Topic: NobleCoin[NOBL] - 8% PoS | 1Yr+ | MARKETPLACE | PAY | GIFT | CHARITIES/MERCHANTS - page 598. (Read 1053158 times)

noblecoin.cryptotycoons.com is stable! with lower fees.


Also, people need to start trading at on this thread https://bitcointalksearch.org/topic/noblecoin-wtbwts-thread-405395
To make it on an exchange, we have to be able to show proof of volume and interest.

Go do a couple microtrades, 10 coins or whatever, anything helps. Right now I think there has only been 1 successful trade. ever.

I like this coin, I need a stable pool, but you stop registration. . .

It gives that pool the ability to bog down the block chain trades, and stop all trades from happening.

At even higher (like ~80%+) hashrate, it can prevent all other pools and solo miners from being able to find a block at all. They'll basically find everyblock, solve it, and move on making all other pools orphan there blocks.

What happens if a pool has more than 50% of the net hash rate?

Compliments for your attitude on this mate, but I don't see it is much of an issue. The coin is still very new, has low total volume still and is not traded anywhere worth mentioning. Not a lot worth manipulating it seems.

yes we are over 51% yet again..... it seems that workers on my pool has stayed stable over night (EST timezone), but now the network hash rate has decreased from 90 Mh/s to 61 Mh/s. The problem seems that the other pools are not retaining the workers they end up leaving my pool when I do a increase in fee to correct the 51% issue. I will do the increase of fee yet again, but other pools need to retain their workers.

A North American pool is live!
Come check it out!
NobleCoin.CryptoTycoons.com


only 1% fee!!!  
Never down
FAST!!

*0% fee special is over due to net hash dropping.


Free escrow service to all active miners!

Use this new Pool

http://noble.plingeling.se

Is valid-error over 50% again? it gives 45 Mhash of 80 net Mhash in the dashboard.

Or is the net difficutly not displayed correctly?

Just a heads up our site has suddenly gone down - we will upgrade hosts if this happens again as it seems to have affected a number of their sites and not just ours. *edit: Nevermind resolved after a couple of minutes.

Rofo, thanks for the honesty ... better have a dev saying he does not master everything ... those one are usually more open to discussion and improvement ^^

I verified the behavior by computing the difficulty by hand using the algorithm in GetNextWorkRequired (and using the block explorer at http://noblepool.zapto.org/block_crawler.php as a source of data)

I can reproduced exactly what is happening on the network (i.e. predicting the next difficulty using the time between the last two blocks), so I am now sure it does only use the time between these last two blocks.

When computing next difficulty, the "30" block value is used as a weighting, but the date of those blocks or the first of the 30 is not checked in any way.

Also the 30*60  value is used when verifying blocks (from the database or from the network, not sure yet).


The variations we can observe (by around a factor of 2) in the difficulty are mostly caused by the variance in the hashcash-like algorithm.
For a given difficulty, it takes *in average* 60 seconds to find a block.
Should one be found in 30, the next difficulty will be greater, should it be found in 90, it will be smaller in a proportional way.
I'm pretty sure this instability would persist at a much higher hashrate.


I don't know what would be best ... for example memorycoin now uses "Kimoto's Gravity Well" https://forum.megacoin.co.nz/index.php?topic=893.msg5742#msg5742 developed for megacoin. But I'm really not sure of the behavior of this algorithm as I did not looked at it in detail yet.

I don't have time today, and it's not really urgent anyway (the instability we experiment is not THAT high with the current parameters), but I will try to take a closer look at this part of noblecoin code this weekend.


As for what the pools say, it probably comes from what it was saying for the coin this pool has been initially developed for, or even from the official post announcing noblecoin ^^

Difficulty retarget seems to work correctly, and the fluctuations should ease once the network reaches high enough hashrate.

Hey Huu, I'm no guru when it comes to the source but have been progressively learning. It is my understanding that line 854 of the main.cpp: static const int64 nTargetTimespan = 30 * 60; specifies re-target time. 60 seconds by 30 being 30 minutes (30 blocks at 60 seconds per). We've seen a number of new coins with this variable seeming to be stable and having no problems.

I do see what you're seeing in that the wallet re-calcuates each block as you do see a new difficulty with each block found. However on pools it specifically states and you can see that it attempts to re-calculate at estimated 30 block intervals. So we're also learning as you are how some of these aspects specifically work. I don't personally enjoy admitting some technical aspects are above me as we'd like to be all over it and understand every detail but there you go. I think NorrisK may be on to something.

Regarding the fluctuations - we had some big miners with 50Mh/s+ jump on and off over the last couple of days between us and others (perhaps Coinye) - so that was sending us from 1 to 3 difficulty a number of times.

Maybe it takes the avarage of the last 30 blocks, but does not reset the counter after changing the difficulty? So for 30 blocks it is ok, but then it changes at 31 (using block 1-31), 32 (2-32), 33 (3-33), etc.

I'm always interested in the technical aspects ...

The official post says there is a 30 block retarget time ("- Difficulty re-targets approximately every 30 blocks.")

From our miners, it clearly appears the difficulty is recomputed at each block.
I did not looked much in the code yet to find what algorithm is used, but maybe an explanation here from the dev could help

Continuous difficulty adjustment has pro and con.

• + It prevents the chain to get stuck for a very long time if the hashrate suddently drops. Think about big switching mining pools (middlecoin & co) that can up the difficulty of a not well designed coin to oblivion before letting it rot. The coin *MUST* be prepared to handle these pools
• + It prevents insta mining that concentrate and devaluate the coin. Long retarget times as in bitcoin/litecoin can't be used for newer coins
• - It can cause difficulty instability. Yesterday I observed the difficulty vary between 1 and 2 over very short periods (~10 to 15 blocks maybe). Variations of this amplitude are not a problem, at least at the current state / total hashrate of the coin. But it's hard to predict how it will behave later. Memorycoin2 had very big problems at launch with their difficulty adjustment, lets not reiterate ^^. At long term, strong variations are also a security / sanity concern ... we could imagine miners connected to different instable coins and mining only the easiest blocks of each one
• - If the continuous  adjustment is too slow, it can also cause problems as in yacoin where it allowed insta-mining at the beginning and subsequent dumping.
  

I'm not sure what is currently used, it seems to me that only the time between the last two found blocks is used to compute the next difficulty. If true, this is the root of the instability, more real blocks should be considered (or the capping factor to the diff. adjustment should be reduced.
(the 30 block value seems to appear only as a weighting, preventing extreme change from block to block)

Where is this part of the code coming from ? Another coin that may be facing this problem ? or is it specific to noble ?

I have to manually change your mining fees. I just updated everyone who has any amount of hashrate! If I missed you let me know, PM me your username and I'll update you immediately!
Sorry for the delay.

Yes transaction section is charging fees still, .1 NOBL to be exact per transaction. This is because the coin has a minimum transaction amount to keep from bogging down the block chain. If there were not a minimum, someone could hypothetically trade a tiny amount of nobl over and over, increasing the size of the block chain until its a few hundred gigs large, and kill the coin.

I'm not positive what the minimum is for this coin, but given the ease of mining, and the high amount of coin, .1 seems about right compared to most other coins.
If its not the minimum i'll reduce it, just pm me.

And they have a real project behind their coins 

Looking good for the long term! Good work dev! 

loving the low difficulty and the plans for this coin. slow and steady wins the race.

Feeling proud, just checked my miner and in the last ~30-35hrs (cant remember exactly when I restarted my miner) I've solved 31 blocks!!!