ComputerGenie
|
|
February 28, 2017, 08:10:29 PM |
|
You altered the history when you told the S9v1 to leave without a long enough sample size. Had they stayed the variance would of caught back up. Any sample size should be bare minimum one year.
He didn't "alter" anything. S9s will forever suffer from the pre-January deficit and would have to span months of overages to "break even"; this is not the same as "altering history", it's forcing historical losses to coincide with current statistics.
|
If you have to ask "why?", you wouldn`t understand my answer. Always be on the look out, because you never know when you'll be stalked by hit-men that eat nothing but cream cheese....
|
|
|
kano (OP)
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
February 28, 2017, 08:33:02 PM |
|
You altered the history when you told the S9v1 to leave without a long enough sample size. Had they stayed the variance would of caught back up. Any sample size should be bare minimum one year.
He didn't "alter" anything. S9s will forever suffer from the pre-January deficit and would have to span months of overages to "break even"; this is not the same as "altering history", it's forcing historical losses to coincide with current statistics. Except that, as I said, people have still been mining with S9v1 on the pool non-stop. Just a lot fewer.
|
|
|
|
tournamentdan
Member
Offline
Activity: 118
Merit: 10
|
|
February 28, 2017, 08:36:18 PM |
|
S9v1
Yep I was asleep for that one Good to know im not the only one who find a block here on kano with S9v1 looks like all those "bad" miners still able to find blocks There's no doubt they find blocks, they've found 95 so far. The issue is the total number S9v1 has found is well below what it should have - CDF[Erl] = 0.991261 with that last one. (They're expected to have found almost 25 more) ... Had all those machines stayed they could have hit those blocks. ... Yes, and they 'could' have hit more, less, or even none. Doesn't matter what they 'could' have done, since that's supposed to be random. Unfortunately, history suggests otherwise. You altered the history when you told the S9v1 to leave without a long enough sample size. Had they stayed the variance would of caught back up. Any sample size should be bare minimum one year. Time isn't a factor in determining block luck statistics. The number of hashes and the number of blocks are the only 2 numbers that matter. I've no idea where you got 'one year' from. But they have also been hashing on the pool since I asked people to switch, just not as many. Time is a factor to figure for block luck statistics. You might have no blocks one day and ten the next. If you calculate the statistics in a short amount of time. You will not have a real number for what the statistics will be over a longer amount of time. But they have also been hashing on the pool since I asked people to switch, just not as many. Yes and that is part of my point. Had the majority of the s9 stayed. They could have gone on a lucky streak and come back even with the expected number. But we will never know now since you asked them to leave based on your statistics taken over a way to short of a time frame. At the same time you asked the s9 to leave. Someone asked about the expected amount for the new avalons. Which you replied there had not been "enough time to get a good idea of the performance of the avalon." Funny how time is a factor when it comes to the avalon. But not the s9.
|
|
|
|
tournamentdan
Member
Offline
Activity: 118
Merit: 10
|
|
February 28, 2017, 08:54:22 PM |
|
S9v1
Yep I was asleep for that one Good to know im not the only one who find a block here on kano with S9v1 looks like all those "bad" miners still able to find blocks There's no doubt they find blocks, they've found 95 so far. The issue is the total number S9v1 has found is well below what it should have - CDF[Erl] = 0.991261 with that last one. (They're expected to have found almost 25 more) You see this stattistic was directly affected by you asking the s9 people to leave. So in reality you can only count the blocks after the mass exodus of s9v1. What is the expected for the S9v1 from after we lost half of our hash rate?
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
February 28, 2017, 09:05:59 PM |
|
S9v1
Yep I was asleep for that one Good to know im not the only one who find a block here on kano with S9v1 looks like all those "bad" miners still able to find blocks There's no doubt they find blocks, they've found 95 so far. The issue is the total number S9v1 has found is well below what it should have - CDF[Erl] = 0.991261 with that last one. (They're expected to have found almost 25 more) ... Had all those machines stayed they could have hit those blocks. ... Yes, and they 'could' have hit more, less, or even none. Doesn't matter what they 'could' have done, since that's supposed to be random. Unfortunately, history suggests otherwise. You altered the history when you told the S9v1 to leave without a long enough sample size. Had they stayed the variance would of caught back up. Any sample size should be bare minimum one year. Time isn't a factor in determining block luck statistics. The number of hashes and the number of blocks are the only 2 numbers that matter. I've no idea where you got 'one year' from. But they have also been hashing on the pool since I asked people to switch, just not as many. Time is a factor to figure for block luck statistics. You might have no blocks one day and ten the next. If you calculate the statistics in a short amount of time. You will not have a real number for what the statistics will be over a longer amount of time. Incorrect. Again, it is ONLY "The number of hashes and the number of blocks" Neither have anything to do with time. But they have also been hashing on the pool since I asked people to switch, just not as many. Yes and that is part of my point. Had the majority of the s9 stayed. They could have gone on a lucky streak and come back even with the expected number. But we will never know now since you asked them to leave based on your statistics taken over a way to short of a time frame. At the same time you asked the s9 to leave. Someone asked about the expected amount for the new avalons. Which you replied there had not been "enough time to get a good idea of the performance of the avalon." Funny how time is a factor when it comes to the avalon. But not the s9. Incorrect. The time fact I was referring to was actually hashes - meaning for the time they had been on the pool, they had not expended enough hashes to be able to give a useful sample. Luck doesn't correct itself. Making any assumption that it does means you clearly have no idea about statistics. I posted in detail about this the other day also.
|
|
|
|
tournamentdan
Member
Offline
Activity: 118
Merit: 10
|
|
February 28, 2017, 09:25:39 PM |
|
Luck doesn't correct itself. Making any assumption that it does means you clearly have no idea about statistics. I posted in detail about this the other day also. Sure it does. Last month you were below the expected. This month you are above the expected. You found more blocks this month than you were supposed to. I wonder if that 20ph of s9v1 that left could have found more blocks than they were supposed to in the last few months to bring the "block found" statistic closer to the "expected block" statistic.
|
|
|
|
philipma1957
Legendary
Online
Activity: 4256
Merit: 8603
'The right to privacy matters'
|
|
February 28, 2017, 09:31:11 PM |
|
@ tournamentdan
do you agree with the world that 1+1 = 2
this is true because the human race made of a set of rules and we stick to it.
the s9 v1 has a cdf of .992
it has made around 95 blocks while it hashed enough to make 120 blocks.
we the human race made up a set of rules that say the cdf is .992
what this means if there was a 1000 groups of correctly functioning s9's v1 that hashed enough to make 120 blocks
we would have lost out to 992 of them and beat 7 of them.
So at what point does shit luck turn into shit software?
I have no idea. but 120 blocks of hashing making 95 blocks means death to me as a miner.
and as kano says the luck is not improving it is still worse then .99
I get your argument that if 20ph left and we now have 5 ph in s9 v 1 they can not move us back to a cdc of .80 to .70 to .60 since they do not have strength of numbers do it. but they are still at .992 which is easy to move to .950 even if the hash is 5 ph and not 20ph in s9's v1
.992 is so fucking bad that .85 cdc would move it to .97 cdf and that has not happened.
so when kano comes to the thread and tells us s9 v1 has improved from .992 cdf to .970 cdf then to .950 cdf
I would say your point could maybe come true.
|
|
|
|
beltsniffer
|
|
February 28, 2017, 09:34:33 PM |
|
Luck doesn't correct itself. Making any assumption that it does means you clearly have no idea about statistics. I posted in detail about this the other day also. Sure it does. Last month you were below the expected. This month you are above the expected. You found more blocks this month than you were supposed to. I wonder if that 20ph of s9v1 that left could have found more blocks than they were supposed to in the last few months to bring the "block found" statistic closer to the "expected block" statistic. We started losing 25ph in September and it was steady decline through to January. It started long before Kano said anything about upgrading s9v1 or asking people to mine elsewhere with those machines if they couldn't switch it.
|
|
|
|
ComputerGenie
|
|
February 28, 2017, 09:38:57 PM |
|
...S9s will forever suffer from the pre-January deficit...
...Luck doesn't correct itself....
See we can agree on some things
|
If you have to ask "why?", you wouldn`t understand my answer. Always be on the look out, because you never know when you'll be stalked by hit-men that eat nothing but cream cheese....
|
|
|
wmabern
|
|
February 28, 2017, 10:32:53 PM |
|
|
BITMIXER.IO Gone Baby, Gone.. ;-) Not any good sig campaigns out there that I want!
|
|
|
|
smutboy420
|
|
February 28, 2017, 11:41:13 PM |
|
All those recent ghs added PLEASE LEAVE!!!! It looks like we find blocks faster when we have less ghs going at it. Turns out there is a work around. You need to have your rigs in "tune with someone with some s9v2s so If you add smutboy420.yourusernameworkername (change workername for each workers name) on all your workers it will help the your luck outcome by a lot. lol well at least will make a YUGE difference in your payouts. J/k hopefully no one takes that to seriously. :-)
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
March 01, 2017, 12:21:56 AM |
|
Well, since the "Monthly Statistics" table shows blocks found during the month, and the next block will be found in March - it's now 20minutes into UTC March - looks like February has helped anyone who mined here in December and didn't do a runner, recover some if not all of their losses! Yay for February
|
|
|
|
beltsniffer
|
|
March 01, 2017, 12:26:45 AM |
|
Yay for February Yay is right! We needed recover from that month!
|
|
|
|
tournamentdan
Member
Offline
Activity: 118
Merit: 10
|
|
March 01, 2017, 04:10:51 AM |
|
Well, since the "Monthly Statistics" table shows blocks found during the month, and the next block will be found in March - it's now 20minutes into UTC March - looks like February has helped anyone who mined here in December and didn't do a runner, recover some if not all of their losses! Yay for February Yes, thanks for February. Which makes December and January hurt a little less in the butt.
|
|
|
|
philipma1957
Legendary
Online
Activity: 4256
Merit: 8603
'The right to privacy matters'
|
|
March 01, 2017, 04:19:44 AM |
|
Well, since the "Monthly Statistics" table shows blocks found during the month, and the next block will be found in March - it's now 20minutes into UTC March - looks like February has helped anyone who mined here in December and didn't do a runner, recover some if not all of their losses! Yay for February Yes, thanks for February. Which makes December and January hurt a little less in the butt. January was fine
|
|
|
|
beltsniffer
|
|
March 01, 2017, 10:39:42 AM |
|
Well, since the "Monthly Statistics" table shows blocks found during the month, and the next block will be found in March - it's now 20minutes into UTC March - looks like February has helped anyone who mined here in December and didn't do a runner, recover some if not all of their losses! Yay for February Yes, thanks for February. Which makes December and January hurt a little less in the butt. January was fine Especially when you add in an average block payout of 13.08
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
March 01, 2017, 10:48:52 AM |
|
Block! by schmokeandapancake S9v2 13.9 BTC ... and 2 payouts
|
|
|
|
smutboy420
|
|
March 01, 2017, 11:26:36 AM |
|
Block! by schmokeandapancake Smiley S9v2 13.9 BTC ... and 2 payouts Ah finally.. That bugger was like a trying to smash a titanium block. lol
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
March 01, 2017, 11:37:11 AM |
|
Block! by schmokeandapancake Smiley S9v2 13.9 BTC ... and 2 payouts Ah finally.. That bugger was like a trying to smash a titanium block. lol Well the CDF says that we expect, on average, 1 in ~22 blocks to be greater than ~308%, so, nothing unexpected.
|
|
|
|
|