If there's a huge drop in hash power..

[Bit_Happy]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

[skooter]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

Well, that's for an extreme 95% hash power leaving the network example.

If 3/4s of the hash power leaves, it'll be 40 minutes / block.

[Bit_Happy]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

Well, that's for an extreme 95% hash power leaving the network example.

If 3/4s of the hash power leaves, it'll be 40 minutes / block.

I had forgotten that BTC had that many days between diff changes.
Really amazing if you think about it.

[skooter]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

Well, that's for an extreme 95% hash power leaving the network example.

If 3/4s of the hash power leaves, it'll be 40 minutes / block.

I had forgotten that BTC had that many days between diff changes.
Really amazing if you think about it.

Well it's supposed to be 2 weeks per change. Right now it ends up being like 12 days because so much hash power is added every day. If 3/4ths of the hash power drops out that increases to 8 weeks. (if the GH drops out at the start of a "cycle")

[Bit_Happy]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

Well, that's for an extreme 95% hash power leaving the network example.

If 3/4s of the hash power leaves, it'll be 40 minutes / block.

I had forgotten that BTC had that many days between diff changes.
Really amazing if you think about it.

Well it's supposed to be 2 weeks per change. Right now it ends up being like 12 days because so much hash power is added every day. If 3/4ths of the hash power drops out that increases to 8 weeks. (if the GH drops out at the start of a "cycle")

Thanks for the details...I'm not going to ask why that "feature" wasn't designed better. (Maybe I'll read those threads someday)
Bitcoin works really well most of the time.

[skooter]

200 minutes per block?
Doesn't sound likely, but we have alt coins to save us. 

Well, that's for an extreme 95% hash power leaving the network example.

If 3/4s of the hash power leaves, it'll be 40 minutes / block.

I had forgotten that BTC had that many days between diff changes.
Really amazing if you think about it.

Well it's supposed to be 2 weeks per change. Right now it ends up being like 12 days because so much hash power is added every day. If 3/4ths of the hash power drops out that increases to 8 weeks. (if the GH drops out at the start of a "cycle")

Thanks for the details...I'm not going to ask why that "feature" wasn't designed better. (Maybe I'll read those threads someday)
Bitcoin works really well most of the time.

Oversight?

Either way, when the time comes where the majority of the hash power is abandoning the network rapidly, you're probably looking at an extreme devaluation of the crypto. Energy prices do increase over time, but that'll cause hash power to gradually leave the network, not abruptly.

[rgm108]

Wouldn't that cause issues?

Like let's say, the difficulty just reset, so we have 2016 blocks to go till the next change. At the current hash power, it'll take ~ 10 minutes / block like it's supposed to.

And for whatever reason, 95% of the hash power is taken off the network (which is NOT totally unrealistic. If there's a rise in electricity costs or a decline in bitcoin prices to the point where power cost > value of bitcoins mined, a lot of ASICs will be taken offline, since even if you wanted the bitcoin, it'd be cheaper to not run the ASIC and buy the bitcoin instead). The difficulty is still the same, and will remain the same for 2016 blocks. If 95% of the GH disappears, then each block will now take 200 minutes to find on average, for the next 280 days.

We the remaining little guys would point our rigs away from alts and point them to mine bitcoin in the false hope of making a buck or 2 whilst securing the network in part.
I would be willing to do this at a cost just to make sure it survives.

[skooter]

Btw, someone said that if 95% of the network hash power left, the remaining 5% would see a 20X increase in mining reward. That's simply NOT true.

It WOULD be true AFTER the difficulty resets to a lower level. But up until that point it would take you just as much electricity to generate 1 BTC as it did prior to the network collapsing.

[amspir]

Here's something to ponder, why a drop in hashrate is not likely:

Assume the current rigs composing the network hashrate were purchased at a price similar to the new rigs coming out, that is at $3 per GH/s, or $3000 per TH/s.  The actual value is much higher, considering the old technology that will still be online for a while.

Today, the network hashrate is 50,000 TH/s, which would correspond to $150 million dollars worth of new mining rigs.   95% of that is $142 million, and I don't see how you can make miners to turn off $142 million worth of equipment.

[lightfoot]

Thanks for the details...I'm not going to ask why that "feature" wasn't designed better. (Maybe I'll read those threads someday)
Bitcoin works really well most of the time.

Read the catcoin threads on this one: Once you start setting low retarget times it's simple for people to game the coin by switching to it when difficulty is low, running the diff way up, then abandoning it to the suckers when the diff is high. Wash rinse repeat, a longer retarget helps to mitigate this.

It's amazing how people will try to game anything possible out there.

[DannyHamilton]

Btw, someone said that if 95% of the network hash power left, the remaining 5% would see a 20X increase in mining reward. That's simply NOT true.

No, I didn't.  You are misquoting, and then arguing against a misquote.  You'd make a great politician.

The original quote was:

If 95% of the miners shut off their equipment, the remaining miners would see an increase of 20 times as many bitcoins for each block that they solve.

Which is true.  There would be a huge gap between solved blocks, but since there would be 95% less miners participating in the pool, the block reward wouldn't need to be split into such small pieces.  Each miner participating in the pool would get a significantly large piece of the reward whenever their pool managed to solve a block.

It WOULD be true AFTER the difficulty resets to a lower level. But up until that point it would take you just as much electricity to generate 1 BTC as it did prior to the network collapsing.

You'll notice that the original quote said nothing about the amount of electricity it would take, only that the received reward would be larger when the block was solved.

[lnternet]

This is definitively a flaw in the design that needs to be fixed. Testnet ran into this problem before. I'm sure the core devs are aware, but it is low priority.

[Klestin]

Since the introduction in ASICs, there's never been a drop big enough where energy cost of mining > value of btc produced.

Ahh but it has happened under the video card era, and there was no drop.  There's no reason to believe ASIC operators would be any different.

You can bet those datacenters spending thousands of dollars in electricity a day will turn off their ASICs if it turns unprofitable.

Only if those data centers are relying on immediate BTC sale to pay their day to day bills. If they're in it for long positions, the daily exchange rate means diddly squat.  

Yes, if the exchange rate drops precipitously and stays low for a long period, some miners will drop off and instead put their money into direct purchase.  Will it be 95%? No.

Also remember that there's more to mining than just acquiring BTC.  If there were not, there would be no ASIC sales. There's not a single machine you can order today that will make more than it costs in BTC over its lifetime. Yet they still continue to sell.

[bryant.coleman]

The hash power is still increasing.

Right now any NEW equipment will likely not generate a positive ROI, but on existing equipment marginal profit (bitcoins mined) still > margin expense (cost of electricity).

From where you'll get NEW equipment. BFL normally takes 9-12 months to ship the mining rigs.

[lightfoot]

Ahh but it has happened under the video card era, and there was no drop.  There's no reason to believe ASIC operators would be any different.

I wouldn't bet on that too much. I've been doing the math, and I will be shunting my 65nm equipment down to "minimum power draw" mode at the end of May. Reducing the voltage to the chips and the clock rate will cut the power use in half, along with the heat generated while dropping the hashing speed by 20-25%. That should keep me going till August, at which point I will shut the miners down.

There is no economic reason to mine once the cost of electricity is more than the bitcoin value generated. It's not even a dollar auction, the hole you're digging is getting deeper because of the electricity costs.

The economic answer is to buy someone else's bitcoins. Which will push up the price of BTC which will cause miners to come back which will re-establish equilibrium.

Anyone still mining btc with a GPU is being totally economically irrational. What the GPU people seem to have done is hit on this whole Litecoin baloney thing which gives their equipment a pseudo reason to exist. Why people pay good bitcoins for Script is not something I quite understand.

Only if those data centers are relying on immediate BTC sale to pay their day to day bills. If they're in it for long positions, the daily exchange rate means diddly squat.

If you're in it for a long position then why not stop mining and buy the bitcoin on the market once it doesn't cover electricity costs? Let someone else pay the difference for you.

Also remember that there's more to mining than just acquiring BTC.  If there were not, there would be no ASIC sales. There's not a single machine you can order today that will make more than it costs in BTC over its lifetime. Yet they still continue to sell.

Exactly. This is not economically rational, people do it anyway. If you do choose to be such, own it and accept it. It's fine, it's entertainment and all that.

C

[skooter]

Since the introduction in ASICs, there's never been a drop big enough where energy cost of mining > value of btc produced.

Ahh but it has happened under the video card era, and there was no drop.  There's no reason to believe ASIC operators would be any different.

You can bet those datacenters spending thousands of dollars in electricity a day will turn off their ASICs if it turns unprofitable.

Only if those data centers are relying on immediate BTC sale to pay their day to day bills. If they're in it for long positions, the daily exchange rate means diddly squat.  

Yes, if the exchange rate drops precipitously and stays low for a long period, some miners will drop off and instead put their money into direct purchase.  Will it be 95%? No.

Also remember that there's more to mining than just acquiring BTC.  If there were not, there would be no ASIC sales. There's not a single machine you can order today that will make more than it costs in BTC over its lifetime. Yet they still continue to sell.

You think someone who has millions to invest in an ASIC datacenter is stupid?

If the cost of mining a coin = $10 in electricity, and the current price of bitcoin is $5, someone who's in it for the long position (and has half a brain) will turn off their ASICs, and spend the money they would've spent on the power bill buying the bitcoin on the market.

Pretty simple math here.

[Tirapon]

Since the introduction in ASICs, there's never been a drop big enough where energy cost of mining > value of btc produced.

Ahh but it has happened under the video card era, and there was no drop.  There's no reason to believe ASIC operators would be any different.

You can bet those datacenters spending thousands of dollars in electricity a day will turn off their ASICs if it turns unprofitable.

Only if those data centers are relying on immediate BTC sale to pay their day to day bills. If they're in it for long positions, the daily exchange rate means diddly squat.  

Yes, if the exchange rate drops precipitously and stays low for a long period, some miners will drop off and instead put their money into direct purchase.  Will it be 95%? No.

Also remember that there's more to mining than just acquiring BTC.  If there were not, there would be no ASIC sales. There's not a single machine you can order today that will make more than it costs in BTC over its lifetime. Yet they still continue to sell.

You think someone who has millions to invest in an ASIC datacenter is stupid?

If the cost of mining a coin = $10 in electricity, and the current price of bitcoin is $5, someone who's in it for the long position (and has half a brain) will turn off their ASICs, and spend the money they would've spent on the power bill buying the bitcoin on the market.

Pretty simple math here.

Or they might keep it running to prevent a drop in hash rate which could threaten their investment.

[vpitcher07]

You bring up a really interesting point which I've also wondered about for a while. People saying that it's impossible to happen but I disagree. If we see a mega crash down to less than $1 per coin I wouldn't be surprised to see a major drop in hashing power. I also think the network is vulnerable to attack this way although it'd be extremely costly and wouldn't exactly bring down bitcoin. Suppose an attacker was able to obtain a large amount of hashing power but not 51%. He then could drive up the difficulty (along with the networks average difficulty increase) and then instantly shut off his miners causing a massive drop in hashing power. That would cause very long confirms. Even if a major pool like BTC guild went offline, we would see way longer confirm times. If we see a major crash, coinciding with an attack, it's very possible. (Just not likely).

[Velkro]

developers already know about this problem, so they won't be suprised and until it happen solution will be there

[DannyHamilton]

Suppose an attacker was able to obtain a large amount of hashing power but not 51%. He then could drive up the difficulty (along with the networks average difficulty increase) and then instantly shut off his miners causing a massive drop in hashing power.

If he has less than 51%?  Than he can only cause a drop of 50% at most.

That would cause very long confirms.

A drop of 50% would only increase confirms from an average of 10 minutes to an average of 20 minutes.  If it happened immediately after a difficulty adjustment, the network would decrease the difficulty 4 weeks later and we'd be right back to 10 minute confirmations.

Even if a major pool like BTC guild went offline, we would see way longer confirm times.

If a pool goes offline, miners wil jsut move their equipment to a new pool.

If we see a major crash, coinciding with an attack, it's very possible. (Just not likely).

The biggest risk would be a drop in exchange rate to something less than $1 in a single day that does not recover back above $1.  This is pretty unlikely, but not impossible.  This is one of many reasons that bitcoin is considered experimental and still in beta test.