New Miner, should I go ASIC

[Taihen]

ive known about bitcoin for a while but up until lately treated it as a hobby on the backburner running cpu mining on a old pc
ive just recently heard about ASICs and have more money than sense for now if i could get in on it ASAP(not on pre order)

what hardware can i get in hand now? i realize being late to the party means a higher barrier of entry so what if im willing to put in $5000-$10000 USD into it? (that said, if $15000 made the a significant difference i might be convinced to go that high)
or have i completely missed the boat and even a high buy-in wont help?


on a side note Arvicco has a Batch 1 Avalon which aparently is capable of 71,000 MHash/s w/ Power consumption around 660Wt
currently its at 280BTC
should i bid on it, and start with that? its one of the few auctions i can see for actual hardware that people have

[Giant]

As far as I've heard, everyone expects ASIC to be used soon, rendering GPU's useless, but ASIC can still only be preordered. Haven't heard of anyone actually using one yet.
So yes, you should go ASIC, but it's hardly available.

[J35st3r]

280BTC is way too much for an Avalon, you'll never make a ROI on that.

Take a look at this thread, lots of good info on expected ROI.  https://bitcointalk.org/index.php?topic=221318.0;all

[MagicBit15]

As far as I've heard, everyone expects ASIC to be used soon, rendering GPU's useless, but ASIC can still only be preordered. Haven't heard of anyone actually using one yet.
So yes, you should go ASIC, but it's hardly available.

Info Incorrect ^^ . ASIC is available, not just preordered. It is new technology being developed and In Use this year and there is a few companys producing units that ARE incredible. You can check out the list here.

https://en.bitcoin.it/wiki/Mining_hardware_comparison#ASIC  Protip:The ones that say Shipping Yes are used there being used as we speak.

[MagicBit15]

280BTC is way too much for an Avalon, you'll never make a ROI on that.

Take a look at this thread, lots of good info on expected ROI.  https://bitcointalk.org/index.php?topic=221318.0;all

Your not taking into account inflation in BTC price, you will absolutely make a Significant ROI. Honestly 1 or 2 years especially when reward halves a second time, you might be in unbeleiveable shape..

Newbies giving advice to Newbies. Never works out I swear. Good Luck out there OP!!

[J35st3r]

280BTC is way too much for an Avalon, you'll never make a ROI on that.

Take a look at this thread, lots of good info on expected ROI.  https://bitcointalk.org/index.php?topic=221318.0;all

Your not taking into account inflation in BTC price, you will absolutely make a Significant ROI. Honestly 1 or 2 years especially when reward halves a second time, you might be in unbeleiveable shape..

Newbies giving advice to Newbies. Never works out I swear. Good Luck out there OP!!

Irrelevant, you're buying it with BTC, so if it does not return your investment in BTC then the exchange rate matters not one whit. If you spend USD to buy it, then you're effectively investing its equivalent value in BTC at the time of purchase. And again if it never returns your investment, you lose.

On your previous post, delivery time is of the essence in calculating ROI. Do your math carefully.

[Flashman]

IMO the miners being sold at sane prices, you're not going to see until September ordering now, those are around $50 a gigahash or less. In the 3 months from right this second until then, they would have paid themselves off about twice over. So it would just be worth paying $100 a gigahash to get in the game earlier than that, but make sure that does really get you in early, pay $100 a gigahash and not get your miner for a month and it was only really worth paying $75 for...

[maxmint]

Check out www.minr.info to compare current ASICs.

[OnkelPaul]

Selfish plug: http://coinish.com/calc is a calculator that takes projected difficulty development and time to delivery into account, you might want to run your options through it (disclosure: I wrote the calc routines, if you find bugs you may keep them but please tell me about them).

The bitcoin exchange rate can make a difference in the ROI calculation in fiat money, so even mining with a graphics card might yield a good return in dollars once BTC skyrockets.
But if the ROI in BTC is negative, you'd be better off to just stay with the BTC in the first place and wait for its value to increase.

Onkel Paul

[J35st3r]

Selfish plug: http://coinish.com/calc is a calculator that takes projected difficulty development and time to delivery into account, you might want to run your options through it (disclosure: I wrote the calc routines, if you find bugs you may keep them but please tell me about them).
Onkel Paul

Cool calculator. I was waiting for someone to call me out on my math, but this is just perfect. Use the expert mode to see the return taking into account difficulty increase. One possible request is to enter the increase per fortnight rather than per day (saves having to use a calculator to calculate the 14'th root of the factor).

[SneakyP]

I've been trying to do the math on whether I should wait or just pay the high prices now for the few ASICs available. 

[J35st3r]

Running the numbers for a batch 1 Avalon through http://coinish.com/calc (NB use expert mode), gives a maximum price of 209BTC before your expected breakeven goes to never.

 Of course this assumes the rate of change of difficulty remains constant, which it won't.

If I plug in my 26% per fortnight (14'th root gives 1.6645% per day) that I used on https://bitcointalk.org/index.php?topic=221318.0;all  its drops to 165BTC, so I was being optimistic (my prediction was for a 250BTC to break even at 26% per fortnight, it clearly won't, it actually needs 1.096% per day or 16.5% per fortnight).

Nice toy (http://coinish.com/calc I mean, but an Avalon would be nice too, but not at that price!)

[MagicBit15]

280BTC is way too much for an Avalon, you'll never make a ROI on that.

Take a look at this thread, lots of good info on expected ROI.  https://bitcointalk.org/index.php?topic=221318.0;all

Your not taking into account inflation in BTC price, you will absolutely make a Significant ROI. Honestly 1 or 2 years especially when reward halves a second time, you might be in unbeleiveable shape..

Newbies giving advice to Newbies. Never works out I swear. Good Luck out there OP!!

Irrelevant, you're buying it with BTC, so if it does not return your investment in BTC then the exchange rate matters not one whit. If you spend USD to buy it, then you're effectively investing its equivalent value in BTC at the time of purchase. And again if it never returns your investment, you lose.

On your previous post, delivery time is of the essence in calculating ROI. Do your math carefully.

You might be the most statistically incompetent person I have ever ran into, in all my days on this forum. Rather than being a tyrannical  A-hole, I will simply prove you wrong *cough*destroy*cough* you will math, ration, and logical thought.


ASSUMING!! Inflation is irrelevant, which it isn't since the entire point of Bitcoin is to create a new, anonymous, decentralized form of CURRENCY. Which at some point is CONVERTED or TRADED which means it holds VALUE!! It is not in any way shape or form irrelevant. No Value, No Worth.

But just to appease your retardery, we will assume it is irrelevant. My god.....

An Avalon selling at 280 btc, which is roughly 33k USD, is RIGHT NOW is considered overpriced. I do agree with this. 20-21k is a much more modest and sought price range at this moment in time. Agreed.

But to say it is infeasible, and you will not have any ROI (Assuming solely BTC), is the most absurd state of nonsense I have EVER read here. It is the most sought after piece/pieces of hardware right now and will CONTINUE to be so for months to come, and that's it!! Well sort of, but again we are assuming value to be irrelevant so I digress. But to keep things simple we will leave it at a single unit, bought right this second, received in the near immediate future, such as the aforementioned auction.

Note: If you can find one and afford it, please, Mine, and Mine F*cking hard right now.
 

Running the numbers for a batch 1 Avalon through http://coinish.com/calc (NB use expert mode), gives a maximum price of 209BTC before your expected breakeven goes to never.

 Of course this assumes the rate of change of difficulty remains constant, which it won't.

If I plug in my 26% per fortnight (14'th root gives 1.6645% per day) that I used on https://bitcointalk.org/index.php?topic=221318.0;all  its drops to 165BTC, so I was being optimistic (my prediction was for a 250BTC to break even at 26% per fortnight, it clearly won't, it actually needs 1.096% per day or 16.5% per fortnight).

Nice toy (http://coinish.com/calc I mean, but an Avalon would be nice too, but not at that price!)

1.) Expert mode is interesting, but that is really all it is, interesting. (Great Calculator BTW, Somehow I have missed it all this time!!)

2.) Secondly your contradicting yourself

Of course this assumes the rate of change of difficulty remains constant, which it won't.

Your right it won't, it hasn't, and it defies Moore's law. So, your wrong just to begin with and it is an important thing to assume it as a constant rate to affirm your next projection. Which you don't which is why your entire argument here and in the other thread is invalid.

3.)

If I plug in my 26% per fortnight

Just stop. Your using a sample size of 14 days. Specifically the last 14 days  .  Just like the kid in the other thread tried to tell you, don't try to estimate nethash, Ever. You can't do it. Assumptions get you No Where!! What you can do is look at data historically and use stats and game theory to project something and make a prediction, and that is about it. Which is why the stock market is still around (Somehow) to this day. Statistics are your best friend!! Your sample size is so small you need to look at the big picture. https://docs.google.com/spreadsheet/ccc?key=0AmcTCtjBoRWUdHVRMHpqWUJValI1RlZiaEtCT1RrQmc#gid=0 and for a lovely network overview. http://bitcoin.sipa.be/

4.) Fact is the difficulty is being destroyed right now, more than it ever has been. The big variable your miss in all your data; network changes are constant and fluctuate GREATLY. 38% in 10 days this past March?! People were flipping out remember? No you don't cause your a noob. I also don't know if you have another username, but I could care less cause that would be even more embarrassing for yourself, so I would probably not mention so.

5.) Last but not least, when it comes down to it, after a 12 month cycle using a single Avalon Unit, at a 71000 mHash, the question is will it be above or below 280BTC, such as the sellers auction was. Will you have a positive ROI. I have my bets the buyer will have had made modest ROI after a 12 month cycle in BTC. Alone. Again, my profitability decline is much more realistic, and statistically held, versus hearsay and bad short terms assumptions. But no one is perfect.

So part one on why your statistical Imbecile is complete. On to a much shorter part 2....

[MagicBit15]

Part 2. Why inflation and price is ridiculously important. And why you are even more of a looney for saying it's irrelevant. (Still in Awe   )

How is one getting the BTC? Is it already in possession, buying, trading etc., This is important not only for obvious reasons, but to determine whether your should hold on the btc or spend it in buying a rig. (Note: get offered one in decent USD, go for it!!) Also you are so into such tiny sample sizes and ROI stats, I think you might as well stick to roulette. Hey you might even get lucky in the short-term!!

Value is important in deciding keeping BTC? (Riding the USD wave) Or mining (Giving it up (Substantially) now to get more later). Which is more important? Which do you think is a better decision? VALUE holds the primary foundation for all of your decisions in all these areas, and inflation is the key in making your final choice.

To sum up everything. The guys price was a little high, but you would be going to constantly make Bitcoins from it in the long run. In a year much less than now, specifically 3 years. But in that time, as price increases, your "risky investment slowly becomes a good one, and all in all everyone is a winner. He gets a short term dump, you get a long term haul. 280BTC for that, at this point of time was not an absurd or ridiculous price. High, but not outlandish.

[J35st3r]

Lovely rant, and the ad hom's were priceless. Can I be your friend, youll be handy in tough places

But seriously, I was rather hoping someone would call me out on my math to see where I may have gone wrong. I especially liked OnkelPaul's calculator as is showed I was actually overestimating the ROI (it conformed my existing bias, so cool by me).

To the point, I agree that nethash prediction is the key. And this was the point I was making, that your ROI must take into account changes in difficulty, which are insanely difficult to predict. The 26% per fortnight that I quoted was actually a back-prediction from the 250BTC price to indicate the maximum rate of difficulty increase that would give a positive ROI. It was actually meant as an illustration of the sort of math you'd need to do to justify that investment (and incidentally was wildly over-optimistic as Paul's calculator shows).

I won't concede your point on inflation (or deflation as is the desired outcome for BTC). Here's why. Mining is like a cookie jar. On day one you buy a magic cookie jar. It cost you 100 cookies in the first place (or you may have got your mom to buy it for you using her dollar credit/debit card), but its worth it because its magic. But you can only take out a few cookies cookies every day (that's mining for you). The first day you get to take 16 cookies, yum. But the second day you can only take 8, and on the third you get 4 cookies. Now how many cookies will you eat before you die of old age (allowing fractions of course)? Did you get your 100 cookies back? That's the nub of the difficulty increase argument.

Now your point is that those cookies are rare, rare treats that nobody can bake any more, the recipe is lost. After a few years they are so valuable that you can command almost any price. Luckily your magic cookie jar is still churning out a few crumbs of dust every day, so you're rich. But you'd be so, so much richer if you'd never bought that damn jar in the first place and just kept your 100 cookies in a nice safe cupboard.

Of course if your magic cookie jar had let you take only eight cookies per day, but halved it every week rather than every day, you'd be in profit, but only just. So can you predict what your jar will allow you to take before handing over your cookies?

That was fun, can we do it again? 

[OnkelPaul]

1.) Expert mode is interesting, but that is really all it is, interesting. (Great Calculator BTW, Somehow I have missed it all this time!!)

Thx 
No wonder you missed it all the time, because it just wasn't there all the time, we just implemented it a few days ago.

Your observation about expert mode is correct - the calculation can't be exact because it simplifies (uses a continuous exponential function instead of the 2016-block stepwise difficulty correction) and of course it can only extrapolate from existing data. The assumption that difficulty growth will be exponential will of course be moot as soon as there is some saturation. All exponential approximations of real-world systems break down when they approach saturation.
The idea of the expert mode is to get a rough feeling how the increase in difficulty will affect future mining results.

For back-of-the-envelope calculations assuming an exponential reward curve, you should find the half-life of your mining reward (time until the reward halves due to difficult increases, not the 4-year block reward halving built into bitcoin). Half of the total reward will be gained before that point in time, and the other half will take forever after that.
At a daily difficulty increase of 1.3 percent, the half-life is about 54 days. So if your mining rig does not yield significantly more than half of its investment cost during those 54 days, it never will pay for itself.

And since mining produces bitcoins, profitability calculations should primarily be done in bitcoins. If the total lifetime yield of your mining rig is fewer bitcoins than what it took to buy and operate it, it's a loss even if fiat exchange rate increases by a factor 10000, because you would have been better off just keeping the bitcoins in your wallet.

Onkel Paul

[Spekko]

Lovely rant, and the ad hom's were priceless. Can I be your friend, youll be handy in tough places

But seriously, I was rather hoping someone would call me out on my math to see where I may have gone wrong. I especially liked OnkelPaul's calculator as is showed I was actually overestimating the ROI (it conformed my existing bias, so cool by me).

To the point, I agree that nethash prediction is the key. And this was the point I was making, that your ROI must take into account changes in difficulty, which are insanely difficult to predict. The 26% per fortnight that I quoted was actually a back-prediction from the 250BTC price to indicate the maximum rate of difficulty increase that would give a positive ROI. It was actually meant as an illustration of the sort of math you'd need to do to justify that investment (and incidentally was wildly over-optimistic as Paul's calculator shows).

I won't concede your point on inflation (or deflation as is the desired outcome for BTC). Here's why. Mining is like a cookie jar. On day one you buy a magic cookie jar. It cost you 100 cookies in the first place (or you may have got your mom to buy it for you using her dollar credit/debit card), but its worth it because its magic. But you can only take out a few cookies cookies every day (that's mining for you). The first day you get to take 16 cookies, yum. But the second day you can only take 8, and on the third you get 4 cookies. Now how many cookies will you eat before you die of old age (allowing fractions of course)? Did you get your 100 cookies back? That's the nub of the difficulty increase argument.

Now your point is that those cookies are rare, rare treats that nobody can bake any more, the recipe is lost. After a few years they are so valuable that you can command almost any price. Luckily your magic cookie jar is still churning out a few crumbs of dust every day, so you're rich. But you'd be so, so much richer if you'd never bought that damn jar in the first place and just kept your 100 cookies in a nice safe cupboard.

Of course if your magic cookie jar had let you take only eight cookies per day, but halved it every week rather than every day, you'd be in profit, but only just. So can you predict what your jar will allow you to take before handing over your cookies?

That was fun, can we do it again? 

haha best post ever...

i must agree with J35st3r tho, a machine now for 250btc will probably not make an ROI in YEARS, seeing as more and more asics comming in,, asicminer growing + 100TH mine possibly comming into play soon.. and when 100TH comes in it will be battles between more and more hashrates of ~100TH comming in to play everytime, ,

just my opionion though, but OP seems to be badass at predicting the future,,, if i had 250btc i would sell half or a quarter at whatever peak i see in the next few months and put the rest in a safe

[p0peji]

I wouldnt do it, you are already late to the game and therefore more likely to lose your money

[lllama]

if you can afford 1, ASIC fo sho

[Lohoris]

what hardware can i get in hand now?

You also have the option of buying ASICMINER shares: they mine for you and give you dividends: much simpler and scalable.