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

[DryMartini]

OK so 237573 and 237575 are now confirmed. My reward for 237575 went from 0.002-ish to 0.0003-ish. 237574 which was mined at almost the same time is still not listed ( http://blockchain.info/block-index/385234/00000000000000f595dabc67b09229d6cd93224da85c46415f5c782e09ab576f )

Luck has been with us, no doubt. I just wish technology were too.

[pikeadz]

18184 is showing a small return for me.  That's 3 in 24 hours.  I am pretty new here but something doesn't seem right.

[mikesheadroom]

OK so 237573 and 237575 are now confirmed. My reward for 237575 went from 0.002-ish to 0.0003-ish. 237574 which was mined at almost the same time is still not listed ( http://blockchain.info/block-index/385234/00000000000000f595dabc67b09229d6cd93224da85c46415f5c782e09ab576f )

Luck has been with us, no doubt. I just wish technology were too.

I also experienced a near 80% rewards drop for block 237575.
Until Slush chimes in, we can only speculate that some sort of hardware error occurred during that window of time including blocks 237573 and 237574.
It's a shame to see BTC just vanish into oblivion.

[rsbriggs]

I also experienced a near 80% rewards drop for block 237575.
Until Slush chimes in, we can only speculate that some sort of hardware error occurred during that window of time including blocks 237573 and 237574.
It's a shame to see BTC just vanish into oblivion.

Am also having big drop - almost 2 orders magnitude - for 237575.  Average reward being .0045  
Reward for 237575 being  .000084421.
 
Similar situation blocks 237642 and 237540.

Work for 237574 gone disappeared.

Rewards on before and after sides of problem blocks seeming normal.

[mb]

Whoa. Finally got through the hoops to be able to post here..

I've been observing the reward calculation and seen some weirdness too. Including the estimated reward for the current block.

Maybe it's not just a matter of databases not being updated correctly, but some data type overflow. Miscalculation seems to happen when pool score passes 4G. If some intermediates are calculated as 32-bit values, bad things can happen to the outcome.

Just hope that Slush can look into this and possibly confirm this bug that has been biting us. More or less since the move to EC2, which could explain different compilers/platforms when compared to the ovh hosted machines.

[bspurloc]

I also experienced a near 80% rewards drop for block 237575.
Until Slush chimes in, we can only speculate that some sort of hardware error occurred during that window of time including blocks 237573 and 237574.
It's a shame to see BTC just vanish into oblivion.

Am also having big drop - almost 2 orders magnitude - for 237575.  Average reward being .0045  
Reward for 237575 being  .000084421.
 
Similar situation blocks 237642 and 237540.

Work for 237574 gone disappeared.

Rewards on before and after sides of problem blocks seeming normal.

What happened is Slush must have adjusted it, but fixed it wrong not realizing 237574 was skipped.
when it got confirmed the Shares that mined 237574 were applied to block 237575 and since 237574 was a very short block we got a very low pay out on 237575. Before that adjustment my pay out for 237575 was too high but not high enough to make up for missing block 237574.
 With block 237574 being completely skipped with no pay outs.

[TheUnk]

Doesn't this scoring method encourage pool hopping?

I mean, what is to stop someone for creating a tool to mine this pool for only the first 15mins or so of each new round and then jump over to another pool, a PPS perhaps.

Lets say your payout average is .25 a block on Slush and .20/hr on a PPS. Lets also say that out of 20 rounds, Slush has 5 10min blocks and 15 2hr blocks . The hopper spends 4hrs or so on the pool and gets full payment for 5 blocks in addition to a consistent rate from their PPS pool in between.

Over the 30 hours:
The hopper gets 1.25 from Slush, + 5.2 from their PPS. Earning 6.47 BTC.
Loyal slush miners get full payment for 20 blocks. Earning 5 BTC.


If slush had 20 rounds all 2+ hours:
The hopper would get nothing for their 5 hours, and they only spend 5 hours even if some of those rounds are 4+ hours, but the other 35+ hours spent on a PPS they earn at least 7 BTC.
Loyal miners would earn 5 BTC over the 40+ hours.

[TomKeddie]

Doesn't this scoring method encourage pool hopping?

I mean, what is to stop someone for creating a tool to mine this pool for only the first 15mins or so of each new round and then jump over to another pool, a PPS perhaps.

Lets say your payout average is .25 a block on Slush and .20/hr on a PPS. Lets also say that out of 20 rounds, Slush has 5 10min blocks and 15 2hr blocks . The hopper spends 4hrs or so on the pool and gets full payment for 5 blocks in addition to a consistent rate from their PPS pool in between.

Over the 30 hours:
The hopper gets 1.25 from Slush, + 5.2 from their PPS. Earning 6.47 BTC.
Loyal slush miners get full payment for 20 blocks. Earning 5 BTC.


If slush had 20 rounds all 2+ hours:
The hopper would get nothing for their 5 hours, and they only spend 5 hours even if some of those rounds are 4+ hours, but the other 35+ hours spent on a PPS they earn at least 7 BTC.
Loyal miners would earn 5 BTC over the 40+ hours.

You need to read more of this thread to understand the scoring.  The scores are weighted towards the end of the round to stop this.  A hopper doing what you describe would get little or nothing from most rounds.

[TheUnk]

Doesn't this scoring method encourage pool hopping?

I mean, what is to stop someone for creating a tool to mine this pool for only the first 15mins or so of each new round and then jump over to another pool, a PPS perhaps.

Lets say your payout average is .25 a block on Slush and .20/hr on a PPS. Lets also say that out of 20 rounds, Slush has 5 10min blocks and 15 2hr blocks . The hopper spends 4hrs or so on the pool and gets full payment for 5 blocks in addition to a consistent rate from their PPS pool in between.

Over the 30 hours:
The hopper gets 1.25 from Slush, + 5.2 from their PPS. Earning 6.47 BTC.
Loyal slush miners get full payment for 20 blocks. Earning 5 BTC.


If slush had 20 rounds all 2+ hours:
The hopper would get nothing for their 5 hours, and they only spend 5 hours even if some of those rounds are 4+ hours, but the other 35+ hours spent on a PPS they earn at least 7 BTC.
Loyal miners would earn 5 BTC over the 40+ hours.

You need to read more of this thread to understand the scoring.  The scores are weighted towards the end of the round to stop this.  A hopper doing what you describe would get little or nothing from most rounds.

What part did I not understand?  They get full pmt for the short rounds they were part of entirely and nothing for the long rounds they left on.

[DryMartini]

Oh well. I'll go back to 50btc and mine my regular 0.04 a day. The 10-50% extra I can make here just aren't worth being on call 24/7. It's a shame though, cause this has always been my first choice.

The bottom line is a business critical system, which this actually is, can not be this unstable. I'm a developer myself and I would be without work if I shipped applications with bugs of this magnitude. I'm not just talking about the last 24 hours, but the last month at least. As far as I know this is totally unprecedented in other pools, and we shouldn't be expected to put up with it.

My advice to you Slush is:
Take the pool offline and fix all problems with disappearing blocks and flucturing rewards. I'm sure you can do it again.

[ewitte]

You need to read more of this thread to understand the scoring.  The scores are weighted towards the end of the round to stop this.  A hopper doing what you describe would get little or nothing from most rounds.

It doesn't stop it its usually fairly easy to tell when there are multiple hours left in a block.  You can come back in the last 30 minutes and get most of the reward.  I don't do so but use the longer blocks to mess around/tweak hardware.

[nottm28]

its usually fairly easy to tell when there are multiple hours left in a block.

is it?

[bspurloc]

its usually fairly easy to tell when there are multiple hours left in a block.

is it?

If you take the assumption that since 3hours have passed then it will end soon. soon as in 3 hours later or in the case last week 8 hours later....   SARCASM.
 You have to get lucky and you will most likely lose more BTC than gain. unless you are doing 20gh/s+ then it doesnt much even matter just misdirected greed

[ewitte]

its usually fairly easy to tell when there are multiple hours left in a block.

is it?

Watching the % completed.   With about 90-95% accuracy I believe I was only wrong once. 

- when your at 20-30 minutes and the % complete is less than the number of minutes
- when it takes 5+ minutes to count up .1 %

[mb]

Nice tried, but.. no. Doesn't work that way.

[ewitte]

Nice tried, but.. no. Doesn't work that way.

*usually* - Like I said only missed it once.  I'm usually only down less than 30 minutes for hardware tweaking so plenty of time to jump back in.

[bspurloc]

Nice tried, but.. no. Doesn't work that way.

*usually* - Like I said only missed it once.  I'm usually only down less than 30 minutes for hardware tweaking so plenty of time to jump back in.

ok we at 3+ hours 92%+ move them all in! cuz it will 100% end between now and the next 8 hours

[Lucko]

Nice tried, but.. no. Doesn't work that way.

*usually* - Like I said only missed it once.  I'm usually only down less than 30 minutes for hardware tweaking so plenty of time to jump back in.

ok we at 3+ hours 92%+ move them all in! cuz it will 100% end between now and the next 8 hours

100% will never happen...

[ewitte]

There is a difference in knowing *when* something will end and just knowing it has a long ways to go.  BTW if you get to 99.xx % and it takes 5 or so minutes to go up .1% your screwed.  It was 99.xx% for like 5 hours on that 11 hour block lol.

[bspurloc]

100% will never happen...
[/quote]

So close though!!!  I am just being sarcastic.

 I rather get the correct payment for block 237575 and also paid for block 237574. and waiting. tap tap tap

edit...

Gotta love the 400+gh/s increase near the "end" of these long rounds. like I said the big boy greed