rather quickly, and then into the buffer.

No u were right with ur ansewer, also metball, and twmz helps me to understand someething that was unclear. Ty!

BTW:
Last 10 Blocks Found
141,643   1   User 791   2011-08-19 16:45
141,639   5   maximF   2011-08-19 16:15
141,625   19   AD / DuffMan   2011-08-19 13:35

This r really good new, arent this ?

Not sure.  oo-oo, what say you?

Either way, things are looking a little better today.

Isn't oo-oo asking how about the future PPLNS and not asking about how it works now?  In PPLNS, you get credited when blocks are found based on your contribution over the past N shares.  That goes in an unconfirmed balance.  When that block gets 120 confirmations, that amount moves to the confirmed balance.  When your confirmed balance gets over your configured threshold, a payment is sent.  There are no buffers in PPLNS.

You need to look at the Pool Stats, then notice Total PPS and Paid PPS Work.  As the SMPPS Buffer gets replenished (i.e. good luck), then the difference between those two gets smaller. (i.e. gettin' paid)

Right now, the difference is -145.18059901.  The buffer only has +54.81940099.  So that still leaves -90.36119802 to be paid out.  
There are 200 BTC unconfirmed, but as those get confirmed, more work gets done, thus increasing the Total PPS Work owed to miners.  
The buffer really needs to be (what, like 5x?) bigger than the difference between Total and Paid for SMPPS to keep up day after day without going into a deficit.

It's like if all the workers are simultaneously letting water out of a rain barrel.  Unpaid Work is just a way to keep the barrel from getting completely empty before it rains again.  Think of this as a block drought.
Every payment method is subject to luck.  

In other words, what Meatball said.  

That's not how it works.  As long as the pool has Confirmed BTC available in the buffer, you get paid for your shares immediately.  If there isn't enough confirmed BTC available you build up your "Unpaid" shares and as new blocks get confirmed those will start getting paid out.

Basically we have to wait for 120 future blocks to confirm a found block before the pool can start using those BTC to payout.

So basically, as soon as we find a block we r going to be paid ?

changing the payment scheme is a big deal and not something you should rush into,

I'd like to continue the "grand SMPPS experiment" and see what happens, but I would very much like to see the pseudo PPLNS system side-by-side to compare the two options.

I've made arsbitcoin my home for the past month or two for a number of reasons:
1) purely sentimental, this is where I found my first (and only) block
2) it has pretty good uptime (compared to some others I've tried)
3) it has a low (zero) fee
4) and I get low stales (around 0.3% lifetime - I have gotten up to 10 times that at some other pools)

I've read a bit about pool hopping and how it can affect payouts for steady miners in proportional pools, I'm curious how it affects us? (I realize that it shouldn't have an effect, but I've also seen that our buffer has been steadily going down ever since the hashrate started spiking like it has in the last few weeks)

whatever payment plan discourages pool hopping the most, while still offering low fees and good performance (low stales) is where I wanna be.

Ah, ok, sorry  Good to know!  I incorrectly mixed the reward system with the payment system.

No, this is exactly the same as Eligius' SMPPS-- we do not prioritize (nor track) any shares.

So, I said that I could try and do PPLNS within 24 hours, but that isn't going to happen.  It turns out there are too many code changes that need to happen and that need to be tested.  The actual PPLNS calculation code is the easy part.  I can easily perform PPLNS payment calculations based on when we find a block now on ArsBitcoin.  However, making sure that everything is 100% correct, making sure we handle blocks that go invalid, showing unconfirmed rewards, confirmed rewards, PPLNS round history, personal user stats that work with our existing stats, etc, is all proving to take longer than anticipated to implement.

On top of that, I have some important deadlines at work tomorrow, and I just can't spend any more time on this right now.

So, now Ars becomes a grand SMPPS experiment as we venture back into unlucky negative buffer territory (which is what we started with, btw, as we went negative the first time a share was submitted and it didn't find a block.)

Hopefully our luck turns soon, but if not, I thought I would again explain how payouts work when we can not pay all of the Total PPS work.

So, various miners have a Total PPS work > Paid PPS Work.  This is because the pool doesn't have BTC to pay anyone.  Once we get more BTC, the following occurs:

Each miner has an 'ideal payment' of Total PPS Work - Paid PPS work.  For example, Miner 1 could have 100 Total PPS Work, 99 Paid.  His ideal payment is 1 BTC.  Miner 2 has 50 Total PPS work, and 48 Paid PPS work, or an ideal payment of 2.

Lets say the pool just confirmed a new block that is worth 1 BTC.  First, we get the ratio of what we can pay everyone.  Total outstanding ideal payments in this 2 miner pool is 3 (2 + 1.)  So, our total ratio that we can pay people is (1 / 3) or 0.333333etc.  

This means that Miner 1 gets (1 * 0.333333etc) BTC and Miner 2 gets (2 * 0.3333333etc) BTC, for a total of 1 paid out (all that the pool has available.)

Note that this isn't exactly like Eligius' SMPPS, where older shares get paid first I believe.  It also isn't exactly like ESMPPS or RSMPPS or whatever.  It pays all miners equally based on their 'ideal pay' when the pool has BTC available to pay.  If someone stops mining and still has some 'ideal pay' amount, it will slowly get paid as the pool gets BTC.  However, people who DO stay will obviously be building MORE 'ideal pay', so they can be sure that they are still getting paid a much higher share of the pools BTC than anyone who stopped mining.

So, in summary.  I still plan to work on PPLNS in the near future, but it will not be done in the next 24 hours.  Our bad luck seems to be continuing, so we will most likely not be able to pay 100% PPS value instantly like we have been able to for the past month+.  However, the pool will obviously still be solving blocks and still pay out as much as it can, just like any other pool out there.  

If time goes on and we still have a negative buffer in the future, I still think we should switch to PPLNS.  What this means is that, at that time, it would no longer be possible to make Total PPS work = Paid PPS work.  Just like if you mine at any (non straight PPS) pool where the pool luck is < 100% over the time you mine there, you will not be able to get paid 100%.  

Sorry that the luck party has decided to visit others for a while (damn Eligius and their 900 BTC+ buffer like we used to have  ).

Probability has been slapping us around a bit. It's no fun...

Kind of in a pickle here.  Going on 9 hours this block while waiting on a ~10,000,000-share block to confirm to replenish the depleted buffer.

Just make sure to roll it out at the beginning of a new block, after you have posted a warning on arsbitcoin mainpage that you are about to do it.  A competing pool switched mid-block without sufficient notice recently and it ticked many people off as an unprofessional thing for a pool op to do.

Even better!

This part at the bottom of the spreadsheet cracked me up:
"Quotes may be delayed up to 20 minutes."

That is because of our shitty luck lately   Based on hashrate, we are a bit higher (3.7%?)

http://bit.ly/hashrate

Neat! Ars is currently at 2.2% of the entire Bitcoin mining power pool.

http://pident.artefact2.com/

True, and it's something every pool will go through. This is the first PPS pool I've been in, and the constant stream due to the huge buffer we had in place was pretty sexy.

If I were you, I'd implement the "pseudo" PPLNS, to show what we'd be making if we were on it. It'd give us a nice comparison.

Well, I think we've created a little too much negative perception ourselves about a negative buffer.  You have to remember that changing to PPLNS won't help us find blocks any faster, it will just distribute the payouts differently when we find them.

A slightly negative buffer has no real ill effect on existing miners.  It is actually mostly just detrimental to new miners looking to join, because they have to help pay miners who already have unpaid credit.

That said, I think I could do a switch to PPLNS sometime in the next 24 hours or so.  Some of the underlying code is already in place.