Bitcoin Forum

http://dot-bit.org/tools/nextDifficulty.php

There is something very interesting happening to namecoin. The surge in interest, followed by the sudden drop a few weeks ago has had the effect that difficulty has remained so high that miners are losing interest and leaving, further postponing the decrease in difficulty.

Given that most of the btc mining is for profit, the same thing might happen to btc if we see something severely affecting miners. If we lose 80% of our mining network for some reason it could take up to 10 weeks for the next difficulty adjustment. If the remaining miners then pull out since all that hard hashing isn't worth it, it would take much, much, much longer.

That can only happen if BTC loses 80% of it's current value, which is defintely a possiblity.

Gets worse, The low power also means every block takes longer, so transactions take a LOT longer to process, making Bitcoins harder to use as a currency.

Want to shut down bigcoins as a government? Do that - alternating pu a server famrm there, take it off. You will get all bitcoins every 2nd "round" and kill the network speed with variance in the meantime.

the namecoin folk could have altered their software to adjust difficulty more often, seems like the 2016 blocks between adjustments was a conscious decission, let them live with that. they will manage eventually to finish the current difficulty's mining round

Well the cool thing is that if that were to happen, then there would be the option of adjusting the source.

Quite ignorant statement, or?

I am not arguing that - and guess waht the ssame 2016 block thing is in bitcoins, but pointing out a danger. if someone adds 75 tera in one block dfficulty jumps up some times. When he goes, we are stuck witha a long slow block. Do that alternating aoud you undermine hthe viability of the currency - and take out a LOT of income from the miners, so they quit, too. The result would be a quite interesting DDOS with serious implications.
Yes, expensive. Not something governments can not just do.

And the same thing will happen after 22/07/2011, when the namecoin difficulty is finally adjusted. Assuming the price of namecoins stays the same, namecoin mining will be significantly more profitable than bitcoin mining, causing a lot of miners to switch for a few days it will take to mine 2016 namecoin blocks, making namecoin mining unprofitable again, and everyone will quit. Rinse and repeat. Same can happen with bitcoin easily enough.

There is no technical reason why difficulty can't be adjusted more often, even after every block (correct me if I'm wrong). You can still count the new difficulty from the last 2016 blocks, but just do it more often. If not every block, then something like every 6 (hour) blocks or so. I can't imagine any downside to adjusting it more often, only benefits. It would completely remove the kind of problems namecoin is having now.

But you're forgetting a major downside to adjusting every 10 blocks instead of 2016 --

Every script kiddie in the universe would have a much higher motivation to initiate DDoS attacks on the pools -- after all, they'd bring difficulty down within a short period of time, and then they could start reaping the rewards.

We don't need that!

Why would people stop mining when many do not care whether they make a profit or not? See the Folding@Home project. They all do it for nothing.

I have 2.25 gigahash/sec mining power. My electricity is practically free. Why would I stop mining? Because its hard? Is that something new? Mining has always been hard, that's why a lot of us purchased Radeon GPU's.

There are hundreds of thousands of people with only a single GPU that will never stop mining just to keep the network alive.

I agree that there are many people who mine for fun, and who will mine "for the cause" to keep the network alive.

I disagree that there are over 10,000 miners total, and that includes casual, part-time, and NVidia card/CPU miners.

Does anyone have evidence to the contrary?

Not true. Even if you somehow would manage to DoS the whole bitcoin network for one day, so that basically no blocks are generated, that would lower the difficulty only by 7% or so, the same as with the current system. So using DDoS to lower difficulty levels is just as effective now (not very) as it would be in a more frequent system. Remember, the new difficulty would still be calculated from the last 2016 blocks, even though the adjustment is done more often.

Well, if we're hypothetically switching to every 10 blocks, why still consider the last 2016 in the calculation?

But even so, I hold that if there were a measurable payoff, DDoS attacks on pools would be that much more common.

Heck, they DDoS now, even though it does no "good" at all (lowering difficulty level).

Because by considering only last 10 blocks, the variance would be so high, that the difficulty would be jumping up and down by 1000%.


If there were, sure, but there isn't.


If in a two week period you manage to DDoS enough, so that 5% of the blocks that would have been solved end up not being solved, the next difficulty level is 5% lower than it would have been otherwise. Same is true for the current system, and my proposal.

I have to say I do think this is an interesting thought but bitcoin has probably more miners that don't switch back and forth where namecoin does.

What I'm curious about is when the payout per block drops from 50 to 25 to see if people start dropping out as with the rise in difficulty some might not be able to turn a profit if that is what they are after unless the price increases.

IMHO Bitcoin already has so much "velocity" that even if we lose 80% of our mining network (I'm not sure if you meant 80% of the people doing mining, or 80% of the total hashing power, but I'll assume the latter) we'll still break through to the next difficulty adjustment before everybody completely bails.  Satoshi should have built in a difficulty adjustment based not only on block count, but also difficulty duration.  Or maybe there is one already?  I dunno.

If if if. If a pink dragon flew by my house and dropped a load of dung on my miner I wouldn't get any more blocks either.

Blocks/hour    6.23 / 578 s

The reality is the bitcoin network is still growing.

With this new patch (http://forum.bitcoin.org/index.php?topic=29074.0;all), namecoin difficulty is only going up. I suggest you start mining now.

You realize that's not even proposed to happen until block 24000 right? (Current NMC block is ~17000), if it actually happens.

As for the original point, it's not very unexpected, I and probably others mentioned this a while back. It's certainly possible, but it's more likely to happen as a result of other catastrophic failures for bitcoin spelling its inevitable doom anyway.


Difficulty adjustment is 100% based on difficulty duration, when it occurs/is calculated however is based on block count. Not sure if that's what you were thinking.

Err.. I guess I should have worded my last entry a little better, and I'm probably showing my lack of understanding how difficulty decreases work here.  What I'm getting from this thread is that difficulty decreases don't happen until block 2016 is found, for a given difficulty period.  What I was trying to say is that a difficulty decrease should take place if we reach the end of a difficulty period even if block 2016 hasn't been found.  Or is that the way it works already?  Please correct me if I'm wrong.

Difficulty could be a sliding window based on the average time of the last 1000 blocks, so it doesn't jump in big leaps or 'get stuck' at a high difficulty, like namecoin, where it can take two hours (http://explorer.dot-bit.org/b/15995") for the next block and two months for a difficulty reset.

Bitcoin could also have 1 minute 5BTC reward blocks - transactions would be confirmed fast even if you triple the required confirmations and increased orphan blockchain lengths, and mining wouldn't be as much a lottery.

Devs would just have to make a client that decides a point far in the future (like block 300,000) is where the new calculations kick in, so that it can be widely distributed before the change.

Yes, I suppose that could have been done, in a sense. There is a hardcoded x4 change maximum, either up or down. It would make sense for the difficulty to change when that value was reached (that is to say 8weeks), and unfortunately I do not believe that is the case. Perhaps it was done so that every node would know precisely when to calculate/update difficulty rather than relying on a possibly varying time-reference.

That is to say it is easy for every node to check the blockchain and get an update "Block xxx was found, recalculate difficulty", it might be a bit more difficult for them to say "Oh XXX time has passed, let's check in and set the difficulty." Not sure if that is the reason, but it seems logical to me.

When pools introduce "merged mining" this will no longer be a problem for Namecoin.
"Merged mining" is: using the same hashes of your GPU's for both blockchains.

http://forum.bitcoin.org/index.php?topic=29074.0

It just makes it a shared problem, so that when hashing power goes down, it goes down for all blockchains. When Bitcoin and Namecoin mining are merged, the total hashing power will still be driven almost entirely by Bitcoin profitability (because you would be getting an equal amounts of bitcoins and namecoins, and bitcoins are much more valuabe). So any scenario that would make Bitcoin hashing power drop suddenly right now (pick your favorite), would cause merged hashing power to also drop, causing problems to both chains.

It's true that merged mining would fix the current problem Namecoin is having, but only until Bitcoin faces similar issues.

It's crazy the amount of people who just have no grasp on Bitcoins.

The core of bitcoins isn't about "miners" making "profits". It's a currency.

If BTC was $2 or it's $20, it doesn't change anything. Someone selling a $10 item sells it for 5 BTC or .5 BTC. The buyer and seller don't care what BTC are worth. The buyer buys BTC, then buys the item. The seller sells his BTC. It's the same regardless of the price.

People run 5-10-20 folding at home rigs FOR FREE. Actually, they PAY to run folding at home rigs. They buy hardware, pay for power, and get nothing in return. If BTC was at $2 it would make more money then folding. People would still do it. If 80% of the miners left it wouldn't mean a thing long term for BTC.

BTC is about people USING it as a currency, not as a way to pay your rent. The people who power Bitcoins are the people buying them, using them, selling goods/services for them. Difficulty adjusts, transactions get processed, if its 2000 Ghashes or 20,000 Ghashes it really doesn't matter. The miners who think they are the people who make the Bitcoin world go round are fooling themselves. All of deepbit could fuck off right now and no one would notice. All of Slush, Deepbit, BTC could fuck off right now and no one would notice after 3-4 weeks. That's 80% of the network, CYA!

Difficulty drops to 400,000, old miners who left probably jump back in, new people join up, people will do it for free, correction people will PAY to mine BTC at $2.

In fact, the miners trying to make quick money are doing more harm to BTC then good. They dump their BTC as quick as they mine them trying to recover costs. They post on the forum that the sky is falling everytime MtGox drops $2. You think the guy sending money to MtGox, buying coins, using coins to buy something, really cares if it's $10 or $20. Everyone uses the usd/euro/etc.. equivalent anyway. If you look at the marketplace no one is selling for XXX BTC firm. It's $$$ at whatever the going rate is for BTC. There's even auto updated APIs that will match a $20 item to the current value of BTCs.

Go check out the "post your hash rate thread". You'll notice the majority of them are under 200mhash. There are lots of people hashing away at 10mhash. The whole "mining" business is about 800 guys trying to make some money and 25,000 people doing it for fun. Trust me the 25,000 small time miners don't need those 800 hardcore guys. Like I said, those 800 guys are probably causing more drama then they're worth.

Do a survey and ask how many people actually treat BTC as a real currency. I guess you'll be disappointed. Many mine it just for profit, and many buy it just for speculation, many hoard it and have no intention to spend it.

Yes, there will always be a group of miners who would mine at loss, and if 80% of the miners suddenly left, the difficulty would eventually adjust. However, there would be a 10 week period, where only about 1 block would be solved per hour. While it would not be the end of Bitcoin, I think we can all agree that it's not a desirable situation.

Also, I think you are significantly underestimating the number of people who are mining just for profits, who would quit mining the very minute it becomes unprofitable, but I suppose your guess is as good as mine. No one really knows.

It's not such a big deal with Bitcoin. There are significant market exit barriers to Bitcoin mining. With Namecoin there are not, because every Namecoin miner used to mine Bitcoins until the "dreaded" tvori chart appeared that said "go mine Namecoins, it's more profitable" and a lot of people (though not everybody, and this is important) switched. After the difficulty change it was easy for everybody to just go back to Bitcoin mining, leaving the young and undeveloped Namecoin chain to starve.

With Bitcoin, what else is there to do? Even if a more profitable method came along, it takes some time for people to switch, due to market entry barriers. Namecoin for example had the problem that there was (and is, afaik) no nifty GUI client, you had to generate a valid address first on the command line, plus no USD/NMC exchanges yet, even more volatile than Bitcoins, etc. etc. That protected Bitcoin from an instant drain of mining power, even though homo oeconomicus would have switched mining to the more profitable (at that time) Namecoin.

So what we would see with Bitcoin mining is a slower decrease in hashing power, dragging the time until the next difficulty change a bit but not too much. Adding to that is the fact that there are a lot of Bitcoin miners at very different profitability levels. You will see people dropping out one after the other. Heck, my energy cost is equivalent to $.30/KWh and I'm still mining! Until it hits the level where, say, $.07 per KWh is too expensive, there will be no "collapse". Bitcoin could technically survive on just one machine generating blocks. Although that would mean certain death in terms of acceptance. But the only real problem with a dropping difficulty is that people might have to wait longer for their confirmations.

That means that a block generation time of 15 minutes might not be unheard of, but I highly doubt that that's enough "lag" to drive people away. 30 minutes or more, sure, maybe. But we are currently not at the point where it matters much how long you have to wait for confirmations. Even with POS systems, they are looking for ways to be totally independent of block generation intervals, so in the long term it won't matter at all, I guess.

tl;dr summary: Bitcoin mining will not collapse for the market exit barriers, thus people will not have to wait for blocks for longer than usual and won't quit Bitcoin altogether.

Wrong.

People run server farms and get paid nothing right now. People will mine BTC because it's fun. People lose money to run folding. People will keep mining regardless of the BTC value.

If we went from 10,000 ghash to 2,000 ghash over 2 weeks, sure block time would go to 30-60 minutes. Within 2-3 weeks difficult adjusts and no one would miss that 8000 ghash AT ALL. Would have no effect. Bitcoin runs just as smoothly at 500 ghash as it would at 50,000 ghash. Bitcoin doesn't need professional miners. Period.

Again, bitcoin doesn't need professional miners doing it for money. Bitcoin ran fine when the network had a total of 1ghash of power. The network doesn't need 10,000 ghash. People who are mining think the network needs 10,000 ghash. Like I said, 80% leave blocks would slow down for 2-4-6-8 weeks then everything returns to normal.

The hardcore miners are doing it for their own personal gain, the network doesn't need them and will happily go on after they've left.

Exactly. 90% of miners leaving would take 3-30 days to go from 11,000 ghash down to 1,500 ghahs. During that time they'd be at least one adjustment, and we'd be close to another. Blocks solved every 30 mins doesn't destroy bitcoins. MtGox the biggest exchange was hacked, $10,000s of dollars stolen, gmail accounts stolen, it was hardly a blip. The network goes on.

Bitcoin will run quite happily at 1,000 ghash. Which means everyone in this mining sub-forum, isn't really needed. We're here to make money for ourselves we could really care less about the network. And the network could care less about us.

I don't think anyone is arguing that bitcoin wouldn't run just fine with a fraction of the current hashing power (security concerns aside).

The point is that bitcoin responds much worse (slower) to a sudden drop in hashing power, than it does to a sudden increase. While some people apparently don't see it as much of a problem, it's still a problem we don't need to have, however small. The only change needed would be to adjust the difficulty more often. It wouldn't make too much of a practical difference in situation where hashing power is increasing or staying the same, but would perform much better in case of sudden loss of miners, however unlikely that is.

But whatever...

Maybe you are technically right about Bitcoin working fine, but it would be naive to think that most people would continue to mine at a loss. I bet a good half of the people who are involved in Bitcoin right now got started because they heard they could use their computer to make money. The ideological reasons are nice, but Bitcoin wouldn't do anywhere near as big as it is today without the initial incentive. Folding@Home has its own incentive, which is to cure diseases. Mining at a loss to support a financial system doesn't make much sense. It's a bit like volunteering at McDonald's because you like their hamburgers.

lol. tell the folding@home guys they do it "for nothing".

it's for advances science and medicine that help fighting diseases! ...

However, they won't drop out all at once. I've seen people exit last week complaning "it's not worth the noise for $2 a day", while others will (and can) happily churn along for months to come. Mileages vary drastically, price of electricity, funds available to spend, rig already paid off or not, cold climate, running rigs in the bedroom vs. the basement... all of these factors make up individual exit points for everybody so that the decrease cannot be drastic.

I'm too lazy right now to calculate, but how long can it take at the most until 2016 blocks are solved given that a sudden event (such as the price of BTC plummeting to $1) immediately after a difficulty change, causing 75%/80%/90% of miners to turn off their rigs? Someone should do a spreadsheet...

Assuming a scenario where the drop happens right after the difficulty change, it's:

two weeks/percentage of the miners left

So for 90% of the miners quitting, it's 2/0.1 = 20 weeks = 4.5 months of solving 1 block every 1h40m.