[4+ EH] Slush Pool (slushpool.com); Overt AsicBoost; World First Mining Pool - page 985.

dietwice

newbie

Activity: 58

Merit: 0

Hi Guys,

I've noticed that it's taking much longer for me to get confirmed blocks now. It seems that my unconfirmed count is increasing while my confirmed count is increasing at a much slower rate.

Is anybody else having this issue, maybe i'm just missing something here.

Thanks.

Sure.
The block confirmation time is measuring in blocks so it depends on the overall speed of the bitcoin network.
For now it's about 5 blocks an hour whilst usually it's about 6-7 blocks. So the 100 blocks confirmation now takes about 20 hours instead of 14-16 as usual.
Also there's a few minutes delay between dropping the block reward from the "unconfirmed" and adding it to "confirmed".
Don't worry, it won't affect your daily average.

slush

legendary

Activity: 1386

Merit: 1097

Quote from: OggerMC on August 04, 2011, 07:13:26 AM

so days like these are just badluck and nothing to do about it but to increase the overall performance?

Increasing overall performance will just make rounds shorter, but it didn't help with having rounds many times longer than expected average. Having many hours round is mostly psychological problem, it does not affect expected payout in longer timeframe (in days).

Personally I'm completely checking pool status everytime when such long round happen, but I'm perfectly calm when I found it isn't any technical problem, but just a "luck". Over a half of a year of running the pool learnt me that one round does not mean anything.

OggerMC

member

Activity: 93

Merit: 10

so days like these are just badluck and nothing to do about it but to increase the overall performance?

slush

legendary

Activity: 1386

Merit: 1097

About screening blocks - at this moment there's no known algorithm to improve probability to find a block faster than is expected average. Ideas like "change block when round is extra long" does not work

, every attempt (every nonce change) is almost perfectly random.

slush

legendary

Activity: 1386

Merit: 1097

Quote from: OggerMC on August 03, 2011, 07:13:28 AM

why did we got yesterday only half of the reward?

There is still no other answer than "the probability"

.

dietwice

newbie

Activity: 58

Merit: 0

Quote from: Xephan on August 03, 2011, 08:03:09 AM

Quote from: zx9r on August 03, 2011, 05:16:05 AM

They have an average of 21 min per block last 24 hours. What is our avg in slush ? My feeling is that it may be hours but didnt make the numbers.

BTW, they "only" have about 2x our hashrate. Too much bad luck here and good luck there ?

Minutes per block average in past 9 days or so
Deepbit 22.3 (Currently 5510 GH/s)
Slush 61.6 (Currently 2045 GH/s)

They've got 2.69x our hashrate, and average of 22.3 is only about 2.7% faster than expected... or we're 2.7% slower than expected

It isn't quite correct to compare the almost instant ratio with the 9-days average.
We had Slush at 19xx GH/s yesterday, and 22xx a day before with almost the same long time average - so the comparison results can differ significantly.

Anyway it would be good to raise the complete pool speed to get better stability and to aviod that multihour blocks.

Xephan

newbie

Activity: 42

Merit: 0

Quote from: zx9r on August 03, 2011, 05:16:05 AM

They have an average of 21 min per block last 24 hours. What is our avg in slush ? My feeling is that it may be hours but didnt make the numbers.

BTW, they "only" have about 2x our hashrate. Too much bad luck here and good luck there ?

Minutes per block average in past 9 days or so
Deepbit 22.3 (Currently 5510 GH/s)
Slush 61.6 (Currently 2045 GH/s)

They've got 2.69x our hashrate, and average of 22.3 is only about 2.7% faster than expected... or we're 2.7% slower than expected

OggerMC

member

Activity: 93

Merit: 10

why did we got yesterday only half of the reward?

cosurgi

sr. member

Activity: 298

Merit: 250

Quote from: Xephan on August 03, 2011, 04:07:12 AM

... like abandoning a block once it takes longer than x% shares to solve the way hoppers abandon pools after 41% ...

heh. in fact a block is abadoned at least once per second, when nTime changes. Add to that extraNonce and block is changing hundred of times per second.

zx9r

full member

Activity: 153

Merit: 100

Quote from: slippyrocks on August 03, 2011, 01:53:49 AM

slush i think deepbit is screening the blocks somehow

probably a bailout threshold like hoppers use

can we implement this pronto

check it

http://deepbit.net/stats

They have an average of 21 min per block last 24 hours. What is our avg in slush ? My feeling is that it may be hours but didnt make the numbers.

BTW, they "only" have about 2x our hashrate. Too much bad luck here and good luck there ?

Clipse

hero member

Activity: 504

Merit: 502

Quote from: Xephan on August 03, 2011, 04:07:12 AM

Quote from: cosurgi on August 03, 2011, 02:40:01 AM

Quote from: slippyrocks on August 03, 2011, 01:53:49 AM

slush i think deepbit is screening the blocks somehow
probably a bailout threshold like hoppers use

can we implement this pronto

we don't need this. our scoring method is designed specifically to be hopper-immune

https://forum.bitcoin.org/index.php?topic=1976.msg50002#msg50002

I don't he's talking about dealing with hoppers but rather he thinks deepbit is using some kind of method to screen blocks so they find blocks faster like abandoning a block once it takes longer than x% shares to solve the way hoppers abandon pools after 41%

I don't think something like that would work because if I'm not wrong, a new block depends on the Merkle root so you can't skip ahead. However, maybe it's possible to choose a different set of transactions and attempt to do a new block with them instead. Maybe slippyrocks would like to explain his theory.

You cant bail on a block, if a block isnt found then everyone mining is still mining for that block.

Each block isnt assigned some arbitrary share value on the network, that is a pool feature to split work up and pay users their share, nothing more.

Xephan

newbie

Activity: 42

Merit: 0

Quote from: cosurgi on August 03, 2011, 02:40:01 AM

Quote from: slippyrocks on August 03, 2011, 01:53:49 AM

slush i think deepbit is screening the blocks somehow
probably a bailout threshold like hoppers use

can we implement this pronto

we don't need this. our scoring method is designed specifically to be hopper-immune

https://forum.bitcoin.org/index.php?topic=1976.msg50002#msg50002

I don't he's talking about dealing with hoppers but rather he thinks deepbit is using some kind of method to screen blocks so they find blocks faster like abandoning a block once it takes longer than x% shares to solve the way hoppers abandon pools after 41%

I don't think something like that would work because if I'm not wrong, a new block depends on the Merkle root so you can't skip ahead. However, maybe it's possible to choose a different set of transactions and attempt to do a new block with them instead. Maybe slippyrocks would like to explain his theory.

cosurgi

sr. member

Activity: 298

Merit: 250

Quote from: slippyrocks on August 03, 2011, 01:53:49 AM

slush i think deepbit is screening the blocks somehow
probably a bailout threshold like hoppers use

can we implement this pronto

we don't need this. our scoring method is designed specifically to be hopper-immune

https://forum.bitcoin.org/index.php?topic=1976.msg50002#msg50002

slippyrocks

member

Activity: 69

Merit: 10

respecttheslider

slush i think deepbit is screening the blocks somehow

probably a bailout threshold like hoppers use

can we implement this pronto

check it

http://deepbit.net/stats

slush

legendary

Activity: 1386

Merit: 1097

Quote from: dietwice on August 02, 2011, 08:22:32 AM

Hello, difficulty!

But a difficulty raise is just about 13% - I think it isn't a reason for a 6 hour block.

Actually 6 hour block is still in "sane" range, with current difficulty and hashrate. It is still about probability - as far as I can tell, there's no technical problem in the pool.

slush

legendary

Activity: 1386

Merit: 1097