bit pit - (LP, ESMPPS, 8-decimal payout, SSL, API, 0% fee, Almost 0% Stales!) - page 12.

hugolp

legendary

Activity: 1148

Merit: 1001

Radix-The Decentralized Finance Protocol

Quote from: zoro on July 15, 2011, 09:58:59 AM

wow! 140GH! with 100GH we have 1 block/day. then prop is working just fine!
i would say to monitor the pool behavior for a while before any changes are made

The problem is that a lot of the miners will go away when the block becomes too long, as had happened in the 10th round. I am waiting for the changes to be implemented to return to the pool.

(I wish I had changed so quick because I missed the quick round and went to BTCGuild that had a horrible luck strike. Oh well.)

zoro

full member

Activity: 226

Merit: 100

wow! 140GH! with 100GH we have 1 block/day. then prop is working just fine!
i would say to monitor the pool behavior for a while before any changes are made

Luke-Jr

legendary

Activity: 2576

Merit: 1186

Quote from: 1bitc0inplz on July 15, 2011, 07:48:28 AM

SMPPS does not "carry debt". If/when the scenario rises that the total payout exceeds the block income plus the reserve funds, then the pool simply divides the block income plus the reserve by however many shares there were that round.

While it doesn't carry debt (because it never rewards more than it has), it does remember the difference, and try to make it up in the future.

Sukrim

legendary

Activity: 2618

Merit: 1007

Ah ok, then I must have misread that part or it was changed since I first took a look at it...

All in all then the pool might just pay less than PPS on unlucky rounds or PPS on all other rounds - seems like a really bad deal to mine there, if pool funds are low! Thanks for the heads-up!

1bitc0inplz

member

Activity: 112

Merit: 10

The way I understand SMPPS from reading here: http://eligius.st/wiki/index.php/Shared_Maximum_PPS
and from talking with luke-jr and reading some of his other posts...

SMPPS does not "carry debt". If/when the scenario rises that the total payout exceeds the block income plus the reserve funds, then the pool simply divides the block income plus the reserve by however many shares there were that round. This leads to a situation where the SMPPS shares are worth less than their advertised price. But, that is all. When the next round starts the the pool has 0 reserves and 0 debt.

The key part to this is the last part of luke-jr's pseudo code:

Code:

else:
   Pay each miner PPSamount / idealPayTotal * availableFunds

Nowhere in there does it mention "remembering" the difference between the ideal payout and the realized payout. It simply switch to "proportional" over the total available funds. So, technically speaking, unless the reserve fund is 0 when the payout exceeds the income, it never does true proportional.

Sukrim

legendary

Activity: 2618

Merit: 1007

How would you pay out on bad luck strikes then?

pseudocode:
IF round_shares > difficulty
pay proportionally
ELSE
pay per share

?
This would mean you would ALWAYS win as a pool operator and otherwise it would be like proportional...

Please explain how this should work "so it does not remembers negative strikes"

hugolp

legendary

Activity: 1148

Merit: 1001

Radix-The Decentralized Finance Protocol

You could make it either way, but honestly I would make it so it does not remembers negative strikes, because otherwise it will stop newcomers from joining the pool. At some point the pool should almost always have a positive credit and rarely touch 0, because variance should even out.

Sukrim

legendary

Activity: 2618

Merit: 1007

You should get paid eventually. If however a lot of people leave the pool it will get very hard (and very expensive...) for the remaining ones to pay the depts from the past.

With round-based payouts you have much higher variance than PPS, but you're not depending on what happened before that round (in SMPPS in essence the "round" goes on forever).

Since it can be expected that a lot of people might quit mining when the payout per block gets cut in half, if a SMPPS pool at that point of time has a big dept towards it's miners from bad luck, it will take a REALLY long time until that is being repaid (or it might not be repaid at all - I would strongly recommend to start at 0 depts/balance after each payout split with SMPPS, as you need 100% more luck to pay past shares.)

luffy

hero member

Activity: 607

Merit: 500

i think in SMPPS you get payed eventually when the luck changes from bad to good in the future? or you loose some shares for good?
i am confused!

Sukrim

legendary

Activity: 2618

Merit: 1007

Quote from: hugolp on July 15, 2011, 05:20:48 AM

In proportional if you have a bad luck streak you get paid less. Fullstop. In SMPPS if you have a bad luck streak you might be able to recover a part.

In proportional (or PPLNS for that matter) if you had a bad luck streak 2 weeks ago and now have good luck, you earn more. In SMPPS you are dependent on the whole payout history of the pool since it was started if you get paid at all for your shares.

hugolp

legendary

Activity: 1148

Merit: 1001

Radix-The Decentralized Finance Protocol

Quote from: Sukrim on July 15, 2011, 05:11:38 AM

Quote from: hugolp on July 15, 2011, 05:04:36 AM

Quote from: Sukrim on July 15, 2011, 04:59:26 AM

SMPPS has the disadvantage that the pool can be over time arbitrarily high in dept towards it's miners. If you had a maximum of hashrate during a bad luck streak, these miners will never get paid at all.

Yes, but they would not be payed in proportional either. So its not like you are loosing anything.

In proportional they would get paid as soon as the next block is found...

In proportional if you have a bad luck streak you get paid less. Fullstop. In SMPPS if you have a bad luck streak you might be able to recover a part.

Sukrim

legendary

Activity: 2618

Merit: 1007

Quote from: hugolp on July 15, 2011, 05:04:36 AM

Quote from: Sukrim on July 15, 2011, 04:59:26 AM

SMPPS has the disadvantage that the pool can be over time arbitrarily high in dept towards it's miners. If you had a maximum of hashrate during a bad luck streak, these miners will never get paid at all.

Yes, but they would not be payed in proportional either. So its not like you are loosing anything.

In proportional they would get paid as soon as the next block is found...

hugolp

legendary

Activity: 1148

Merit: 1001

Radix-The Decentralized Finance Protocol

Quote from: Sukrim on July 15, 2011, 04:59:26 AM

SMPPS has the disadvantage that the pool can be over time arbitrarily high in dept towards it's miners. If you had a maximum of hashrate during a bad luck streak, these miners will never get paid at all.

Yes, but they would not be payed in proportional either. So its not like you are loosing anything.

Sukrim

legendary

Activity: 2618

Merit: 1007

SMPPS has the disadvantage that the pool can be over time arbitrarily high in dept towards it's miners. If you had a maximum of hashrate during a bad luck streak, these miners will never get paid at all.

luffy

hero member

Activity: 607

Merit: 500

i like how the SMPPS pool arsbitcoin.com is working. it has ~0.5% stales though.
this pool will be better Wink

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Sad news... Apparently ~60% of our hashrate are pool hoppers. Time for SMPPS, me thinks.

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: luffy on July 15, 2011, 02:17:48 AM

you should set a poll in order to vote what system we want Wink

now i see that this ~0% stale pool would be ideal for SMPPS Wink

it has the potencial to be the best pool ever Cheesy

The problem with that kind of democracy is that people often don't fully understand the implications of what they're voting for.
Ideally we should just reach a consensus within a discussion here. It seems like SMPPS is being favored, so now the question would be which of the variants that I outlined above.

luffy

hero member

Activity: 607

Merit: 500

you should set a poll in order to vote what system we want Wink

now i see that this ~0% stale pool would be ideal for SMPPS Wink

it has the potencial to be the best pool ever Cheesy

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: owowo on July 14, 2011, 08:51:30 PM

I think SMPPS or PPLNS would discourage the hoppers, especially pplns.. since the block right now takes longer ppl are leaving and we dropped down 50% hashrate... I personally am more fascinated with prop, as it has more like a goldwashing experience ;o) ,... but for the pool to grow SMPPS or PPLNS would be better.

See http://forum.bitcoin.org/index.php?topic=12181.msg253239#msg253239 for an explanation why Score/PPLNS isn't good for a growing pool. (PPLNS will behave very similarly to Score, and actually I simplified Score to the PPLNS case in that post, so it should be applicable for both.)

Obsi

newbie

Activity: 30

Merit: 0

Quote from: 1bitc0inplz on July 14, 2011, 08:12:35 PM

Hey,

I just left a comment on that issue. I hope that helps explain the issue.

I am going to try and find cdhowie and perahps between the two of us we can find something. Also, though I don't think I mentioned it at github, we're also working with the nginx team to see if perhaps they know of a way to remove the erroneous trailing CRLF before it even enters Node.

Excellent, thanks!

Topic: bit pit - (LP, ESMPPS, 8-decimal payout, SSL, API, 0% fee, Almost 0% Stales!) - page 12. (Read 80771 times)