Moving from slush to where with 29 Th/s ? NOW: competition pool vs. pool.

[gnu123]

I did like this :

created 1 new account on slush with a new payoutaddress.

created 1 new account on kano with a new payoutaddress.

Started 1 S7 Antminer on each account 2015-12-28 13:30.


Now i will see what happends when the day goes by.

Payouts so far :

Slush : 0.0 BTC

Kano : 0.00048526 BTC


Both miners are hashing about 4,8 TH/s so this will be an intresting competition to follow up.
Both miners live in the same serverroom and they eat swedish high quality electricity.



But as you know, greed is a powerful thing..... So today i created another account on kano and put another S7 in work there becouse i have a feeling that kano will produce better. Maybe im wrong, time will tell. I will not be the first person in history making bad decissions couse of greed. ;-)

But the payout is not the most important, what disturbs me most is that i dont understand, and can not change the strange luck on slush with about 82% luck over a very long time. It just doesnt feel right.

Then we will se how this competition goes on, I guess that slush will win the first week, but after 2 weeks kano will lead the race. If im wrong, atleast i have tried and maybe learned a bit about the PPLNS payout system, i have read a little bit about it, but i dont understand so much, yet. My English is not the best...

[gnu123]

This is how it looks tonight running since 2015-12-28 13.30

First kano


Then slush




Right now the payout :

slush : 0.01161327 BTC

kano : 0.00048526 BTC

One S7 Antminer on each account.
If i add more miners i will do it on other accounts with other payoutaddresses just to keep this statistic.

Some Minerporn :

Bellow we se one quad KNC saturn built from junk out of a scrapbox, hashes about 500 Gh/s and is very quiet.

It´s  name is "The Beast"


Bellow one Antminer S7, not a very good copy, hashes about 4,8 Th/s clocked at 700Mhz, do not like to be overclocked. NOT very quiet.
It´s name is "Nummer Ett" , Number One.



How do i post pictures on this forum directly so there is no need to click them for viewing ?

[gnu123]

Day 3 of hashing.

The payouts :

kano : 0.0150691 BTC

slush : 0.02318437 BTC

Slush has gained more coins, for now. i Hope, and think that kano will catch up and tke the lead , but in how many days ?


The "luck" at slush is still very bad, with better luck it would have gained even more coins.

Both miners are doing well, besides a little break of about 1,2 hours yesterday when kano was down due to a ddos attack.

Do people really have such a boring life that they are doing ddos attacks on mining pools ? WTF ! And i thought my life was shitty.

Any input appreciated, the PPLNS system is not fully understood by me.
And i dont know what to expect.

[jonnybravo0311]

Here's how I explained PPLNS on my own pool's thread...

Sorry about the confusion Rozo.  Let's see if I can explain this without even using Bitcoin. 

Assume you have an empty glass in front of you.  Now, you start pouring water into the glass.  Every now and then I'm going to come by and take a look at that glass and measure how much water you've got in it.  The first time I come by, you've got the glass only 1/5th of the way full.  I mark that down and pay you for 1/5th of a glass of water.  I come by a while later and you've got the glass 2/5 full.  OK, I mark it down and pay you for 2/5 of a glass of water.  This continues until the glass is full of water.  Once your glass is full I pay you for a full glass of water, but I also now open a valve at the bottom of the glass so that water drains out as fast as you're pouring it in.  In this fashion, your glass will always be full of water and every time I stop by, I see you've got a full glass and pay you for it.

Eventually, you stop pouring water into the glass.  However, that valve is still open, so the water is draining at the same rate you were pouring.  When I stop by the next time, I see you've only got 4/5 of a glass, so I pay you for that much.  The time after that, the glass is only 3/5 full, so I pay you for that.  This continues until the glass is empty.

In my story above, every time I stopped by is equivalent to the finding of a block.  I see how much water you have in the glass and pay you for however much is there.  My story assumed I stopped by at even intervals of time, so you'd always know when to expect me.

So, how does luck play into it?  Well, every time I stopped by, I made sure to calibrate that valve on the bottom of the glass.  If I stopped by more often than expected, I'd mark you down and pay you.  If I stopped by less than expected, the valve would clog up and the water you were pouring in would slosh over the edge of the glass and be wasted.  When I finally did stop by, I would notice you had a full glass, and pay you for it, then fix the valve so that the pour rate equaled the drain rate.

OK... maybe I shouldn't try explaining things after I've been drinking

[swilla]

Once you hit your 5N, at current pool hash rate you will be getting ~.02 per block with kano. So it will take ~10 blocks on slush to reach the same amount. On days like yesterday, when we hit 3 that takes ~30 blocks with slush. Stick with kano and you won't be sorry. Since march of this year my little 5T setup has gotten me over 10 bitcoins with kano. Not bad at all IMO.

[gnu123]

Thnx jonnybravo !

That was a very good way of explaining the PPLNS system !



Thnx swilla for info, I intend to stick with kano.

Right now i have moved all S7 miners to kano without the one that must stay on slush so i can be able to compare.
Totaly 24 Th/s for now.

[kano]

Yeah it's the usual comment about what the 5N actually means

I put this on the help page a while ago:
https://kano.is/index.php?k=payout

PPLNS acts like the reward 'ramps up' when you first start mining.
What actually happens?

The 5Nd means it takes that long to reward your shares.
The ramp isn't missing rewards, it's delaying them to reduce variance.
Each share is rewarded in all the blocks found in the 5Nd after the share.
That's simply how it reduces variance. Each share's reward is averaged out over the 5Nd after it.
The pool hash rate for the last day is roughly 5.42PHs which means the 5Nd 'ramp' is roughly 4days 7hrs.

So your total reward for each share will currently depend on blocks found after each share for "roughly 4days 7hrs"
But that's just dependent upon the pool hash rate, it's the 5Nd that decides the rewards.
So, e.g. if you stopped mining and the pool was 10.8PHs, it would take "half as long in time" to payout your remaining shares.
If you stopped mining and the pool was 2.7PHs, it would take "twice as long in time" to payout your remaining shares.

The amount of the payouts as you start and finish are not affected by the the 5Nd.
The 5Nd just says how long it takes.

The thing that affects the indivdual/daily/whatever you choose payouts is the pool luck.
The pool had very good luck just after you started, so your current total is of course actually higher than expected for such amount of mining vs the 5Nd

Looking at the payouts based on a short period is actually not all that different to saying:
"I rolled a 6 on this dice on the first roll, so I will always roll a 6"

The best example of that is 'red october' where for the full month the block luck was 50%
Anyone who only mined during october and no other time got paid shit. Oh well, that's how luck works.
Conversely, anyone who was mining from 19.5wks go (16-Aug) until now should be above the 100% mark ... and that includes mining in october.

--

Now for the two things that it would appear that a lot of miners have no idea about
1) You can't predict random block luck, the future random block luck is unaffected by previous random luck.
There is no statistical expected difference at all chasing luck vs not moving around.

2) Pools all include factors in their luck that are not luck.
They like to hide this fact coz it makes them look bad.
Them's the breaks.
Those controllable factors are lowest at ckpool.org and kano.is vs every other non-SPV pool.

[gnu123]

Thnx, i think i understand.

 I also read : http://www.etcwiki.org/wiki/Bitcoin:_PPLNS_vs_PPS

[gnu123]

Day 4 of hashing

The payouts :

kano : 0.0150691 BTC

slush : 0.04655816 BTC


Happy new year to You !

[gnu123]

Day 5 of hashing

The payouts :

kano : 0.05815938 BTC

slush : 0.05831845 BTC

Almost the same !


If nothing strange happends, slush will take the lead and keep lead for 2-3 weeks, then kano will pass.
That is only what i think becouse of that the first payouts from kano was very low due to the PPMLN system, and that will take a week or two
to catch up for kano.


Right now the "luck" on slush is :

day 106%
week 80%
month 79%

Strange for a 39 Ph/s pool with that bad monthly luck for a long time, maybe the statistics is wrong and slush will outperform kano in a month ?
Dont know.

[kano]

Well yeah with kano.is after you stop mining, you expect, on average, i.e. if luck was always 100%, to get a total of 2 more full payouts, matching previous blocks, at the rate you were mining for the 5Nd before you stopped:
4/5 + 3/5 + 2/5 + 1/5 = 2
As usual, that depends on luck during the 5Nd after you quit, but those are the average expected rewards

Edit: Also note, in PPLNS, this 'ramp' does exist, no matter what anyone else says, when you start and stop, no matter how the pool does their N.
It must always exist since, when you start mining at a pool, you can never guarantee that you started at or before the start of the oldest shift used in the payout of the next block.
Even if their N is as tiny as only 1 hour, if you happen to start at half way through the N, and a block is found while your first shift is still included, you won't have a full shift of shares for the N so your first reward will of course be lower than having a full shift of shares.
That is of course expected since you are being paid for less shares, but the side effect is that each reward appears to 'ramp' up when you start (and 'ramp' down when you exit)
That 'ramp' can in fact sometimes be zero, but that would depend on the N size in PPLNS, the pool size and the pool luck at the time.
No one knows when the next block is, so yeah if some pool you are mining at thinks they do know in advance when they will get blocks, expect problems mining there

[gnu123]

Well yeah with kano.is after you stop mining, you expect, on average, i.e. if luck was always 100%, to get a total of 2 more full payouts, matching previous blocks, at the rate you were mining for the 5Nd before you stopped:
4/5 + 3/5 + 2/5 + 1/5 = 2
As usual, that depends on luck during the 5Nd after you quit, but those are the average expected rewards

That i fully understand.

What i want to prove with this little competition beween the two pools is mostly that something seems to be strange with a big pool doing about 80% for a long time.

Maybe im wrong..... maybe im just a foilhat seeing a conspiracy or scam everywhere, time will tell. ;-)

If both pools where doing an average of around 100% over a three month period and one pool take 2% fee and the other 0,9% fee the payout difference would be small.

I know people saying that there can be no problem with slush becouse its so big, have existed for so many years and so many people are mining there so any problem would be discovered and handled with fast. Maybe.... Maybe the luck will turn and stay at 120% for six month.....

[kano]

Yeah, I was just pointing out for anyone looking at the comparison and didn't realise, that when it ends there will be a period after that of more BTC from kano.is

[gnu123]

 

Day 6 of hashing

The Payouts

kano : 0.05815938 BTC

slush : 0.07002296 BTC

[dodziu]

Day 6 of hashing

The Payouts

kano : 0.05815938 BTC

slush : 0.07002296 BTC

is it daily payout or total?

kano : 0.05815938 BTC   slush : 0.07002296 BTC  

if daily, please tell me how did you do it

[xleejohnx]

Day 6 of hashing

The Payouts

kano : 0.05815938 BTC

slush : 0.07002296 BTC

is it daily payout or total?

kano : 0.05815938 BTC   slush : 0.07002296 BTC  

if daily, please tell me how did you do it

Haha this is going to be total for the 6 days

[d57heinz]

This is how it looks tonight running since 2015-12-28 13.30

First kano


Then slush




Right now the payout :

slush : 0.01161327 BTC

kano : 0.00048526 BTC

One S7 Antminer on each account.
If i add more miners i will do it on other accounts with other payoutaddresses just to keep this statistic.

Some Minerporn :

Bellow we se one quad KNC saturn built from junk out of a scrapbox, hashes about 500 Gh/s and is very quiet.

It´s  name is "The Beast"


Bellow one Antminer S7, not a very good copy, hashes about 4,8 Th/s clocked at 700Mhz, do not like to be overclocked. NOT very quiet.
It´s name is "Nummer Ett" , Number One.



How do i post pictures on this forum directly so there is no need to click them for viewing ?

Right away i see an issue with the way you started this test.  By looking at you image of slush i can clearly see you was on that pool prior to being on kano. It takes a couple hours to ramp up on slush. could you please show the entire reward history for that account.. Then i can see how much earlier you started on slush then kano.. But as it is. there is no way you could have gotten full reward on that first block at 13:00 on slush had you not been mining minimum of two hours prior. what i think you should do is  make two new accounts on both sites and move the miners to those sites at the exact same time. or as close as possible.. Within couple minutes..  you need to show us the shifts when you started kano mining and when you started slush mining.   then you will have an accurate test.. wait a month or two then take both machines off both pools and see where you stand after both pools show 0 ths.

best Regards
d57heinz

[gnu123]

It´s the total rewards, its only one S7 on each pool. If it was daily out of a S7 we would all be rich :-)


d57heinz :


As you say, something seems to be not right with the time. I dont know how/why becouse both miners were started within 20 minutes of eachother.
I have absolutely no intention of giving any of the pools an advantage but 20 minutes more or less doesnt matter after two weeks.
Maybe different timezones or wrong time on pool i dont know but the hashing has been running on both pools started amost the same.

What matters more is that no S7 hashes exactly the same as one other. I got 6 copys of S7 , I choose the two who hashes almost the same, but still after some days it will not be 100% accurate.


What i´m fishing after in this test is the rather big difference in luck between these two pools. Smaller differences is not possible to measure with this type of simple test. Also kano was down due to dDos for 1,2 hours before yesterday.....

This is no exact science, but if the 20% difference in luck realy exists it will show up.

I can not restart the test just to get thoose 20 minutes right becouse the owner of the serverroom, who also is paying the electricity which i get for free, he dont like slush anymore so he wants me to move the last S7 from slush quite soon, and as long as he pay the electricity he´s word goes.
So some more weeks il run the test, then all miners must work at Kano´s labor camp or they will get no more food.

BBT with more pics and payoutdata, and now i will write that it is total 

[usukan]

Thanks for running this test gnu123

Very interested - been at Slush for a long time but jumped to Kano-ck a couple of weeks ago - and glad I did!

I sat with Slush thinking luck would turn - but it just does not even till now.  Some days it looks OK - then back to 12hrs plus to wait for blocks - its really quite unbelievable that it goes on for months.

It all went to shit at Slush when the hash went up from 15 PH/s

Slush seems to have an effective hash of 15-20 Ph/s - not what is reported at 37-40?

I just cannot understand why the faithful stay when its clearly just not working anymore.

Early days at a couple of weeks for me but no comparison with Slush vs Kano so far.

Look fwd to your ongoing updates - great work!

It might be worth benchmarking the results against a PPS pool like Antpool PPS or F2Pool. 

For instance at current difficulty F2Pool payout (with fee of 4%) is 0.00464763 BTC / Thash/s
https://www.f2pool.com/help - but you need an account to check payout updates

Cheers - usukan

[vortexz]

Something fishy happens on slush, I belive they may show more hash rate than there actually is ...