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

[The Avenger]

As I mentioned a couple of pages back, I've managed to get an average of 100% of expected reward by mining at Slush's pool. The scoring algorithm penalises pool hopping more than standard proportional, but there are still gains to be made if you know exactly how far into a new round you should leave.

[1534547652]

1534547652

[PostMixer]

Check page 742. It's all there if you read carefully.

[The Avenger]

Check page 742. It's all there if you read carefully.

organofcorti suggests a figure can be calculated of what duration into a round is a good time to leave if you have not found a block. I'd assume, as it's time, it would be minutes or seconds or something like that. Page 742 says

The "RESULT" is the Erlang distribution cumulative distribution function p-value.

which, as I said, does not make any sense to me, as I'm not a stats expert.

I'm just asking because I'm curious. If it's some kind of Fight Club rule not to spell it out, then fair enough.

[organofcorti]

Y U NO EQUATIONS AND VARIABLES !

It's all in the blog post. You will have to figure out 'c' for yourself, since it changes over time. In the post I link to an R script that does it for you - you don't have to do it by hand. Look for "Navigator" in the post.

'c' changed from 300 to 200 about two years ago, but I don't know what it is now.

Edit: I think PostMixer is referring to this post on page 739.

https://bitcointalk.org/index.php?topic=1976.msg5053668#msg5053668

[organofcorti]

The "RESULT" is the Erlang distribution cumulative distribution function p-value.

which, as I said, does not make any sense to me, as I'm not a stats expert.

I'm just asking because I'm curious. If it's some kind of Fight Club rule not to spell it out, then fair enough.

It doesn't explain how you can mine strategically, but in case you were wondering, this post contains a bunch of quotes asking and answering the p.value and CDF questions in a different way and might be easier to understand: https://bitcointalk.org/index.php?topic=77000.msg3614424#msg3614424

[misternanyte]

I'm so sick of this. I leave slush and go to another pool, check on slush without pointing the miners back to slush, see that the block times are like 1 hour and less for the most part. Proceed to point the miners back to slush, and now we are back in the fucking 11's and 15 hours.

I'm so sick of this conspiracy bullshit.

And then 4 in a row, super quick. Damn this conspiracy 

I'm just saying it kept happening to me over and over again. I'm tired of it. Thank god... it's back to normal.

[The Avenger]

It doesn't explain how you can mine strategically, but in case you were wondering, this post contains a bunch of quotes asking and answering the p.value and CDF questions in a different way and might be easier to understand: https://bitcointalk.org/index.php?topic=77000.msg3614424#msg3614424

I appreciate the replies, but I just can't figure it out.

In the hitchhiker's guide to the galaxy, the answer to "life, the universe and everything", according to Deep Thought, was 42.

On that basis, I will assume the answer to your hop point equation is, give or take, 42 minutes.

Tell lambert_w to go home  

[PostMixer]

I just HAD to know 


The organ of Corti (or "spiral organ"), found only in mammals, is part of the cochlea of the inner ear and is provided with hair cells or auditory sensory cells. [1] It evolved from the basilar papilla found in all tetrapods, except for a few derived species that have lost it.

[Trefdraeth]

I'm so sick of this. I leave slush and go to another pool, check on slush without pointing the miners back to slush, see that the block times are like 1 hour and less for the most part. Proceed to point the miners back to slush, and now we are back in the fucking 11's and 15 hours.

I'm so sick of this conspiracy bullshit.

And then 4 in a row, super quick. Damn this conspiracy 

I'm just saying it kept happening to me over and over again. I'm tired of it. Thank god... it's back to normal.

Dude, it's always been 'normal', that's the point  .

We roll with the punches.

Forget the witchcraft and nonsense, average is average.

We are all, pretty much average.

IF there was a magic formula, it'd have regularised and published in open source years ago. Everything else is ego and nonsense.

Onwards and upwards. Before it becomes too difficult to mine for anyone other then Th/s corporates....

[organofcorti]

I'm so sick of this. I leave slush and go to another pool, check on slush without pointing the miners back to slush, see that the block times are like 1 hour and less for the most part. Proceed to point the miners back to slush, and now we are back in the fucking 11's and 15 hours.

I'm so sick of this conspiracy bullshit.

And then 4 in a row, super quick. Damn this conspiracy 

I'm just saying it kept happening to me over and over again. I'm tired of it. Thank god... it's back to normal.

You have to expect variance. Here's a few links to show you how the average/expected can vary:

Average over one round: http://www.wolframalpha.com/input/?i=Erlang+distribution%2C+k+%3D+1+%2C+rate+%3D+1

Average over three rounds: http://www.wolframalpha.com/input/?i=Erlang+distribution%2C+k+%3D+3+%2C+rate+%3D+3

Average over ten rounds: http://www.wolframalpha.com/input/?i=Erlang+distribution%2C+k+%3D+10+%2C+rate+%3D+10

[organofcorti]

It doesn't explain how you can mine strategically, but in case you were wondering, this post contains a bunch of quotes asking and answering the p.value and CDF questions in a different way and might be easier to understand: https://bitcointalk.org/index.php?topic=77000.msg3614424#msg3614424

I appreciate the replies, but I just can't figure it out.

In the hitchhiker's guide to the galaxy, the answer to "life, the universe and everything", according to Deep Thought, was 42.

On that basis, I will assume the answer to your hop point equation is, give or take, 42 minutes.

Tell lambert_w to go home  

Argh!

OK, a worked example. Slush might hate me for doing this, but:

Code:

D = 2621404453.0646
hashrate = 769984.646
assume c=200 (although it might have changed - you'll have to find out for yourself).

Then hop point = 0.0164293+1.14254/(1.8747*2621404453.0646/(769984.646*200)+exp(1))
                     = 0.04942183


Shares mined before 0.04942183*D = 86486769 d1 shares are submitted will, on average,
be worth more than expected. Shares mined after this point will be worth less than expected.

I should reiterate the above is only valid if c is still 200. Since c influences variance to a great extent, it will be decreased when Slushes proportion of the network increases, and be increased when  Slushes proportion of the network decreases. Since the Slush's proportion has decreased significantly since it was changed to 200, c may be back to 300 by now.


Edit: That's more like 8 minutes rather than 42 minutes.

[organofcorti]

I just HAD to know 


The organ of Corti (or "spiral organ"), found only in mammals, is part of the cochlea of the inner ear and is provided with hair cells or auditory sensory cells. [1] It evolved from the basilar papilla found in all tetrapods, except for a few derived species that have lost it.

Try telling that to a creationist.

[boumalo]

8 Blocks in 4hours today, pretty good! Of course my miner was not hashing at full speed so I only earned 1/6th of what I should have

[The Avenger]

Edit: That's more like 8 minutes rather than 42 minutes.

Looking at the last 30 blocks on Slush, if you left after 8 minutes, you'd only have hit one block. Certainly not a way to maximise earnings by hopping, so I'm sure Slush won't mind the example 

With the hitchhiker, err, "Deep Thought coefficient of pool hoppi-ness", you'd have hit 8 blocks before jumping.

I think you may have inadvertantly proved science fiction trumps statistics I jest!

I really appreciate the worked example, thanks organofcorti.

[gourmet]

Got a clue why do i lose 5/6% hashing power, so strange.

Do you lose five sixths of a percent of your hashing power? (That's 0.008333... of your power?)
How have you determined this strange fraction?

[misternanyte]

I'm so sick of this. I leave slush and go to another pool, check on slush without pointing the miners back to slush, see that the block times are like 1 hour and less for the most part. Proceed to point the miners back to slush, and now we are back in the fucking 11's and 15 hours.

I'm so sick of this conspiracy bullshit.

And then 4 in a row, super quick. Damn this conspiracy 

I'm just saying it kept happening to me over and over again. I'm tired of it. Thank god... it's back to normal.

Dude, it's always been 'normal', that's the point  .

We roll with the punches.

Forget the witchcraft and nonsense, average is average.

We are all, pretty much average.

IF there was a magic formula, it'd have regularised and published in open source years ago. Everything else is ego and nonsense.

Onwards and upwards. Before it becomes too difficult to mine for anyone other then Th/s corporates....

Just seems like all these sites HATE ME PERSONALLY. It's like, I'm trying to sell stuff in an auction online, and these assholes who have JUST REGISTERED continue to block the transactions by clicking "buy" and never paying. So I gotta wait at least 4 days to repost the auction. It's happening over and over and over again. I'm tired of it. I was using paypal to send my BTC to USD money to because paypal is the easiest way to do it, and then guess what all of a sudden they shut it down. Then the attacks happen, and blah blah blah.

At least they didn't destroy us.

[gourmet]

I think I read somewhere a while back that it is possible for the difficulty level to actually go down. If so, I wonder if the recent issues with the exchanges, the price drop, and what seems to be an unlucky streak that is allegedly not exclusive to just this pool could cause a slow down or reversal of the difficult change? Any thoughts?

OoC has explained the mechanism (for the 15^th time). :-)
I'll only add that IMHO it's not too likely in the near future for the difficulty to go down, as new ASICs are still produced and new people are entering into Bitcoin. The difficulty increase may slow down a little bit, but I don't believe it would reverse any time soon.

[PostMixer]

How quickly I forgot what a 9+ hr block feels like :-)

[gourmet]

I'm so sick of this. I leave slush and go to another pool, check on slush without pointing the miners back to slush, see that the block times are like 1 hour and less for the most part. Proceed to point the miners back to slush, and now we are back in the fucking 11's and 15 hours.

I'm so sick of this conspiracy bullshit.

And then 4 in a row, super quick. Damn this conspiracy 

I'm just saying it kept happening to me over and over again. I'm tired of it. Thank god... it's back to normal.

You're sick of moving away and back again? :-)))
(Looks like the simpliest solution would be to stop doing so...)

[gourmet]

It's all in the blog post. You will have to figure out 'c' for yourself, since it changes over time. In the post I link to an R script that does it for you - you don't have to do it by hand. Look for "Navigator" in the post.

'c' changed from 300 to 200 about two years ago, but I don't know what it is now.

Edit: I think PostMixer is referring to this post on page 739.

https://bitcointalk.org/index.php?topic=1976.msg5053668#msg5053668

On the News page of the Slush site, dated 2013-09-25, Slush has posted:

While working on completely new and quite exciting rewarding system, we've changed some settings of the current one in order to improve stability (e.g. decrease variance) in miners' income. It has already been deployed.

The change is about altering shape of used scoring function. The function is now less steep than before which means that your submitted shares influence your score for a longer time period. Therefore your score is less sensitive to instant luck of your miners and leads to a lower variance in results.

I've interpreted it as increasing the c constant for myself at that time. Or might the altering shape of used scoring function mean something more substantial?