Topic: Inaba's request - page 3. (Read 8511 times)

Don't forget that this is not 5% a month perpetually -- it will diminish as the difficulty increases. How fast is anybody's guess.

Indeed, I would be very happy with 5% a month. My bank gives me less than 5% a year!

People in the Bitcoin community are very short sighted when it comes to returns.  1% per day sounds reasonable...     5% per month sounds low?   

If we could get 5% per month, we'd all be rich very, very soon.

Red bold emphasis is mine.  

Q: Where in the "real world" will you find MONTHLY returns like that?
A: No where.

This is why I feel investing in bitcoin is a great investment, beyond all the other benefits that bitcoin will provide.

Thanks for your time, Keefe! 
Your explanations speak for themselves! Specifically for the European market! 

Alright, one more scenario, more extreme:

Let's assume $12/BTC, $0.25/kwh power cost, and 1000 TH network hashrate after reward halving.

Hypothetical device that does 60 GH, costs $1300, and uses 60W:
$100K buys about 4600 GH of hardware which will use about 4.6kW of power and earn about $5980 per month, minus about $830 for power, for 5.15% monthly ROI.

Hypothetical device that does 54 GH, costs $1070, and uses 405W:
$100K buys about 5100 GH of hardware which will use about 38kW of power and earn about $6630 per month, minus about $6885 for power, for -0.26% monthly ROI. Uh oh.

Hypothetical device that does 54 GH, costs $1070, and uses 120W:
$100K buys about 5100 GH of hardware which will use about 11kW of power and earn about $6630 per month, minus about $2040 for power, for 4.59% monthly ROI.

But someone with $100K to invest, ought to find a better place to setup than where power costs $0.25/kwh.

And if ~5% monthly ROI were attractive to professional miners, why has mining been historically much more profitable than that, except when the exchange rate fell toward $2? I think 10-20% monthly ROI will continue for the next year or so at least. Though a price war among mining devices could really screw ROI up.

The guy with $0.25/kwh power can quit mining and let the professional miner in the right location make up the slack.

and now please with $0.25/kwh 

I think your latest post in regards to power and ROI is much more clear about it to the point where it would take a real idiot or shill to misinterpret. I swear though, I feel like I need to record any conversation I use conditional statements in. Some people have a real hard time processing them.

I have no pre-orders for any ASICs from any company, I just have some sympathy for someone dealing with FUD-like issues.

I apologize for not making it more clear that the wattage figure was purely hypothetical and came from a competitor.

So let's put some figures together...

Let's a say a professional miner has $100K to invest in mining hardware and is in a good area where power costs only $0.06/kwh.

In the GPU era, with 23 TH network hashrate and 50 BTC block reward, he can buy about 190 GH of GPUs which will use about 63kW of power and earn about $21.5K per month, minus about $2700 for power, for 18.8% monthly ROI.

In the ASIC era, with 280 TH network hashrate and 25 BTC block reward, he can buy about 4600 GH of BFL ASICs which will use about 4.6kW of power and earn about $21.4K per month, minus about $200 for power, for 21.2% monthly ROI.

$200/mo for powering BFL ASICs is quite an improvement over $2700/mo for GPUs.

Now let's imagine a 54 GH bASIC were to use 405W (which I don't believe will be the case):
He can buy about 5100 GH of bASICs which will use about 38kW of power and earn about $23.7K per month, minus about $1600 for power, for 22.1% monthly ROI. Hey, we're ahead of BFL still! But putting that point aside, we would only be twice as power efficient as GPUs

Now let's try a more realistic scenario, where a 54 GH bASIC might use 120W (this is my personal wild guess):
He can buy about 5100 GH of bASICs which will use about 11kW of power and earn about $23.7K per month, minus about $480 for power, for 23.2% monthly ROI. Even better ROI. And now we've cut our power bill to 1/6 of GPUs, very nice.

DISCLAIMER: We really don't know yet how much power the bASICs will use. The wattage figures for bASIC above are purely hypothetical and probably wildly wrong.

Don't underestimate people's ability to ignore the words "if" "possibly" "perhaps" "supposedly" and such. I think Cablepair has a legitimate worry of his ASIC product being unfairly associated with a made up power number.

Also interesting how a standalone post can go off topic so quickly, I believe this thread was about demonstrating Inaba/BFLJosh's identities are not well firewalled and citing possible spurious statements of his.

On topic: Has BFL provided their operating address yet? Inaba/BFLJosh indicated he would be able to do so "soon" as of September 26:

Total network hash power right now is around 23 TH. That's like ~$12M of GPUs, earning ~$2.6M per month (at $12/BTC) minus power cost. When the block reward is halved, $6M of GPUs earning $1.3M per month would have the same ROI. When $1300 buys a 60 GH unit, $6M will buy ~280 TH, and will have a similar ROI. So when looking at power draw of this new hardware, you should account for going from 23 TH to 280 TH. You should adjust a unit's hashrate by 12x to get a fair comparison.

I believe that the predominant factor determining how much hardware is mining is ROI. And I believe that it's decided at the margin, by those with cheap power and lots of capital to work with. Those people can buy and setup additional GPUs and FPGAs as the exchange rate goes up and difficulty lags, until price/difficulty is back to where they are comfortable with the ROI.

The cost of power, especially for FPGAs and ASICs, isn't that significant compared to the hardware cost. If you spend $1000 on a unit, and only $50 on power for a year, and you earn $2600 in a year, power doesn't make a big % impact on ROI.

I don't think botnets and casual miners affect the equation. If there is 2 TH of botnets, that just means the professional miners will likely run 2 TH less hardware, because they're looking at ROI. And if some people decide to stop buying more hardware, as long as there are a few professional miners who will take up the slack it won't matter.

Now I do agree with at least one of Gmaxwell's points: GPUs have value outside mining, which can make a big difference in the risk calculation. So maybe professional miners will want a better ROI to compensate. Also, ASICs will likely drop in price (per GH) alot over the next year, as companies like BFL and BTCFPGA fight for market share with products with huge marginal profit (after NRE). Those two factors could make a big difference in where price/difficulty stabilizes.

Wow, I didn't expect Cablepair to blow up at me of all people. I've posted in the past on Cablepair's side against Inaba's claims that the bASIC won't be competitive. Are you upset about my latest post just because I referenced Inaba's 405W figure? I made sure to say "IF" a bASIC uses 405W. As Gmaxwell said, I am making a logical argument, despite oversimplifying. FWIW, I've ordered 4 of the bASIC 54GH units, so Cablepair currently holds ~$4400 of my money. I feel a little insulted that I get called a troll in return.

Edited to protect the innocent

I'm just glad to see it put in its own special thread rather strewn through other threads that were really meant for other things.

Thanks for that, OP.

Some interesting information posted here. But still this is the hardware forum...

While I don't agree with Keefe, he's actually making a coherent argument here.

First, as you know— the earnings from mining don't depend on your absolute hashrate. 1MH/s, 1GH/s, 1TH/s whatever. What matters is your rate relative to the other miners. Keep this in mind.

If you assume that all mining is done by initial purchase cost limited miners who bought cards just for mining, and e.g. that as many people who spent $1070 on GPUs will spend $1070 on ASICs  ...  then eventually your $1070 will have the earnings equivalent of what $1070 in GPUs has today, but a higher operating cost because it will draw more power than $1070 in GPUs. (well maybe, but thats the argument).

Of course this is all crazy and wrong, because (1) there is some unknown but non zero amount of botnet miners who are going to become insignificant, (2) there is some unknown but non zero amount of casual miners using their primary GPUs who are going to become insignificant, (3) OPERATING COST IS A MAJOR DRIVING FACTOR IN RUNNING A GPU FARM, it's probably best to assume that large miners would match their operating costs, not their initial costs (4) some miners have already spent their money on GPUs and won't continue anymore, (5) some miners felt more comfortable investing in GPUs which have some resale value if bitcoin fails. (6) People holding out on ASICs until they are for sale for immediate delivery or for second-generation products or for the effect of competion to lower prices. etc.  There are also reasons we could expect things to be less profitable: people holding out on GPU purchases in anticipation of asics, horded gpu earnings being released to buy asics, big asic farms should be dollar per dollars much easier to manage (GPU FAILURES SUCK) encouraging bigger ones. (not to mention the subsidy decreases!)

The model is all wrong, but it's not crazy or abusive.

Please don't let the people here stress you out.  The proof is in the pudding. Your product will amaze people. Spend your limited mental cycles on making it the best you can. If all goes well, or if all blows up.. it will be what it will be and none of these arguments matter.

In a few months, Your 54GH/ s seems to be the same as 800 MH/s today! Thus, the comparison W/$ on the purchase of an ASIC seems already correct!
So everything is over 150W/1000$ seems for the European market not so great anymore!
Just my impression!