Difficulty post ASIC?

[Unacceptable]

The way I see it:What's easier to 51% attack,a 10TH network or a 250TH network...............................

I just don't see Bitcoin being centralized at any point,sorry.

BTW,I was hacked on MTgox & was told "call the police & get a report & send it to us",guess what the police said........wtf are BTC?...what do you want us to do?.............I don't think any gov agency will have anything to do with BTC anytime soon,(unless SR gets too big & money laundering gets out of hand)maybe many years down the road,MAYBE.

The local authorities are not an issue and sadly like you experienced, not much help either.  A 12+THash (current avg) network is actually impressive by using off the shelf hardware and some specialized hardware.  

I am unsure the costs at this stage are going to be worth the network speed increase.  We are assuming bigger is better.   Profit Incentive is a very important aspect of the BTC concept.  It is what allows many disassociated parties to collectively work towards a common goal in not a purely socialistic or communistic manner other than the important fact that we all share in the benefits of the network.

I believe it IS worth it.I have done some figureing(up to 240th),I feel I can be as competitive with 50gh ASIC as I was with my 2gh of GPU's,for about 4 months,maybe less,but we'll see.Then I'll roll some of my profits into another 40gh & so on...................

The very low power consumption makes it VERY worthwhile.Up to 70% of my mining proceeds currently pay the electric bill  I will be INCREASING my profit margin,ALOT 

Until someone forks the chain This would be very destructive to the integrity of Bitcoin,at least for me, it will show that folks have lost faith in this concept.

[SgtSpike]

I'd just like to make a few points.

- It is likely that, as faith in Bitcoins increases, and price stability increases, more professional companies who are willing to look at ROI in terms of years instead of months, and who already get excellent incentives on electricity, will start making larger investments in mining equipment.
- There will always be an aspect of volatility.
- There will always be an aspect of the "average joe" being able to mine and make a profit.  If it is profitable for a company, it can be profitable for an average joe.  The break-even only has to do with electric rates, and the ROI will settle around a certain number of years of break even compared to electric rates.
- The ROI won't increase to years until we either have more faith in Bitcoin, or a significant price decrease in BTC.

- Why is a years-long ROI a bad thing?
- Why are large scale mining operations a bad thing?
- Why can't we push the merits of Bitcoin as a transactional system instead of a magic money making machine?

If ROI really is pushed out to years-long, then that means Bitcoin will be better secured than if ROI is months long, because it means more people and companies are mining. Isn't that a very GOOD thing? It means it is that much more difficult for someone to stage an attack on the network, does it not?

If large scale mining operations are put into place, isn't that also a very GOOD thing?  Again, more hashing power protecting Bitcoin, and more people taking it seriously (as evidenced by people making large-scale investments).  It also means that those mining companies would work hard to keep the price of Bitcoin stable - they want to have consistent returns, not returns where they might be losing money one day, and making twice as much the next.  If mining companies are holding back on the sale of coins to help stabilize the price, is that not a good thing?

On top of all of this, it will NOT be centralized mining.  It will be a variety of companies or business groups making a variety of investments in a variety of sizes of mining operations.  Just because you can't profitably mine at residential electric rates doesn't mean it is centralized.  It just means that mining has moved from a hobby to the real business world.

Centralized is the buzzword around here, but everyone is using it wrong.  If there was only one person or company in the world mining Bitcoin, then I would call that centralized.  But a group of businesses competing with each other is the opposite of centralization.

This is not the end of Bitcoin - this is just the beginning.  Sorry you don't like the way it is going, but deal with it.

[Raize]

Sgt, don't misinterpret my words as disgust regarding Bitcoin as a whole. I just think there are a LOT of miners that are not thinking things through and don't understand just how difficult mining is going to be getting, and that they could still become miners again after second-gen ASIC blows first gen out of the water.

I'm really worried about these sorts of things, because I get the impression that a lot of miners will be under the impression they are owed something for the work they do, and they aren't going to be very pleased with the reward halving for sure. It could be this translates into a hoarding of Bitcoin at some unspecific point around the beginning of the difficulty adjustments, but given that so many people will have already wrapped up funds in ASIC hardware, it could be these folks that miss out the most...

[SgtSpike]

Sgt, don't misinterpret my words as disgust regarding Bitcoin as a whole. I just think there are a LOT of miners that are not thinking things through and don't understand just how difficult mining is going to be getting, and that they could still become miners again after second-gen ASIC blows first gen out of the water.

I'm really worried about these sorts of things, because I get the impression that a lot of miners will be under the impression they are owed something for the work they do, and they aren't going to be very pleased with the reward halving for sure. It could be this translates into a hoarding of Bitcoin at some unspecific point around the beginning of the difficulty adjustments, but given that so many people will have already wrapped up funds in ASIC hardware, it could be these folks that miss out the most...

Ok, so your post is more of a warning to miners than anything else?  Fair enough, I can agree that that isn't a bad thing to do.  Plus, it disincentivises mining, thereby increasing your own mining profits in the meantime. 

[Raize]

Yes, but believe me, that's not my fullest intent. I do mine with FPGA and ASIC will "crush" my operation, but I have the financial capability to invest in first gen ASIC and I'm not because I see the early adopter advantage being very small, especially given the length of time any funds invested in such an operation will be held up. Too high of a risk vs. reward IMHO, but I suppose we'll see if I'm wrong.

I see a revisiting of the period of time when GPU miners first started considering buying more than one unit to mine with. Say, March 2011. And the followup of 6 months that took the price per coin from a low of $.60 to $30+. It wasn't profitable to buy GPU miners as much as it was to buy coin during that difficulty increase. In fact, had you bought just $500 of coin at $.80 in March, sold at $20 in July and started a GPU farm of 15 mining machines netting you an average profit of 12 Bitcoin/day (actual profit over this period of time was quite a bit higher where I live, and nearly double that when the difficulty hit its bottom), then today you would have at least 3,500 coin, enough to have paid off those GPU miners, invested in FPGA, and pocketed a nice profit instead of buying a dual GPU unit at that $500 when it was first viable. Still better would have been the guy that sold at $20 and rebought at $2, but folks still say you can't time the market (especially if you ignore when difficulty is falling and when it is recovering). I disagree.

I guess what I'm saying is that if you expect the price to be falling drastically from $6 then it's time to invest heavily in miners, and I just don't see that happening right now. Not with the inevitable difficulty increases ahead of us and the block reward halving. Especially with graphs like these showing Bitcoin interest and use at all time highs:
https://blockchain.info/charts/n-transactions-excluding-popular?timespan=1year&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

https://blockchain.info/charts/market-cap

So if difficulty isn't falling and is, in fact, increasing exponentially, now is definitely not the time to be buying miners. It's too risky and we don't know where/when the difficulty increase is going to stop/decline.

[ldrgn]

- Why is a years-long ROI a bad thing?

Risk.  On top of that, remember that bitcoin mining is perfectly competitive - the longer this time period gets the closer you'll be getting to the magical long-term point where you are no longer profitable.

[sadpandatech]

BFL Trade-In impact on Bitcoin Difficulty
Speculation based on my complete mastery (not)

Time frame of 5 months.  Assuming BFL delivery matches past performance.
Per SC chip output velocity will be greater with at least 10x chips per SC Single.

25 BTC block reward.  It's too close to call whether BFL can get SC delivered prior to reward halving.

BFL Singles sold (estimate)

500

BFL Mini Rigs sold (estimate)

20

Network capacity represented

Singles - ~ 400 GH
Mini Rigs - ~ 500 GH


BFL Trade-In network capacity impact
using a trade-in scenario of 4 Singles (FPGA) converted to 3 SC Singles, 2 Jalapenos (90% trade-in conversion to SC)
Mini Rig.  Straightforward conversion.  1:1 plus capital.

4 Singles trade-in credit - $2,400
Trade-in cash obligation  - $2,400
$4,800

3 SC Singles - $4,500
2 SC Jalapenos - $300
$4,800

1 Mini Rig trade-in credit - $15,000
Trade-in cash obligation - $15,000
$30,000

SC Mini Rig - $30,000


Single capacity loss (FPGA)

360 GH

Mini Rig capacity loss (FPGA)

500 GH


SC Single/Jalapeno capacity gained

14.3 Terahash

SC Mini Rig capacity gained

20 Terahash

Current network capacity (Mini Rigs/Singles not yet delivered and online added)

~ 12.5-14.5 Terahash

Minus Single (FPGA) capacity

~ 12.1-14.1 Terahash

Minus Mini Rig (FPGA) capacity

~ 11.6 - 13.6 Terahash

Minus GPU attrition (Let's say an initial conservative loss of 70% GPU capacity where GPU is assumed to represent 85% of current network capacity)

~ 3.6-5.6 Terahash

Minus non BFL FPGA attrition (Same attrition estimates as GPU)

~ 2.7-4.7 Terahash

(As you can see the advent of ASICs for bitcoin hashing decimates prior technology.  In and of itself is fine.)

Plus BFL Trade-in capacity

SC Singles

~ 17-19 Terahash

SC Mini Rigs

~ 37-39 Terahash


Resulting hashing difficulty increase

~ 181%

Average pre-existing BFL miner capacity increase (using trade-in figures above)

~ 39x increase in capacity

Current vs. BFL SC profitability per GH  (including block reward halving)

 -82%

BFL SC Profitability/ROI (Excluding prior investments.  Exchange rate remains constant.  Block reward - 25 btc.  Excluding all other factors that would lead to difficulty adjustments)

~ 62 days


What we can immediately take away from this (considering only the figures above)

1.  GPU currently represents a large portion of network capacity.  Immediate effects of BFL trade-in activity could be blunted by GPU attrition.  Depends how sensitive to electricity costs.
2.  Miner profitability will be, at a minimum, approximately 82% less profitable per GH.
3.  Once additional BFL SC sales beyond trade-in transactions occurs we are looking at a significant additional difficulty increase.
4.  If one hopes that GPU actors will not participate in BFL SC ROI calculation looks very good.


Now.  Let's see what happens when GPU actors participate in BFL ASIC.

Assumptions (attempting to be conservative with estimates)

1.  40% of former GPU actors (that's 40% of the actors representing 8 Terahash of current capacity - those that suffer attrition) leave mining.  60% remain to rebuild their operations using BFL SC.
2.  These 60% of GPU actors choosing to rebuild operations spend approx. 60% of their initial GPU capital outlays on BFL SC units.


Equating current GPU capacity to capital

60% (of those suffering attrition) of 8 TH.  4.8 TH at ~ $1,200 per GH. $5.7 million. 60% capital re-up - $3.4 million.

BFL SC approximate cost per GH.  $38

Additional network capacity from re-uping former GPU actors

~ 89 Terahash

Total network capacity

~ 127 Terahash

Resulting hashing difficulty increase (from current difficulty)

~ 840%

Current vs. BFL SC profitability per GH  (including block reward halving)

 -94%

BFL SC Profitability/ROI (overall.  Excluding prior investments.)

~ 205 days

Quoting this so it gets seen again.. Will you have my babies? ;p

[SgtSpike]

- Why is a years-long ROI a bad thing?

Risk.  On top of that, remember that bitcoin mining is perfectly competitive - the longer this time period gets the closer you'll be getting to the magical long-term point where you are no longer profitable.

You have only answered the question as to why years-long ROI is bad for miners.  But miners are self-balancing.  If years-long ROI's are too risky, then miners won't continue to invest to the point where years-long ROI's are the norm, thus keeping it down into the months.  If miners are willing to risk buying even with years-long ROI's, then they might be called stupid or crazy, but they're still helping to give Bitcoin more hashing power to protect it.

It's just like ANY other business out there.  The more people invest into a particular market, the less room there is for competitors.  The less room there is for competitors, the less attractive it is for investing, and the less people will invest.  Those who already invested in the market have sunk costs, and might choose to ride it out or sell out and take the hit.  Those who sell out will help increase the profit of those who don't, and the ROI would eventually sink back to months instead of years again.

It's like you guys are complaining that gas stations open up near each other and compete, and it makes it more risky because it'll take longer for gas stations to recoup their investments, and some of them might not be profitable.  Ok... so??

So again, I ask the question, why is a years-long ROI a bad thing?  And I'll make it more specific:  Why is it bad to Bitcoin?

[ldrgn]

You have only answered the question as to why years-long ROI is bad for miners.

Yes, sorry about that.  Was just meaning to pop in and make a correction.

  But miners are self-balancing.

See the bit about perfectly competitive markets again, in the long term they stop balancing.  Endgame for Bitcoin looks like an oligopoly of a few very large, very efficient miners with microscopic/no profit margins.

So again, I ask the question, why is a years-long ROI a bad thing?  And I'll make it more specific:  Why is it bad to Bitcoin?

Relating to the first point I'm not arguing if it was good or not, just pointing out the way the system works.  Whether or not it is good depends on your idea of what a digital currency should be: if you highly value decentralization and a large pool of small miners Bitcoin is not designed to satisfy you in the long run.

[proudhon]

If everything is acurate then they will have chip with base potencial of 3.5GH/s @ ~2W. With little "overclocking" it will be 4GH/s @2.5W. 10 of these will be in Single= 40GH/s and 250 in MiniRig=1TH/s. I doubt they design 3 different chips. How many they can realese to wild in one week? 100 for sure. Lets assume 500 per month = 2TH/s. So difficulty will double in summer next year... With cheap electricity GPU miners don't have to worry for at least one year.
Thats of course speculation, that don't include others ASIC manufacturers.
My guess is that price of bitcoin will drop signifficantly becuse they will be "made" cheaply, very cheaply.

This. 

[flower1024]

If everything is acurate then they will have chip with base potencial of 3.5GH/s @ ~2W. With little "overclocking" it will be 4GH/s @2.5W. 10 of these will be in Single= 40GH/s and 250 in MiniRig=1TH/s. I doubt they design 3 different chips. How many they can realese to wild in one week? 100 for sure. Lets assume 500 per month = 2TH/s. So difficulty will double in summer next year... With cheap electricity GPU miners don't have to worry for at least one year.
Thats of course speculation, that don't include others ASIC manufacturers.
My guess is that price of bitcoin will drop signifficantly becuse they will be "made" cheaply, very cheaply.

This. 

but that does not mean that there are more bitcoins available.

i am not sure, but i still think price will go down after reward halfing (only through speculation - too rip people off which invested before the reward halfing)

i dont think asics will have a huge influence on price.

[crazy_rabbit]

Well we know that at *Least* a quarter million dollars of ASIC was bought yesterday with Bitcoin.

http://www.marketwatch.com/story/bitpay-shatters-record-for-bitcoin-payment-processing-2012-06-26

So that means a minimum of about 8.3TeraHashes is coming down the pipe in October. Anyone care to crunch the numbers about what that represents for difficulty adjustment?

[sadpandatech]

Well we know that at *Least* a quarter million dollars of ASIC was bought yesterday with Bitcoin.

http://www.marketwatch.com/story/bitpay-shatters-record-for-bitcoin-payment-processing-2012-06-26

So that means a minimum of about 8.3TeraHashes is coming down the pipe in October. Anyone care to crunch the numbers about what that represents for difficulty adjustment?

5 posts up from yours, contains about all the math crunching one can hope for at this point.

[Raize]

So again, I ask the question, why is a years-long ROI a bad thing?  And I'll make it more specific:  Why is it bad to Bitcoin?

I think you're missing the point of concern here.

People buy a product like a BFL device and expect it to:
1. Be delivered by X date,
2. Pay itself off within Y months, and
3. Hash at Z rate.

BFL has over-delivered on point #3, but their horrible performance on #1 and #2 means people are setting up for disaster.

Case in point: which is shipping first, the SC Jalepenos or SC Singles? How are two different ASIC devices going to ship simultaneously?

So why is a "years-long" ROI bad? Well, because the first people to get them aren't going to have that "years-long" ROI, they are going to pay off their BFL device within weeks. It's all the other fools that get their devices last that are going to have "years-long" ROI, and if you think that isn't going to be bad to see that kind of divergence, you're just being silly.

However on the very narrow point you mention, is it bad for Bitcoin? Absolutely not. Now's the time to buy, IMHO, time to take advantage of all the suckers that just wrapped up a quarter of a million USD in non-refundable deposits!

[SgtSpike]

So again, I ask the question, why is a years-long ROI a bad thing?  And I'll make it more specific:  Why is it bad to Bitcoin?

I think you're missing the point of concern here.

People buy a product like a BFL device and expect it to:
1. Be delivered by X date,
2. Pay itself off within Y months, and
3. Hash at Z rate.

BFL has over-delivered on point #3, but their horrible performance on #1 and #2 means people are setting up for disaster.

Case in point: which is shipping first, the SC Jalepenos or SC Singles? How are two different ASIC devices going to ship simultaneously?

So why is a "years-long" ROI bad? Well, because the first people to get them aren't going to have that "years-long" ROI, they are going to pay off their BFL device within weeks. It's all the other fools that get their devices last that are going to have "years-long" ROI, and if you think that isn't going to be bad to see that kind of divergence, you're just being silly.

However on the very narrow point you mention, is it bad for Bitcoin? Absolutely not. Now's the time to buy, IMHO, time to take advantage of all the suckers that just wrapped up a quarter of a million USD in non-refundable deposits!

...

So your ONLY point is that people who buy later than the first adopters will get a lower return?  Well, duh, they're coming to the party later!  And anyone who doesn't do the math before buying DESERVES to have a lower/longer ROI than they expected.

People are making a huge deal out of nothing, IMO.  It's the end of the world because some people might make bad investments that take a while to recoup?  Big freaking deal...

[nedbert9]

So again, I ask the question, why is a years-long ROI a bad thing?  And I'll make it more specific:  Why is it bad to Bitcoin?

I think you're missing the point of concern here.

People buy a product like a BFL device and expect it to:
1. Be delivered by X date,
2. Pay itself off within Y months, and
3. Hash at Z rate.

BFL has over-delivered on point #3, but their horrible performance on #1 and #2 means people are setting up for disaster.

Case in point: which is shipping first, the SC Jalepenos or SC Singles? How are two different ASIC devices going to ship simultaneously?

So why is a "years-long" ROI bad? Well, because the first people to get them aren't going to have that "years-long" ROI, they are going to pay off their BFL device within weeks. It's all the other fools that get their devices last that are going to have "years-long" ROI, and if you think that isn't going to be bad to see that kind of divergence, you're just being silly.

However on the very narrow point you mention, is it bad for Bitcoin? Absolutely not. Now's the time to buy, IMHO, time to take advantage of all the suckers that just wrapped up a quarter of a million USD in non-refundable deposits!

Yes,  BFL needs to up their game on point 1.  I want them to answer the specific question of how many staff will be operating each assembly line.

I don't believe their assurances that they can get these units out the door quickly.  And with the ridiculous amount of cash they will be pulling in BFL can certainly afford cheap final assembly labor.

I'm optimistic on ASIC chip production.  The chips are produced at a small fraction of the cost of FPGA and it's not like BFL will have any sourcing problems.  They alone are the customer.

So, if BFL doesn't up their delivery game it makes them look much, much worse.  I hope people get this.

[Raize]

So your ONLY point is that people who buy later than the first adopters will get a lower return?

One person bought a SC Jalepeno, one person bought an SC Bitforce Single, both yesterday. What is your bet that both receive product on the same date? That's not an "early adopter advantage" that's a vast difference in ROI.

[SgtSpike]

So your ONLY point is that people who buy later than the first adopters will get a lower return?

One person bought a SC Jalepeno, one person bought an SC Bitforce Single, both yesterday. What is your bet that both receive product on the same date? That's not an "early adopter advantage" that's a vast difference in ROI.

And? If either one of them didn't account for that possibility, they only have themselves to blame.

[Raize]

And? If either one of them didn't account for that possibility, they only have themselves to blame.

So you admit that wasn't, in fact, my only point. No purchaser of BFL products today knows which ASIC is going to be run first, that's pretty imperative for preordering without the chance for a refund.

A cursory glance at the forums seems to show that they clearly don't think they are the only ones to blame.

Let me ask you this question to show how silly your argument is: Why was the allinvain hack bad for Bitcoin?

After all, it wasn't an underlying security issue with the blockchain. Plus, if he should have stored most of his Bitcoin in an offline wallet or multiple offline wallets, right? Duh, everyone should know that by now and should have known by then, too.

If you think new Bitcoin miners "screwing themselves" over because of a lack of understanding on difficulty adjustment isn't a big deal for Bitcoin, I guess you're just an ass. I'm all for survival of the fittest here, but to not see even a need to educate people on this? That's asinine.

[SgtSpike]

And? If either one of them didn't account for that possibility, they only have themselves to blame.

So you admit that wasn't, in fact, my only point. No purchaser of BFL products today knows which ASIC is going to be run first, that's pretty imperative for preordering without the chance for a refund.

A cursory glance at the forums seems to show that they clearly don't think they are the only ones to blame.

Let me ask you this question to show how silly your argument is: Why was the allinvain hack bad for Bitcoin?

After all, it wasn't an underlying security issue with the blockchain. Plus, if he should have stored most of his Bitcoin in an offline wallet or multiple offline wallets, right? Duh, everyone should know that by now and should have known by then, too.

If you think new Bitcoin miners "screwing themselves" over because of a lack of understanding on difficulty adjustment isn't a big deal for Bitcoin, I guess you're just an ass. I'm all for survival of the fittest here, but to not see even a need to educate people on this? That's asinine.

I never said there wasn't a need to educate people about the potential for long-term ROI's on these ASIC investments.  I just think it's silly to highlight the fact that there is risk involved, when that should be common sense for just about anyone entering any sort of investment to start with.