Diff trap... Again.
-
Nutnut is correct I believe. With more frequent difficulty adjusts at less variance it should encourage more loyal miners and would actually deter multipools if the time between adjusts was small enough. I am used to working through high difficulty with Litecoin but that was before we had so many migrant miners. These miners means that we move by the maximum each time and 41% feels to great a number for that.
Ideally we want a faster adjust like Ghostlander suggests. However, with very small difficulty adjusts people seem to find ways to exploit or abuse it which is my concern. We do not want to end up like Terracoin having to issue several hard forks in quick succession. Hard forks are something that should be done as infrequently as possible. We have only ever gone through this process once at block 33,000 when we got hit by the 400% difficulty adjust that we use to have.
If we go for something somewhat conventional like half or a quarter or even an eighth of what we have right now we could perhaps bundle this with the ACP release. Any grand changes to the way that difficulty is calculated will have to be vigorously tested.
-
[quote name=“Bushstar” post=“25730” timestamp=“1376665564”]
Ideally we want a faster adjust like Ghostlander suggests. However, with very small difficulty adjusts people seem to find ways to exploit or abuse it which is my concern. We do not want to end up like Terracoin having to issue several hard forks in quick succession. Hard forks are something that should be done as infrequently as possible. We have only ever gone through this process once at block 33,000 when we got hit by the 400% difficulty adjust that we use to have.
[/quote]Terracoin was exploited through a time travel bug. Their 1 hour retarget window (30 blocks at 2 minutes each by default) also was a breach for attackers. Sunny King has had these fixed in April, but TRC can be abused anyway in these ASIC days. It cannot compete with Bitcoin for resources. Neither PPC can. Our situation is different, though we can learn from their mistakes, too.
-
The OT of this thread is “Diff trap … again”.
I still believe software solutions can’t fix this “problem”. It’s educational work, promotions, PR that convert junk-coin miners to FTC miners.
Lets take a look a coinchoose right now:
[b]Number 1 is Alphacoin[/b]: 8 MH/s network speed and an exchange rate of 0.00000733:
When I look at the Ask side on Cryptsy there are not a lot of people willing to buy Alphacoin.
A few oders at 0.00000733, some at 0.00000732, etc.
The Ask side from 0.00000733 all the way down to 0.00000500 is maybe 1.5BTC in total!In other words: I have no chance to make the 750% profitability that coinchoose promises because I can’t transfer my Alphacoin to BTC. Or fiat.
[b]Number 2 is Nibble[/b]: 1.49 MH/s network speed and an exchange rate of 0.00006558:
1.49 MH/s network speed? Seriously? Even my Office PC can start a 51% attack on this coin. Why would I risk mining a coin that could be worthless in the next 5 minutes?… and so on and so on …
The first coin to take serious comes at [b]number 15 (Novacoin)[/b]. Exchange rate 0.04729 and 2220 MH/s network speed.
[b]Summarized: We have to educate the alt coin miners[/b]. In the end it’s about stability. Same as in stock markets: You can have a lucky shot with some high tech start up. Followed by 5 unlucky shots. The way to “get rich” is with stability (equity funds for example). A stable growth of 5, 6, 7% or more per year.
It’s this “I can beat the market” that makes people (the fast majority!, not Goldman Sachs & other big players) actually lose money in the stock markets. And, imho, the same is true for alt coin mining: This constant jumping around makes you lose money in the end - unless maybe you have a very good strategy and strictly and disciplined follow it. Because 90% of the alt coins are in a downward spiral (difficulty and exchange rate wise) you have to dump them as fast as you can.
Feathercoin is different! Difficulty constantly around 140 - 200. Stable exchange rate.
[b]The facts are there people! We just have to show them to the alt coin miners.[/b]
-
Agreed, chasing the most profitable coin does not work as there is often not enough volume to support those prices. If people focused on the long term then I like to think that they would mine Feathercoin. However if they have kit to pay for or mining no longer covers the costs, then they may need to shift their mining elsewhere to a coin that does have the volume to support the price. Right now this will be when the difficulty is high.
I believe that we should push for greater adoption and talk to people about the merits of Feathercoin, but also, if we can make Feathercoin a better coin for miners then we should pursue that too. Making the difficulty changes more frequent to a lesser variance will make the difficulty swings less violent and fairer to loyal miners.
-
[quote name=“Radacoin” post=“25770” timestamp=“1376738900”]
The OT of this thread is “Diff trap … again”.I still believe software solutions can’t fix this “problem”. It’s educational work, promotions, PR that convert junk-coin miners to FTC miners.
Lets take a look a coinchoose right now:
[b]Number 1 is Alphacoin[/b]: 8 MH/s network speed and an exchange rate of 0.00000733:
When I look at the Ask side on Cryptsy there are not a lot of people willing to buy Alphacoin.
A few oders at 0.00000733, some at 0.00000732, etc.
The Ask side from 0.00000733 all the way down to 0.00000500 is maybe 1.5BTC in total!In other words: I have no chance to make the 750% profitability that coinchoose promises because I can’t transfer my Alphacoin to BTC. Or fiat.
[b]Number 2 is Nibble[/b]: 1.49 MH/s network speed and an exchange rate of 0.00006558:
1.49 MH/s network speed? Seriously? Even my Office PC can start a 51% attack on this coin. Why would I risk mining a coin that could be worthless in the next 5 minutes?… and so on and so on …
The first coin to take serious comes at [b]number 15 (Novacoin)[/b]. Exchange rate 0.04729 and 2220 MH/s network speed.
[b]Summarized: We have to educate the alt coin miners[/b]. In the end it’s about stability. Same as in stock markets: You can have a lucky shot with some high tech start up. Followed by 5 unlucky shots. The way to “get rich” is with stability (equity funds for example). A stable growth of 5, 6, 7% or more per year.
It’s this “I can beat the market” that makes people (the fast majority!, not Goldman Sachs & other big players) actually lose money in the stock markets. And, imho, the same is true for alt coin mining: This constant jumping around makes you lose money in the end - unless maybe you have a very good strategy and strictly and disciplined follow it. Because 90% of the alt coins are in a downward spiral (difficulty and exchange rate wise) you have to dump them as fast as you can.
Feathercoin is different! Difficulty constantly around 140 - 200. Stable exchange rate.
[b]The facts are there people! We just have to show them to the alt coin miners.[/b]
[/quote]Do you want to work with me getting this messaging out there?
How hard would it be to do our own coinwarz? But with our own algo?
-
[quote name=“Nutnut” post=“25716” timestamp=“1376653584”]
[quote author=chrisj link=topic=3327.msg25715#msg25715 date=1376650620]
[quote author=Nutnut link=topic=3327.msg25706#msg25706 date=1376640837]
Even I am now getting fed up of mining the least profitable coin so others can drop by for a quick win and I’m considering taking on the attitude of if you can’t beat em, join em.Smaller more frequent changes mean would only fluctuate by a couple of positions on coinchoose rather than 10 or 15 and therefor encourage more miners to stick with us.
MOST MINERS ARE NOT LOYAL. They want to make money, as do most people involved in crypto.
[/quote]Personally I am very grateful for your loyal support in mining during low difficulty because it helped me to work on the SMS Address rollout and the Homeless project I am working on.
Please don’t “join 'em” Nutnut, don’t go to the darkside.
Is there anything you need asside from more stability in the coins that we might be able to help you with?
For example I can join your pool. I can buy you a pint in Oxford on the 7th with some FTC :)
Anything else?
[/quote]Haha, dude, i think you misunderstand me. It’s not a pint or mining on my pool that will make me stick around it’s smoothing out the hashrate/difficulty spike so i don’t have to mine it back down for the arseholes in the troll box to decsend on us for 8 hours only to ramp it back up again.
I personally think (as do a LOT of miners - particularly the hardcore) that smaller more frequent diff changes WILL make miners stick around as it will save them mining at a high diff for days only to get a brief 8 hours of low diff. I feel like the bullied kid at school who has to do everyones homework or get his flushed down the crapper.
My point is, as the Director of Mining i listen to the miners, speak for the miners and mine myself so i know how we think and work. Miners NEED the diff swing to be reduced. We don’t mind if the diff naturally increases due to prices/economy cos that means we’re still covering the leccy bills and we’re getting where we want to be as THE alt currency. When it’s up and down there’s simply very little in it for us during the tough times.
I’m not going anywhere just yet but i need to be heard and i speak for the loyal miners.
Bushstar once called me a professional miner because I have HUGE outlays and HUGE leccy bills to cover and therefor maybe i [b]should[/b] move around or stick to litecoin (this was back in may)… maybe i am (and not to belittle the smaller miners) but it’s the pro-miners that bring the big guns to the party so we need to look after them. A 40% swing in diff/coin return is HUGE to a “pro miner” and could be worth 150+ coins per day, i think you see why they get itchy feet?
The more of them we have full time, the more secure we are.
[/quote]+1
Im getting tired of mining the diff down on ftc. I have no issue with the diff increasing. that just means fewer ftc pr day and most likely higher value because of that.
But this constant diff rollercoaster is driving me nuts (get it ;D)
I would be much better off financially if i just directed my 13.8 (soon 16)Mhash towards the multipools and just bought ftc with those earnings…However, im a firm supporter of ftc, and will continue to support both the network and the forum.
I need however to know that this is taken seriously, and there will be an end to this stupid diff bungy ::) -
[quote name=“svennand” post=“25786” timestamp=“1376764367”]
I need however to know that this is taken seriously, and there will be an end to this stupid diff bungy ::)
[/quote]I am a miner myself and I hear what you are saying.
But take a look at Bitcoin. They have a much bigger difficulty adjustment time. Yet no difficulty swings. The only way is up.
This is where we have to go!
Prio 1: Establish Feathercoin as a currency, create and support real world use cases
Prio 2: Make Feathercoin the No. 1 Scrypt coin (and in the future overtake Bitcoin)
Prio 3: Fine tune Feathercoins properties (like the difficulty adjustment algorithm)
…[quote]
I would be much better off financially if i just directed my 13.8 (soon 16)Mhash towards the multipools and just bought ftc with those earnings…
[/quote]This also supports our cause, because it would stabilize/support the exchange rate.
-
[quote name=“Radacoin” post=“25790” timestamp=“1376766679”]
But take a look at Bitcoin. They have a much bigger difficulty adjustment time. Yet no difficulty swings. The only way is up.
[/quote]I do not agree, when bitcoin started there where no other alternative, they didnt have to worry about coin hopping pools/miners.
And now they have so much hashpower on network that if a pool desides to go elsewhere it would even create a dent in the diff.
Feathercoin is “fighting” against alot of other alt coins for hashrate attention.
41% diff adjust is alot! especially given the “founding network hashrate”, by that i mean the faithful miners, and do not include the multipools/coin jumpers.
This leads to a massive 41% increase within hours of diff dropping, then it takes 1-2days to get the diff down before it jumps way to high again.The networks on the worst days go all the way down to 1000Mhash. when the diff drops we´re seeing hasrate up to 10000Mhash…
thats 9000mhash in pure “get in get out” mining.[quote]
This also supports our cause, because it would stabilize/support the exchange rate.
[/quote]My point is that if its more lucrative to mine a coin until its to difficult, then mine some other coin to trade for ftc.
I will say the system is broken, and should be dealt with.
Now the only people getting punished are the ones that stick around…it should be like that.The exchange rate will represent the confidence people have in the coin,
good confidence == small sell order && High buy order.
Low confidence == high sell order && Low buy orderIn my opinion the best way to get confidence for the coin is to constantly show that we care and evolve it for the greater good of the people
holding it, and the people mining it. If that means a smaller diff adjust in shorter time then its the thing we should do.When we get a “stable” coin with great security/reliability its much easier to sell/expand the concept to merchant/3rd party dev and so on.
Just my 50cent :D
-
[quote name=“Nutnut” post=“25716” timestamp=“1376653584”]
Haha, dude, i think you misunderstand me. It’s not a pint or mining on my pool that will make me stick around it’s smoothing out the hashrate/difficulty spike so i don’t have to mine it back down for the arseholes in the troll box to decsend on us for 8 hours only to ramp it back up again.
[/quote]Actually it’s playing out as I predicted before the faster change.
It used to 6 weeks stuck high maybe a week low. Now it’s a few days high, then 8 hours low. If we make it faster, it will be 2 hours high, 10 minutes low.
Granted that 40% change limit is a life saver. But this is a slippery slope. I’m going to make a slight adjustment to my design plans.
I overshot trying to be nice to everybody. Screw that. Time to get real.
-
[quote name=“zerodrama” post=“25800” timestamp=“1376781331”]
It used to 6 weeks stuck high maybe a week low. Now it’s a few days high, then 8 hours low. If we make it faster, it will be 2 hours high, 10 minutes low.
[/quote]There will be high or low no more if done properly.
-
[quote name=“ghostlander” post=“25802” timestamp=“1376788845”]
[quote author=zerodrama link=topic=3327.msg25800#msg25800 date=1376781331]
It used to 6 weeks stuck high maybe a week low. Now it’s a few days high, then 8 hours low. If we make it faster, it will be 2 hours high, 10 minutes low.
[/quote]There will be high or low no more if done properly.
[/quote]There won’t be any incentive to invest in hardware.
-
[quote name=“zerodrama” post=“25804” timestamp=“1376789568”]
[quote author=ghostlander link=topic=3327.msg25802#msg25802 date=1376788845]
[quote author=zerodrama link=topic=3327.msg25800#msg25800 date=1376781331]
It used to 6 weeks stuck high maybe a week low. Now it’s a few days high, then 8 hours low. If we make it faster, it will be 2 hours high, 10 minutes low.
[/quote]There will be high or low no more if done properly.
[/quote]There won’t be any incentive to invest in hardware.
[/quote]Nope, in fact since calculating your return should be a steady curve not a sine wave. I usually prefer to get a steady payout to pay my cost then a fluctuation in payout.
at some point 2-3 weeks ago we got a soft spot where we got 2-3 diff change of nearly 0 and had a hashrate of 2.8gh/s and diff around 144. then price spike and goes again with the 41% limiter.
I see the 41% like a rev limiter for a car you should have it to protect you not always hit it. the actual problem with coin hoppers is that the actual calculation make full gas or full brake and always overshoot. (so a possibility can bee to apply a damping factor of 0.5 to the retarget knowing that the retarget is not 1 for 1 as it is for Bitcoin and Litecoin. note Litecoin has 10% swing also now depending on other coins)
for the hoppers we need to understand that most don’t just get the most profitable, but use a select 4-5 and mine the one most profitable from that, many doing it by hand for the day or part of the day. ex. Look at the hash rate switch from LTC to FTC when FTC becomes more profitable then LTC ~1Gh/s switch over the 3-4 hours that follow the switch in profitability. so not been on first top (as we have not been in many of the recent high we got 4-5 on coinchoose and low in 15-20) a 7-8 spot should give us a likely 3-3.5Gh/s steady with some hoppers that will becomes more steady as we will most of the time meet their criteria. (steady price with a market and profitability adj for stale. some also look at 7 days profitability)
what is important i think is that we keep a range near 41.4% max adjustment for 504 blocks
so the 12 blocks looking back 504 blocks should make a small change like 1% each (42*1% composite would make like 47-48% max). this would keep time warp attack to at least actual efficiency and probably make it more efficient.a good timing of time warp and big retarget over 12 block even if looking at 504 back can make a diff increase that can be worst then what we have even with 30-40% of the hash rate. the probability of success are just proportional to your hash rate in the network and can be retry every 12 blocks. so you place block in future at some point getting diff down a bit artificially when you hit the retarget. so at some point after you make 504 block after block with time in the past as mush as you can. if you get both side on a retarget you can get ~3h lower time so mostly a max up retarget.
I would suggest 1% with a damping factor of .25 so been off by 4% would retarget by 1% been off by 1% would retarget by 0.25%. since we retarget over 504 blocks this would minimize the effect of luck/badluck of the timing for the 0&504 time difference to smooth it.
a more complexe formula that take 41.4% over 504 blocks and allocate for each 12 blocks a retarget greater then 1% can be done, and can have a better result. but this come with a risk of opening other attack vector for time warp
-
1 word (or is it 2) that sums up what no-one really took into account way back when…
MULTIPOOLS
They were our savior once now they could kill us. More multipools are popping up daily and it’s those that we need to manipulate to keep our hashrates more stable.
-
[quote name=“Nutnut” post=“25812” timestamp=“1376815361”]
1 word (or is it 2) that sums up what no-one really took into account way back when…MULTIPOOLS
They were our savior once now they could kill us. More multipools are popping up daily and it’s those that we need to [b]manipulate[/b] to keep our hashrates more stable.
[/quote]Manipulate?
I agree that Multipools might be the doom of all alt-coins, because they put them in a downward spiral (difficulty and exchange rate wise). But manipulate them? Really?
What we have to do is create a stable upward trend - so no one keeps looking at those multipools, choinchooses and coinwarzes any more. And only mines FTC.
-
OK, you go create that uptrend. You need stability to create an uptrend.
Meanwhile, i’ll hope bush adresses the re-targets that will smooth out the HUGE drops in profitability that causes coinchoose etc to lower our rank that causes the multipools to swing away.
When i say manipulate i mean keep out profitability stable so that the multipools keep mining us.
We’re going round in circles now. There’s clearly 2 different schools of thought so can we just agree to disagree.
-
[quote name=“zerodrama” post=“25800” timestamp=“1376781331”]
[quote author=Nutnut link=topic=3327.msg25716#msg25716 date=1376653584]
Haha, dude, i think you misunderstand me. It’s not a pint or mining on my pool that will make me stick around it’s smoothing out the hashrate/difficulty spike so i don’t have to mine it back down for the arseholes in the troll box to decsend on us for 8 hours only to ramp it back up again.
[/quote]Actually it’s playing out as I predicted before the faster change.
It used to 6 weeks stuck high maybe a week low. Now it’s a few days high, then 8 hours low. If we make it faster, it will be 2 hours high, 10 minutes low.
Granted that 40% change limit is a life saver. But this is a slippery slope. I’m going to make a slight adjustment to my design plans.
I overshot trying to be nice to everybody. Screw that. Time to get real.
[/quote]you do realise that if the 40% hadent been implemented it would be 8hour low then 6weeks high again? the hashrate is higher now. the 33000 coldfork was due to ONE large mining pump.
i do not concur in the statement 2hours high/10 min low.
with a faster diff adjustment fewer miners will keep switching between ftc and other coins
since the diff will change with smaller increament, and with shorter time periode. -
[quote name=“Nutnut” post=“25815” timestamp=“1376816461”]
OK, you go create that uptrend. You need stability to create an uptrend.Meanwhile, i’ll hope bush adresses the re-targets that will smooth out the HUGE drops in profitability that causes coinchoose etc to lower our rank that causes the multipools to swing away.
When i say manipulate i mean keep out profitability stable so that the multipools keep mining us.
We’re going round in circles now. There’s clearly 2 different schools of thought so can we just agree to disagree.
[/quote]Radcoin and you are right where you agree that not all of this can be solved with tech, you have asked for stability and you are right that’s why on the PR team we need to get out there and get FTC in to more hands so that it’s controlled by a few. We are working hard to do this and if you have any ideas, or have heard some that you think could be getting more attention please tell me.
In the mean time thank-you so much for your patience, we couldn’t do this without you.
-
[quote name=“groll” post=“25806” timestamp=“1376799747”]
what is important i think is that we keep a range near 41.4% max adjustment for 504 blocks
so the 12 blocks looking back 504 blocks should make a small change like 1% each (42*1% composite would make like 47-48% max). this would keep time warp attack to at least actual efficiency and probably make it more efficient.a good timing of time warp and big retarget over 12 block even if looking at 504 back can make a diff increase that can be worst then what we have even with 30-40% of the hash rate. the probability of success are just proportional to your hash rate in the network and can be retry every 12 blocks. so you place block in future at some point getting diff down a bit artificially when you hit the retarget. so at some point after you make 504 block after block with time in the past as mush as you can. if you get both side on a retarget you can get ~3h lower time so mostly a max up retarget.
[/quote]We can surely set a 1% maximum on difficulty retarget every 12 blocks to keep performance close to current 41.42% over 504 blocks. Additional restriction on future time stamps and increased nMedianTimeSpan reduce negative effects of time warp attacks. Advanced checkpointing addresses the problem of valid blocks being orphaned by attackers and decreases necessary number of confirmations for transactions to just a few. Is there anything else worth to be implemented?
-
As it seems appropriate to this thread (to me anyway), and because I don’t really feel like I got any good answer as to why it isn’t feasable, I will inject this comment again from one of the previous threads:
[quote name=“SixGun” post=“16421” timestamp=“1371649501”]
Why must DiffIncrease be equal to DiffDecrease? The issue I see is that we are having difficulty maintaining the 2.5 minute block target. What would be the negative potential if the 41% increase algorithm was separated from the algorithm that determines the diff decrease based on dropping hash power. The goal being to maintain the block time, not the profitability factor. Thinking long term, the survival of any currency depends on a consumer’s ability to use it. What I see hurting the FTC value is the long block times in period of reduced hash power.Wouldn’t this set FTC apart from other coins (not just alts)? Thinking of a currency from a payment processing perspective, digital currencies will [i]never[/i] be able to reach parity with actual cash {not thinking of bank to bank transfers} unless there was [b][i]instantaneous[/i][/b] confirmation. Part of the reason credit cards work so well for consumers is that they [i]perceive[/i] the confirmation to be instantaneous, just like if they handed over paper bills. Let’s not get into the semantics of credit as that is whole other discussion.
My point is that consumers/users of a medium of payment don’t think about the technical aspects. If we are trying to build a long term currency/economy we must remove the variability of payment processing in some way. Since transaction confirmation depends solely on the miner nodes on the network…it follows that the time period to confirm must be as fast as possible nearly irregardless of the technical backbone (miners) of the system.
The way I understand it the positive benefit of digital currency is to move [i]large[/i] amounts of currency quickly, with moderate anonymity. Cash is king because I can go down the street and hand over a bundle to my neighbor to buy his car. The legality, and proper reporting, of that transaction isn’t and shouldn’t be up to the producer of the currency.
I’m not speaking about the diff change and it’s impact on miners necessarily, rather I am thinking about it’s affect on the viability and perceived stability/staying power of currency (FTC). What I perceive as killing the network’s strength right now is that we had a huge jump in hash power which necessarily raised the difficulty (and by proxy the profitability and market value), however once the difficulty adjusted accordingly and miners jumped off the network strength suffered.
This may be a very naive point of view from an admitted n00b, but please help me understand why this wouldn’t be beneficial for long term success. We need to think ahead to the days when FTC is humming along at 5 TH/s and suddenly a major internet backbone goes dark and we lose a significant portion of the network nodes for a short period. The network needs to be able to maintain it’s confirmation times or [i]consumers[/i] lose faith.
Again, please be kind…
[/quote]A couple of notes:
-
as pointed out by zerodrama a little farther down on the OT, I should have said “The goal being to maintain the block time, [b][i]and subsequently[/i][/b] the profitability factor.” instead of “not”. I didn’t mean that they way I expressed it.
-
Also, after some more time mining {can’t mine currently, but hope to be back up this weekend} and more time reading these posts and learning from more knowledgeable here {I greatly respect both NutNut and Svennand}, I am definitely thinking about the difficulty and it’s affect on miners. I realize now that I was before, but didn’t really know it. :-\
Anyway, that’s my thoughts. Been lurking for a while, but this topic is dear to me as I have been fighting internally about whether to invest into more mining gear.
-
-
Will you look at that, we´re at 12kM/hash again ::)
