This Difficulty Retarget Problem
-
Thanks Christian :)
[quote name=“zerodrama” post=“1822” timestamp=“1368109397”]
I think it should be at least 4 difficulty drops before we decide to reset the retarget rate.The community interaction powers the coin. Look at BTC-e. Hacks and DDoS have not dented their reputation the slightest. They respond to mistakes (sending litecoins to a bitcoin address - did that a couple times) very quickly.
Take this drop as a gut tester.
Faster retargets will kill mining. And then prices will rise too fast creating bubblecoin.
We should just chill and build and then come back with a vengeance in this time frame.
Don’t forget that there’s a cult amongst cryptocoin users who hate anything other than 1Tr3uC01n70b1nd7h3m. They see you forking early and they will attack like wolves smelling fear.
[/quote]I’m not to worried about the cultists, if I was I would never have started Feathercoin in the first place.
I am concerned on your point about faster targets killing mining. I would like to think that a difficulty change targeted for once a day might be welcomed over having to wait the current four days. I did not pay enough attention to some of the other coins which had fast difficulty adjust like Terracoin and the sorts of problems that created, this needs to be looked at before we can do anything.
Can you expand on your point as I’m want to do some further research into this?
-
[quote name=“ChristianRiesen” post=“1837” timestamp=“1368110844”]
@Bushstar: I’d advise you to do something that I have counseled many very successful people. Well successful, after I advised them. Listen to what the users are saying, but react to what they are doing. Look at hash rates, transaction volumes etc, the real stuff, things that actually matter, not what a hand full of people on this forum say, me included. We all have different opinions about this and other crypto currencies, but if you act according to what users are doing, it will always be successful.
[/quote]+1 for that.
-
Hi All - new poster here…
So I’ve been reading with interest the different opinions on the difficulty re-targets and from a miners perspective, all i do is check the most profitable coin to mine and then point my 15Mh/s to the relevant pools. At present it’s LTC. Reasoning is obvious, i have a lot of money tied up in mining rigs and a large electricity bill to cover. Whilst it’s easy to say that you want the miners to stick around for the rough and the smooth, it’ll never happen. We’re a ruthless lot that are in it for the money. Simples.
In a couple of weeks the FTC diff will drop massively, all the mega-miners (read: BigRigs) will jump back on it pushing out the smaller miners, flood the market with mined coins which will further devalue it and it will spike again. You’ll end up with a massive difficulty again but coins worth even less than they are now. Meanwhile LTC diff will drop again and the profitability will rise and we all go back to LTC again. rinse and repeat.
In the meantime FTC is getting quite a bad rep on the various forums and trollboxes and i’m concerned that the damage being done may be irreparable. We need to act fast.
I actually think there is room for another Alt-coin such as feathercoin (shame no-one though to register the other TLDs though!) and think it can work but we need to somehow stop the bounce.
Is it even possible to make the re-targets live? I.e. tied to the current hash rate so there is no spike or set it to every 50 to 100 blocks or something so that it’s much more cumbersome to move your miners around? just needs to be more responsive. Or slow down the adjust so it’s only half or even a quarter of the full retarget?
Just my 2ftc…
-
We are not Bitcoin when GPUs were just being repurposed for mining. We have miners jumping chains. These are not people figuring out the settings or adding new hardware for mining. They already do it. The result is that they can come and go fairly quickly.
So you have two tiers of GPU miners: prowlers jumping chains at the difficulty thresholds and community builders buying new hardware. Difficulty has to stay at a given amount because the people who bought their hardware need to make a profit on their investment. The prowlers / flash miners are most likely going to dump anyway. If difficulty changes faster, the price will have to go up in order for people to make profit. This will further cause shocks in the market.
Profit = Mining output * Price * Price drop percentage due to profit taking - Cost of hardware.
To compensate for a faster retarget, gpu miners will try renting out their hardware so they can keep buying more to keep up. They will fall behind due to scams, difficulty jumps, poor configuration, hardware failures. They will then buy up coins to meet their debt to customers. This will drive the price up for a while. But instead of being able to create long term business and community, the prices will keep rising just to have the miners break even. When this is no longer sustainable, we’ll hit another difficulty stall like we have now. And that one will be followed by a price crash.
But even without these issues, prowlers can engage in the same High Frequency Trading tactics as they do on Wall Street. The average creative person who doesn’t give a hoot about overclocking will be fumbling with settings while these big shots come in, snipe the currency, and run off. And they’ll keep flushing the currency as much as they can get away with it.
There is a solution, but I’d like to talk to you at length about it.
-
Maybe something to look into, if and when the difficulty issue becomes a large enough problem, is the PPC difficulty (which was integrated into scrypt with Novacoin).
The PPC difficulty alogo appears strongly resilient to large hash swings as it is able to adjust difficulty more frequently but take in consideration an average hash rate over a set number of blocks.
TRC was basically saved from complete destruction by a “difficultly” attack by partial implementation PPC difficultly alogo.
I suggest this only as something to look into in case a hard fork is ever needed to save the coin, not as something that needs (or might ever be needed) at the moment.
Also, I know the “51%” attack is a common concern, but many times its threat is overblown. There have been multiple times a pool has had up to 61% of the BTC hash rate, and nothing happens unless said person(pool) has malicious intent. There are two phases to a “51%” and the hash rate, generating a chain in private and merging that chain are just the first part. The second part of a real “51%” attack involves creating trades or spending the coin to get goods or service and then reversing the transaction in effect stealing the goods or service traded for.
Though executing the first part of the attack and going public with it (not really a true “51%” attack) can destroy customer confidence, coins can recover and have (Example: BBQ coin)All new alt-coins communities need to be more concerned about difficultly attacks, TRC took a devastating blow and has not fully recovered, though it is recovering slowly.
-
I agree a fork should be a last resort, but would a planned change to the algorithm @ a set block# require a fork? (I really don’t know the answer to this btw?)
How I see things now, and what I predict will happen (I really hope I am proved wrong though) but when the difficulty finally drops and the loyal have ploughed through it, the big-rig miners (who have since departed after dumping their quickly made coin) will return to make more quick coin and dump once again, this will drive the price lower and lower on each return visit, and we could be stuck in a cycle of 1 step forward and 2 steps back. :-\
Or am I missing something?
-
@Bushstar. I have a short term solution.
First things first. PPC and Novacoin are deliriously wrong in varying the reward amount. I have tried to profit in the past on Terracoin and Novacoin and the result was:
PPLNS ramp too slow to deliver coins, reward changes too fast to get profits that are projected on say dustcoin.com.What we can do instead of shortening the retarget, which as with CHNcoin (60 seconds wtf?), will only make the bounce rate faster and drive away long term creative people, is actually EXTEND the retarget time when it comes back down and SHORTEN it when it goes back up.
-
Here is my opinion
In fact,the main problems is less miners cause more block time and it also case the lose of miners and more block time ,and also less coin made.
I just wish if what ever the miners have,what ever the hash rates have,the coin madespeed is in the same rate?
So,diff change every time or every number of block is one of way to keep the coin made speed in the same rates,which cause this problem. there must other way to let the coin made speed at the same rate,which will keep the coin value.
Any coin follow btc in the diff change dynamic will follow btc and will also face this problems, more or less.
So it is time for us to creating a new world,keep the same principle that the total coin is fixed and the coin made speed is nearly fixed or in a fixed rate.
I will continue thinking on it.I am not understand why NFC changes its block value with its difficulty in the same direction.
I mean,when diff up,block value should be more.I do not understand why it is going down.in my opinion,change block value can be done in the same time when the difficulty changes as usual.
Block value should change dynamic to ensure the fix coin made speed.only in the situation hash rate did not match the diff.
At the normal time,the difficulty should follow the hash rate so the block value is normal.
When the difficulty is lower for the hash rate,usually the block time is less than estimate, so I think the block values should all changes to ensure the coin made rate is fixed.
Forexample if 1000hash rate equal s 100 diff,when the diff is 100,if hash rate is 1000,every block value should be 200 as usual. if the hash rate up to 2000 before diff changes to 200(100 still),each block value should be 100(for example,may be other number calculated).
When the diff is still 200, but the hash rates decreased to 50( similar like current situation), the block value should be 400(or other no),to ensure the miners earn as usual so the miners won’t run away.This is my personal opinion,consider and talk about it.
-
Coin value comes from community trade.
-
[quote name=“zerodrama” post=“1900” timestamp=“1368117906”]
Coin value comes from community trade.
[/quote]
You are right,but I mean coin made value.For example, in diff 188, if I only got 100 coin per day,with hash rate decrease(I mean network hash rate decrease,not mine.just suppose I have the same hash rate all the time),I may only get 20 coin per day. how can I still get 100 per day in now situation is what I try to solve.
With what I say,there is may cause a little effect to pool for pplns gain.but it dissent matter,there is no effect for the stable miners.
-
1. If you are going to fork FTC, I think the sooner it’s done, the better. I mean the further the more there will be Feathercoiners, the more would need to know about the update, and the more of them would miss it. But I think retarget time should be set in some smarter way (some ideas already are on this topic). Because if you just divide the current block number by 4 that doesnt change much. As FTC and all cryptos get more popular difficulty will get much higher. And those 504 blocks can still be damn hard to mine if miners would continue to hop on and off FTC.
2. Maybe this retarget problem isn’t so bad as i thought at first. LTC will retarget after 1,5 day, its difficulty will rise. Other coins also will adjust their difficulty. Then some miners could get back to FTC. Maybe we won’t need to wait the currently stated 18 days for FTC retarget. ;)
-
leotian, that is an interesting idea that I believe would be better suited for a new coin. It would be a lot to implement and test which would be best tried out in a coin dedicated to it. One of the first issues that springs to mined is the reliability of the network time with is a median of peer nodes.
[quote name=“zerodrama” post=“1866” timestamp=“1368114457”]
What we can do instead of shortening the retarget, which as with CHNcoin (60 seconds wtf?), will only make the bounce rate faster and drive away long term creative people, is actually EXTEND the retarget time when it comes back down and SHORTEN it when it goes back up.
[/quote]Do you mean overall number of blocks till the difficulty retarget or target block time?
If it is the block target time then it is already the difficulty that makes it easier and harder to mine. You can adjust the time taken to mine blocks by adjusting the difficulty alone. If it is blocks till the next retarget then what happens if the difficulty always goes up like Bitcoin?
[quote author=d2 link=topic=305.msg1863#msg1863 date=1368113790]
I do like the 504 retarget idea since everything else is 4x as fast but also would recommend an adjustment to the diff change amount during re-target proposed by nutnut:[quote author=Nutnut link=topic=305.msg1851#msg1851 date=1368112073]
Or slow down the adjust so it’s only half or even a quarter of the full retarget?
[/quote]Re-targeting to hit the exact 2.5 minute block time is what causes such massive spikes and drops in the first place. If you were to speed up the re-target to every 504 blocks but only allow 10-20% dif +/- it would be a relatively quick re-target and moving in small increments would get you to your goal block time in a reasonable time-frame.
[/quote]This idea seems very sensible to me as a four fold change, four times more frequently would cause trouble and potentially be abused by large mining groups. A block adjust of 504 blocks but with quarter of the adjustment makes difficulty changes more fine grained but with less drastic swings. It will still retain the same overall potential but be more dynamic and I believe safe.
-
Something I never thought I’d see: Royal Coin has 480 block retargets (vs 504) and 3 minutes confimations (vs 2.5).
It is producing blocks and getting hit by orphans faster than anyone can download the constantly extending chain.
https://bitcointalk.org/index.php?topic=199839.0
Do not lower the block retarget. The controlled difficulty increase +/- 10 sounds like a plan. I have a few other ideas.
-
[quote name=“zerodrama” post=“1678” timestamp=“1368069897”]
Quick retargeting makes it pointless to invest in new mining hardware in the first place. The price isn’t going to react fast enough.
[/quote]Not exactly. TRC had a problem with difficulty rising and falling too fast since retargeting occurs every 30 blocks, so they introduced averaging window of 720 blocks. Now there are several difficulty jumps a day which is quite acceptable. It makes TRC somewhat more ASIC resistant than other sha-256 coins. There may be other ways though.
The network lost about 90% of hashpower by this moment which results in 19 days to retarget at deliriously low difficulty of 30. I guess the situation will get even worse over time, so it’s better to figure out something until the network explodes with hashpower again.
-
[quote name=“zerodrama” post=“1971” timestamp=“1368136002”]
Something I never thought I’d see: Royal Coin has 480 block retargets (vs 504) and 3 minutes confimations (vs 2.5).It is producing blocks and getting hit by orphans faster than anyone can download the constantly extending chain.
https://bitcointalk.org/index.php?topic=199839.0
Do not lower the block retarget. The controlled difficulty increase +/- 10 sounds like a plan. I have a few other ideas.
[/quote]
ohrly?Yes, it will be a much worse scenario for FTC when the difficulty becomes 20 in the next few weeks, I’m pretty sure we will see 5 Ghash/s on the network within 24 hours of the difficulty change, just for them to mine for a day and then hop back off to Litecoin/Novacoin/etc.
If the retarget is lower, then many of those miners will constantly be “missing the next bus” smoothing out day-to-day hashrates.
I’ll admit- I’m mining RoyalCoin at the moment…but not for long. I own over 1100 FTC and plan on converting the rest of my LTC to FTC… I’ll be back on this network in the next 2 weeks to help out with network hashrate regardless of profitability.
-
There’s miners on litecoin who get 56Mhs or more. They won’t miss a thing. The only people missing anything will be the peasants.
-
The solution that the Devs implement MUST hinder hop-on hop-off strategies.
Or else the coin will become a trampolin.
I think that maybeone or more of the following could solve this:
1)Increasing/Decreasing difficulty x% when hashrate drops y% (Retarget works as usual 2016 blocks)
2)Automatically increasing reward for higher difficulty / decreasing reward for lower difficulty -
[quote name=“sumantso” post=“2022” timestamp=“1368168640”]
Its a shame that all those who stick through the high difficulty phase can’t reap the reward when the difficulty falls.If we are changing the difficulty scheme with a hard fork, I would say do it quickly. The reason being that it will save the diffculty crashing after the next retarget.
PS: Bushstar, it was great to see you pointing out the good support that FeatherCoin has on the Trollbox. I am trying my part in keeping the spirit up but you can’t save idiots from selling at 0.0017.
[/quote]TBH the way the coin is at the minute, I can understand people selling all the way down to the very bottom, as they will more than likely be able to buy back in much lower when the next diff-change+dump occurs…Like you say the ones who stick through the high difficulty will be the ones who won’t really reap a reward, because as soon as the difficulty drops and the hashrate increases, those big rigs will be able to mine in 2-3 hours what has taken most of us 2-3+ weeks to acquire. :-\
-
I think if we are going to fork, it needs to be done now.
The hashrate has dropped to practically nil which exposes us to the 51% attack and the re-target time is just getting further away. The further away it gets the more people abandon it. Me may never hit the re-target. At present there are very few people on the network (as can be seen with the amount of time it takes to get confirmations) and the number is falling as it becomes less profitable to mine and/or trade.
The price is falling rapidly as people are loosing faith and there seems to be growing calls on the trollbox for the altcoins (ftc & cnc) to removed from BTC-e.
I say we take the opportunity to fork now whilst there are very few people mining and trading the coins from their wallets. It wont go down well but it will only get worse in time.
EDIT: Thinking about it, a fork would show that we are proactively supporting and maintaining ftc rather than leaving it to rot. People like things that are actively developed.
-
Is it possible to hire peoples to mining FTC for a week to short the re-target time?
