Pages:
Author

Topic: âš’[CGA] Cryptographic Anomaly - The Elusive Coinâš’ - page 57. (Read 226291 times)

legendary
Activity: 980
Merit: 1004
I think it would be cool if CGA hooked up with a site like cryptovest.com

CGA would look uber sexy in some type of holographic metal coin.

Sounds pretty cool, but you don't want to trust your private keys to a third party. Guy might keep a record of the private keys and cash in all the coins ten years from now.
sr. member
Activity: 448
Merit: 250
I think it would be cool if CGA hooked up with a site like cryptovest.com

CGA would look uber sexy in some type of holographic metal coin.
sr. member
Activity: 322
Merit: 250
Spray and Pray
Try not to look at the pool stats to determine if there is a problem with the code, instead look at the block explorers(they should be accurate, except that they round the diff so the math may not play out as planned) and compare the result to what the pool stats claim. To be honest I have no idea how the pools display their stats but since the beginning there has been issues due to the way CGA functions.

Also, estimates are just that, estimates, and depending on how you do it depends on how accurate it will be but in crypto-coins nothing is certain, the time span you are looking at my be too small and the pools maybe having a string of bad luck and a large solo-miner maybe getting most of the reward blocks (out of luck or manipulation, which will be fixed in the update I am about to announce).

I do not understand why, for example: 58,041 % 3.89795634 = 0.43, did not have a reward, It maybe because of the way the "double" rounds the number after a certain decimal place. But "double" can handle very long decimals so I fear that isn't the problem. Anyway I was aware of such an issue but hadn't had it confirmed... The new update will not have this problem.

I will be releasing the update today. And will announce it here once it is on Git. It may take a day or two to get the Windows version but Mac and Linux will be available.
Hi Sawedoff,
Who decides whether there will be a reward or not?
As soon as the new code is released, you will have some nodes say that there should be a different reward than other nodes.
Will we need a fork to cleanly switch from one reward system to the other, or is it going to be pure luck whether you receive 1 CGA, 1/3 CGA or nothing after finding a block, since it will be rewarded differently under the two versions?
Do we need to make sure that the majority of all CGA nodes update at the same time and is there the risk of creating a fork when there is not a clear majority and the two camps disagree?

The new block reward will start on block 62,000, so there wont(shouldn't) be a problem with forks unless the pools or solo miners don't update on time. But block 62,000 should give everyone time. I will be making it a priority to let the exchanges know as well.

Edit: Block 64k that gives everyone a little over 48 hours (from now).
hero member
Activity: 595
Merit: 500
Try not to look at the pool stats to determine if there is a problem with the code, instead look at the block explorers(they should be accurate, except that they round the diff so the math may not play out as planned) and compare the result to what the pool stats claim. To be honest I have no idea how the pools display their stats but since the beginning there has been issues due to the way CGA functions.

Also, estimates are just that, estimates, and depending on how you do it depends on how accurate it will be but in crypto-coins nothing is certain, the time span you are looking at my be too small and the pools maybe having a string of bad luck and a large solo-miner maybe getting most of the reward blocks (out of luck or manipulation, which will be fixed in the update I am about to announce).

I do not understand why, for example: 58,041 % 3.89795634 = 0.43, did not have a reward, It maybe because of the way the "double" rounds the number after a certain decimal place. But "double" can handle very long decimals so I fear that isn't the problem. Anyway I was aware of such an issue but hadn't had it confirmed... The new update will not have this problem.

I will be releasing the update today. And will announce it here once it is on Git. It may take a day or two to get the Windows version but Mac and Linux will be available.
Hi Sawedoff,
Who decides whether there will be a reward or not?
As soon as the new code is released, you will have some nodes say that there should be a different reward than other nodes.
Will we need a fork to cleanly switch from one reward system to the other, or is it going to be pure luck whether you receive 1 CGA, 1/3 CGA or nothing after finding a block, since it will be rewarded differently under the two versions?
Do we need to make sure that the majority of all CGA nodes update at the same time and is there the risk of creating a fork when there is not a clear majority and the two camps disagree?
hero member
Activity: 595
Merit: 500
Try not to look at the pool stats to determine if there is a problem with the code, instead look at the block explorers(they should be accurate, except that they round the diff so the math may not play out as planned) and compare the result to what the pool stats claim. To be honest I have no idea how the pools display their stats but since the beginning there has been issues due to the way CGA functions.

Also, estimates are just that, estimates, and depending on how you do it depends on how accurate it will be but in crypto-coins nothing is certain, the time span you are looking at my be too small and the pools maybe having a string of bad luck and a large solo-miner maybe getting most of the reward blocks (out of luck or manipulation, which will be fixed in the update I am about to announce).

I do not understand why, for example: 58,041 % 3.89795634 = 0.43, did not have a reward, It maybe because of the way the "double" rounds the number after a certain decimal place. But "double" can handle very long decimals so I fear that isn't the problem. Anyway I was aware of such an issue but hadn't had it confirmed... The new update will not have this problem.

I will be releasing the update today. And will announce it here once it is on Git. It may take a day or two to get the Windows version but Mac and Linux will be available.
Hi Sawedoff,
Good to hear that.
I did verify my findings with the block explorer and I did find the same and more blocks that should have been rewarded but were not.
I only looked at a few pages of blocks (20 per page) and on every page was at least one block that was
- divisible by 3
- diff below 3
- reward = 0
So, it is a real problem in the CGA code.
Good that after today it will be history, I hope that the pools will update quickly.
hero member
Activity: 595
Merit: 500
I see. If no anomaly with diff<3 only thats OK
As I understand it, the soon-to-be-released version has a fixed 1/3 CGA per block reward *if* the diff was <= 3,
else the reward is depending on the diff: the higher the diff, the lower the reward, which makes it double-hard to
mine once the diff increases much above 3: both higher diff (less likely to find a block) and lower rewards.
The calculation is very simple: take a copy of the actual difficulty and call it d
if d < 3 then d = 3 //(make d minimum 3)
reward = 1/d        // block reward is inverse related to difficulty
So, you get a max reward of 1/3 CGA or less if the diff is higher.
Every block will be rewarded, some less than others (more miners => less reward)
sr. member
Activity: 322
Merit: 250
Spray and Pray
Cryptographic Anomaly =  CGA = 异常币

一个与众不同的币 Grin

我希望我们救了我们面前,但这些其他翻译:隐情,异常加密,神秘的币。

有几个人,但我不记得了。
sr. member
Activity: 322
Merit: 250
Spray and Pray
Quote
I'll do my best here to explain it from a technical point of view.
The block reward can either be 0 (zero) or 1 (one) CGA coin.

    1. The block reward is determined by first checking if the difficulty is above or below 3 (three). If it's BELOW 3, then go to the next step. If it's ABOVE 3, go to step 3.
    2. Divide the current block number by 3 (three), and check the remainder.
    2a.  If the remainder is equal to 0 (zero), then the block rewards one CGA coin. Done.
    2b.  If the remainder is not equal to 0 (zero), then the block does not reward any CGA. Done.
    3. Divide the current block number by the current difficulty, and check the remainder.
    3a.  If the remainder is LESS THAN 1 (one), then the block rewards one CGA coin. Done.
    3b.  If the remainder is NOT LESS THAN 1 (one), then the block does not reward any CGA. Done.

We know that the difficulty increases as the network's hashrate increases, so more mining will make this coin more elusive, but not impossible to find.
s4w3d0ff and others, please correct me if I am wrong. I hope this helps you understand the coin better!
can you explain step 3 more on the examples:
current block is 57452 and diff is 4.26714924
current block is 57453 and diff is 5.94869535

57452 % 4.26714924 = 3.36978188 > 1 = No anomaly
57453 % 5.94869535 = 0.50030969999 < 1 = Anomaly
Sawedoff,
I hope that I am wrong, but I think we have a problem in the current code....
For days I have been mining and noticed that all pools are paying *much* less than what I would expect from calculation,
so I decided to look into it more. First I checked how many blocks were found by pools to see if that would show that the
coin works as expected and what I saw is that the nr of found coins is *much* lower than can be expected from the
ratio between pool and network hash rates. I did take into account that the pool does *not* show blocks that have no reward,
the only blocks that you will see on a pool are the 1 CGA reward blocks and the blocks with 0 CGA but carrying TX fees (generally 0.001 CGA). What I would expect if the Diff is lower than 3, a pool should find 1/3 of the expected blocks with 1 CGA reward and an unknown nr with only TX fees but essentially the other 2/3 have 0 reward. If the diff is >3 then the pool should show on average a lower nr of blocks with 1 CGA reward and higher nr with no reward.
What I found was that even on the biggest pool (anomalypool) which has 2/3 of network hashing power, less than 10% of the blocks have 1 CGA reward, even though the diff is around 3 and you'd expect that more than 30% of the blocks carry a 1 CGA reward.
I verified with another smaller pool and a very honest pool operator, who did not know nor could explain why it regularly took more than 1000% of expected shares to find a block (with reward, since zero-reward blocks are not shown) and generally the stats were about 3 times lower in payout and finding then I could calculate.
Then I decided to take the test, grab a random block without reward (which can show up on the pool due to carrying 0.001 CGA TX fee)
and do the calculations for myself by hand. Take block with height 58,041 and verify the diff: 3.89795634
Now 58,041 % 3.89795634 = 0.43 !!!!
As far as I can see, this zero-reward block *should* have been rewarded with 1 CGA.
Can you verify the code that you are putting in the new release, because I am afraid that there is a bug in the CGA code so it does not pay out as expected and since network hash rate has gone up, we are seeing even fewer payouts, anomalies are not as much as calculated. My estimations are that we are seeing only half to as low as a third of the expected CGA rewards. That makes mining CGA unattractive...

Try not to look at the pool stats to determine if there is a problem with the code, instead look at the block explorers(they should be accurate, except that they round the diff so the math may not play out as planned) and compare the result to what the pool stats claim. To be honest I have no idea how the pools display their stats but since the beginning there has been issues due to the way CGA functions.

Also, estimates are just that, estimates, and depending on how you do it depends on how accurate it will be but in crypto-coins nothing is certain, the time span you are looking at my be too small and the pools maybe having a string of bad luck and a large solo-miner maybe getting most of the reward blocks (out of luck or manipulation, which will be fixed in the update I am about to announce).

I do not understand why, for example: 58,041 % 3.89795634 = 0.43, did not have a reward, It maybe because of the way the "double" rounds the number after a certain decimal place. But "double" can handle very long decimals so I fear that isn't the problem. Anyway I was aware of such an issue but hadn't had it confirmed... The new update will not have this problem.

I will be releasing the update today. And will announce it here once it is on Git. It may take a day or two to get the Windows version but Mac and Linux will be available.
full member
Activity: 219
Merit: 100
I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.
I disagree, I think this is uniq feature of this coin. But code must be checked on errors, no doubt
Unique or not - there are plenty coins with varying rewards (Doge to name a well established one)
but the all-or-nothing reward that the coin had until now is hard on small pools and solo miners,
only larger pools will find plenty blocks fast enough to average out the rewards, so why not give
the average straight away? then solo miners and small pools will not have the risk to get no rewards
for a long time while the whole idea of Anomaly remains: the more the coin is mined, the higher the diff and the *lower*
the coin release rate by dividing the rewards by the diff. No net change for the coin's issue rate, but a protection for the small guy.
Since I found a serious bug in the all-or-nothing reward algorithm not working as designed, it is better to switch sooner rather than later.
yes, it is better to switch if there is a bug, but if bug will be repaired - we should stay I think
Sawedoff already said that he released a version earlier with this change, but a problem was found so it was reverted back to the
0-or-1 reward until he makes a new release, but this is going to happen soon anyway. See post 2011:
https://bitcointalksearch.org/topic/m.5663581

I see. If no anomaly with diff<3 only thats OK
full member
Activity: 182
Merit: 100
Cryptographic Anomaly =  CGA = 异常币

一个与众不同的币 Grin
member
Activity: 66
Merit: 10
need date   new release!?!
hero member
Activity: 595
Merit: 500
I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.
I disagree, I think this is uniq feature of this coin. But code must be checked on errors, no doubt
Unique or not - there are plenty coins with varying rewards (Doge to name a well established one)
but the all-or-nothing reward that the coin had until now is hard on small pools and solo miners,
only larger pools will find plenty blocks fast enough to average out the rewards, so why not give
the average straight away? then solo miners and small pools will not have the risk to get no rewards
for a long time while the whole idea of Anomaly remains: the more the coin is mined, the higher the diff and the *lower*
the coin release rate by dividing the rewards by the diff. No net change for the coin's issue rate, but a protection for the small guy.
Since I found a serious bug in the all-or-nothing reward algorithm not working as designed, it is better to switch sooner rather than later.
yes, it is better to switch if there is a bug, but if bug will be repaired - we should stay I think
Sawedoff already said that he released a version earlier with this change, but a problem was found so it was reverted back to the
0-or-1 reward until he makes a new release, but this is going to happen soon anyway. See post 2011:
https://bitcointalksearch.org/topic/m.5663581
full member
Activity: 219
Merit: 100
I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.
I disagree, I think this is uniq feature of this coin. But code must be checked on errors, no doubt
Unique or not - there are plenty coins with varying rewards (Doge to name a well established one)
but the all-or-nothing reward that the coin had until now is hard on small pools and solo miners,
only larger pools will find plenty blocks fast enough to average out the rewards, so why not give
the average straight away? then solo miners and small pools will not have the risk to get no rewards
for a long time while the whole idea of Anomaly remains: the more the coin is mined, the higher the diff and the *lower*
the coin release rate by dividing the rewards by the diff. No net change for the coin's issue rate, but a protection for the small guy.
Since I found a serious bug in the all-or-nothing reward algorithm not working as designed, it is better to switch sooner rather than later.
yes, it is better to switch if there is a bug, but if bug will be repaired - we should stay I think
hero member
Activity: 595
Merit: 500
I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.
I disagree, I think this is uniq feature of this coin. But code must be checked on errors, no doubt
Unique or not - there are plenty coins with varying rewards (Doge to name a well established one)
but the all-or-nothing reward that the coin had until now is hard on small pools and solo miners,
only larger pools will find plenty blocks fast enough to average out the rewards, so why not give
the average straight away? then solo miners and small pools will not have the risk to get no rewards
for a long time while the whole idea of Anomaly remains: the more the coin is mined, the higher the diff and the *lower*
the coin release rate by dividing the rewards by the diff. No net change for the coin's issue rate, but a protection for the small guy.
Since I found a serious bug in the all-or-nothing reward algorithm not working as designed, it is better to switch sooner rather than later.
hero member
Activity: 595
Merit: 500
Quote
Koontas: compcentral, let see if BCX wants to test out CGA...we'll see how long it lasts....good luck. Adios.
compcentral: Koontas, ok.  BCX is a little bitch who hasn't done shit in ages, can't touch CGA.
Gotta love some the shid I get in pm.
80 second retarget and people think this coin isn't wide open for a brutalization?
~BCX~
Thought I'd let you know who Koontas is:
https://cryptocointalk.com/topic/7578-ebtcoin-dev-wants-to-drop-support-and-keep-premine/
I clicked on that link and all I see is that I am logged in, but "you do not have permission to access that topic".
It looks like someone covered up...
full member
Activity: 219
Merit: 100
I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.

I disagree, I think this is uniq feature of this coin. But code must be checked on errors, no doubt
full member
Activity: 219
Merit: 100
hero member
Activity: 595
Merit: 500
Sawedoff,
It appears that also for low diff there is a problem.
Take for example the block with height 57,996 with diff 0.8
57996/3 = 19332 so this is an Anomaly and
this block should have 1CGA reward, but the pool shows 0 reward (and a small TX fee)
I am afraid that many more blocks *should* have carried a reward, but incorrectly show 0 reward...
So, the earlier released description of the rewards is how it *should* have worked, but not what is actual...
That is the reason that there is such excessive times between finding a block with reward!
(and the fact that pool sw does not show the 0-reward blocks, you must be lucky to find an incorrect block
that is carrying TX fee so it shows up) All other truly 0-reward blocks found are invisible, so it is hard to see
that the algorithm is broken...

I hope you can release the new update soon where every block is rewarded with some CGA and not the faulty 0-or-1 CGA.
hero member
Activity: 595
Merit: 500
Quote
I'll do my best here to explain it from a technical point of view.
The block reward can either be 0 (zero) or 1 (one) CGA coin.

    1. The block reward is determined by first checking if the difficulty is above or below 3 (three). If it's BELOW 3, then go to the next step. If it's ABOVE 3, go to step 3.
    2. Divide the current block number by 3 (three), and check the remainder.
    2a.  If the remainder is equal to 0 (zero), then the block rewards one CGA coin. Done.
    2b.  If the remainder is not equal to 0 (zero), then the block does not reward any CGA. Done.
    3. Divide the current block number by the current difficulty, and check the remainder.
    3a.  If the remainder is LESS THAN 1 (one), then the block rewards one CGA coin. Done.
    3b.  If the remainder is NOT LESS THAN 1 (one), then the block does not reward any CGA. Done.

We know that the difficulty increases as the network's hashrate increases, so more mining will make this coin more elusive, but not impossible to find.
s4w3d0ff and others, please correct me if I am wrong. I hope this helps you understand the coin better!
can you explain step 3 more on the examples:
current block is 57452 and diff is 4.26714924
current block is 57453 and diff is 5.94869535

57452 % 4.26714924 = 3.36978188 > 1 = No anomaly
57453 % 5.94869535 = 0.50030969999 < 1 = Anomaly
Sawedoff,
I hope that I am wrong, but I think we have a problem in the current code....
For days I have been mining and noticed that all pools are paying *much* less than what I would expect from calculation,
so I decided to look into it more. First I checked how many blocks were found by pools to see if that would show that the
coin works as expected and what I saw is that the nr of found coins is *much* lower than can be expected from the
ratio between pool and network hash rates. I did take into account that the pool does *not* show blocks that have no reward,
the only blocks that you will see on a pool are the 1 CGA reward blocks and the blocks with 0 CGA but carrying TX fees (generally 0.001 CGA). What I would expect if the Diff is lower than 3, a pool should find 1/3 of the expected blocks with 1 CGA reward and an unknown nr with only TX fees but essentially the other 2/3 have 0 reward. If the diff is >3 then the pool should show on average a lower nr of blocks with 1 CGA reward and higher nr with no reward.
What I found was that even on the biggest pool (anomalypool) which has 2/3 of network hashing power, less than 10% of the blocks have 1 CGA reward, even though the diff is around 3 and you'd expect that more than 30% of the blocks carry a 1 CGA reward.
I verified with another smaller pool and a very honest pool operator, who did not know nor could explain why it regularly took more than 1000% of expected shares to find a block (with reward, since zero-reward blocks are not shown) and generally the stats were about 3 times lower in payout and finding then I could calculate.
Then I decided to take the test, grab a random block without reward (which can show up on the pool due to carrying 0.001 CGA TX fee)
and do the calculations for myself by hand. Take block with height 58,041 and verify the diff: 3.89795634
Now 58,041 % 3.89795634 = 0.43 !!!!
As far as I can see, this zero-reward block *should* have been rewarded with 1 CGA.
Can you verify the code that you are putting in the new release, because I am afraid that there is a bug in the CGA code so it does not pay out as expected and since network hash rate has gone up, we are seeing even fewer payouts, anomalies are not as much as calculated. My estimations are that we are seeing only half to as low as a third of the expected CGA rewards. That makes mining CGA unattractive...
hero member
Activity: 595
Merit: 500
I've been meaning to finalize and announce this for a long time, but so far the p2pool CGA source is still experimental to me (mostly because of the lack of a proper SUBSIDY_FUNC). As such I put it in the experimental branch on https://github.com/0x0aNL/p2pool-0x0a/tree/experimental which you can clone with
Code:
git clone https://github.com/0x0aNL/p2pool-0x0a -b experimental p2pool-cga
Greetings from another Dutchman!
Is your pool SW different than tf2honeybadger's P2P pool at inceptioncraft.net?
Or are you using the same?
I have mined on that P2P pool a short while, but my miner did not reach the usual hash rate
due to the restarts for every new block detected, probably this is something unique for CGA P2P pool
since I rarely see a work restart on other pools.
Pages:
Jump to: