BFL Minirig seems like a spectacular bargain?

[misterbigg]

Assuming:

1) The BFL Minirig ships
2) Network hash rate stays under 500Th/s
3) Bitcoin price remains stable
4) Actual performance of the minirig is close to expeted performance

Then, it seems to me that having a Minirig is basically having a license to print Bitcoins. It would pay for itself after a few short months, and by the time a year has passed you have thousands of extra coins. Maybe even 10,000BTC?

Is this right, or is my math off?

[bcpokey]

Assuming all those conditions, the same statement is true for any number of ASICs, with the given specs, or combination thereof, up until you control 500TH/s,

The problem is that
1) Network hashrate will not stay under 500TH/sec, if ASICs do ship, guaranteed.
2) Bitcoin price will likely not remain stable.

Otherwise, yeah, I guess it's a magical money machine.

[alexkravets]

Assuming all those conditions, the same statement is true for any number of ASICs, with the given specs, or combination thereof, up until you control 500TH/s,

The problem is that
1) Network hashrate will not stay under 500TH/sec, if ASICs do ship, guaranteed.
2) Bitcoin price will likely not remain stable.

Otherwise, yeah, I guess it's a magical money machine.

OK, we don't need to worry about 2) above because over longer periods of time volatility is gonna be factored out of the equation, unless you believe in total collapse of bitcoin to zero in which case you have no business mining at all.

As for 1) above, may I ask what are you basing your very big claim that total hash rate will NOT stay under 500 THash/sec.  Do you have a magic hash rate prediction machine or is there a list of total ASICs already ordered, paid for and about to ship whose total hash rate adds up to way above 500 THash ?

[Puppet]

As for 1) above, may I ask what are you basing your very big claim that total hash rate will NOT stay under 500 THash/sec. 

The simple fact a minirig would be so insanely profitable will guarantee hashrate will go up until it no longer is. And preorders alone may well get you there.  Your only hope is that BFL and its competition has trouble producing and delivering enough of them, *and* that you are one of the lucky few to receive yours early.

[alexkravets]

As for 1) above, may I ask what are you basing your very big claim that total hash rate will NOT stay under 500 THash/sec. 

The simple fact a minirig would be so insanely profitable will guarantee hashrate will go up until it no longer is. And preorders alone may well get you there.  Your only hope is that BFL and its competition has trouble producing and delivering enough of them, *and* that you are one of the lucky few to receive yours early.

Right ... this kind of reply is what I call "Conventional Wisdom Strikes Again".  What you say is of course true but ONLY about the whole network in aggregate ... i.e. assuming your hash power only keeps up with the average joe schmo miner then you'll experience average joe schmoe near zero returns ... however while in aggregate mining is always gonna be a very low margin business, *some* people will be a lot more profitable than the average ...

[Puppet]

Yeah, but its not the miner buying a minirig, its the miner getting them before everyone else does.

[alexkravets]

Yeah, but its not the miner buying a minirig, its the miner getting them before everyone else does.

Nope, wrong again, sorry.  The timing of getting the minirig or any other kind of ASICs equipment only matters in the very short term. 
In the long run, the miners able to increase their own hash rate at a rate faster than the average (i.e. the aggregate hash rate of the network) will collect the lion share of mining profits.

The name of the game is to stay ahead of the amateurs not to be first out of the gate.

[Puppet]

Yeah, but its not the miner buying a minirig, its the miner getting them before everyone else does.

Nope, wrong again, sorry.  The timing of getting the minirig or any other kind of ASICs equipment only matters in the very short term. 
In the long run, the miners able to increase their own hash rate at a rate faster than the average (i.e. the aggregate hash rate of the network) will collect the lion share of mining profits.

Nonsense. Whether you order 1 or 100 minirigs, the cost per GH remains the same, and the profits will actually go down a bit, as you start competing with yourself. If you order more than is good for you, you might be able negotiate a discount that BFL isnt advertising, but if that is were you are hoping to get your profit from, good luck.

OTOH,  if you buy now or 1 year from now, the cost per GH will likely be several times lower then. You may not understand ASICs, but while they are being sold as if they are made from pure gold,  the variable cost to produce them is close to ~$0 per chip.  BFL and the others will keep producing them (BFL alone have said to have 75K chips on order so far, good for ~500TH) and keep selling them at whatever price the market will bear until the price gets close to their variable cost. When that happens, you will be mining with a minirig that cost at least 10x, possibly 100x as much. Good luck earning back your investment if you havent yet by that time.

[thoughtfan]

Yeah, but its not the miner buying a minirig, its the miner getting them before everyone else does.

Nope, wrong again, sorry.  The timing of getting the minirig or any other kind of ASICs equipment only matters in the very short term.  
In the long run, the miners able to increase their own hash rate at a rate faster than the average (i.e. the aggregate hash rate of the network) will collect the lion share of mining profits.

The name of the game is to stay ahead of the amateurs not to be first out of the gate.

I believe you are mistaken on this.  You are correct when you say timing counts 'only in the very short term' but that 'only' makes a massive difference to the overall profitability of the equipment.

More importantly is this idea of 'lion's share' and 'amateurs'.  The implication is that those buying on an industrial scale will do much better than those who don't.  My point on this is per $ for anyone buying from the current ASIC offerings there is very little difference whether you're buying $1,000 or $100,000's worth of kit.  The margins and ROI is not far off the same offering virtually no economies-of-scale advantages to big spenders.  The ASIC manufacturers have very little incentive to offer big discounts for those buying bulk because each unit out of the door decreases the value of the remaining units.

Sorry to all those reading me saying this for the umpteenth time but I don't like potential 'hobbyist' ASIC miners being put off by the idea that ASICs means we'll get squeezed out by the big boys.  We may make very little money other for the few who receive early (depending ultimately on power costs) but rest assured the big spenders are in the same boat unless they have sourced cheap power.

Edit:  Just saw Puppet's post which more-or-less says the same.  No harm in repeating though

[alexkravets]

Guys,

You both make valid points.
1. It's true that cost per Ghash will be plummeting rapidly because those ASICs chips are very inexpensive to manufacture and will become less expensive over time, so YES
    a year or two from now something with 1.5Thash capacity like the minirig might cost $3000 instead of $30,000.

2. It's also true that it does not matter if you buy a mini-rig or a jalapeno b/c they have a nearly identical cost of Ghash of capacity, so mini-rig buyers do not have any inherent advantages over smaller guys buying jalapenos.

However, neither point invalidates the crux of my argument, namely: In the aggregate the total mining network will have very low profit margins because it's so ruthlessly competitive it's basically a model of pure competition, still there will be a bell-curve like distribution of *rates of growth* of hashing power between mining operations and any mining operation that can stay ahead of the average network growth hashrate by rotating depreciating equipment out and buying newer more powerful ASCIs equipment in, will end up capturing larger and larger share of total hashrate and therefore larger and larger share of all the coins being mined at that roughly constant rate (for the next 4 years).

It doesn't matter if you get a jalapeno or a minirig today, what matters is your turnaround time to get grow hash rate capacity relative to the average miner and thereby capture larger and larger share of the remaining coins.  Another hidden assumption in my argument is that you can continue running your operation WITHOUT necessarily having to cash out your coins for fiat to keep financing it, because not having to cash out your coins would allow you to ride out any volatility.  Eventually of course (one per year ?) you'd need to sell some coins to recoup the cost of previous year's investments ...

[Puppet]

Being an amateur or a pro provides zero benefits. Assuming you have the skills needed to plug it in, everyone will enjoy or suffer the same ROI, there is no benefit of scale, nothing to compete on, there is nothing you can do to get a better profitability than anyone else with the same hardware (and ultimately, same electricity cost which will become an issue again down the line). The only variable that really matters is time. Get your gear early, and you might get rich. Get them late and you will burn money.

As for continually upgrading; it doesnt matter. If a new faster/cheaper ASIC product comes out, either it will have a positive ROI or it wont. Who owns it, or how many he owns is irrelevant, its the same for everyone.

[alexkravets]

Being an amateur or a pro provides zero benefits. Assuming you have the skills needed to plug it in, everyone will enjoy or suffer the same ROI, there is no benefit of scale, nothing to compete on, there is nothing you can do to get a better profitability than anyone else with the same hardware (and ultimately, same electricity cost which will become an issue again down the line). The only variable that really matters is time. Get your gear early, and you might get rich. Get them late and you will burn money.

As for continually upgrading; it doesnt matter. If a new faster/cheaper ASIC product comes out, either it will have a positive ROI or it wont. Who owns it, or how many he owns is irrelevant, its the same for everyone.

Allow me to define my terms a little better:

Amateur is someone who buys a piece of mining equipment and sits on it watching its returns decline and his equipment become worthless.
Pro is someone who keeps upgrading his average mix of equipment to sustain a growth rate of his hashing power faster than an average "amateur" or the network as a whole ( which is pretty much the same thing )

Now, given the above definitions, is it not the case that the pro will eventually be taking a larger share of the "fixed" pie ?

Yes it's true that for any given equipment purchase there is no advantage ... but it's the growth rate of hash power that determines the outcome not a particular snapshot in time

[Puppet]

Pro is someone who keeps upgrading his average mix of equipment to sustain a growth rate of his hashing power faster than an average "amateur" or the network as a whole ( which is pretty much the same thing )

So you mean the pro is the one that keeps throwing good money after bad? If the investment proves unprofitable for either miner, buying more isnt going to solve it. Growth rate is meaningless.  If you try to grow your hashrate along with the network, so will your profits or losses, no different than the "amateur". Hashrate is fungible, costs and profits scale linearly. Well, actually, only costs scale linearly, as profits scale  less than linearly if you are talking large enough numbers.

If you are referring to BFL's upgrade program, dont count on that ever existing for ASICs, and even if it does, it will be like now where you have to double down your investment to get the discount.  The only reason BFL and its competitors offered upgrades is because otherwise no sane person would have kept buying FPGAs after they announced the ASICs. Regardless, betting your business on someone offering you a freebee that they never promised is rather silly.

[alexkravets]

If you try to grow your hashrate along with the network, so will your profits or losses, no different than the "amateur". Hashrate is fungible, costs and profits scale linearly. Well, actually, only costs scale linearly, as profits scale  less than linearly if you are talking large enough numbers.

This is true, but ONLY if you measure your profit or loss in USD.  If BTC/USD rate spikes, you simply purchase new equipment with previously mined coins, if BTC/USD price collapses, so will the network hash rate (as it did during the bubble pop) and then you'll capture more BTCs as your share of the hash rate will increase. 

If you have to keep selling BTCs to pay rent, electricity or to buy more mining equipment, then you cannot play the pro version of this game.

[meowmeowbrowncow]

Pro is someone who keeps upgrading his average mix of equipment to sustain a growth rate of his hashing power faster than an average "amateur" or the network as a whole ( which is pretty much the same thing )

So you mean the pro is the one that keeps throwing good money after bad? If the investment proves unprofitable for either miner, buying more isnt going to solve it. Growth rate is meaningless.  If you try to grow your hashrate along with the network, so will your profits or losses, no different than the "amateur". Hashrate is fungible, costs and profits scale linearly. Well, actually, only costs scale linearly, as profits scale  less than linearly if you are talking large enough numbers.

If you are referring to BFL's upgrade program, dont count on that ever existing for ASICs, and even if it does, it will be like now where you have to double down your investment to get the discount.  The only reason BFL and its competitors offered upgrades is because otherwise no sane person would have kept buying FPGAs after they announced the ASICs. Regardless, betting your business on someone offering you a freebee that they never promised is rather silly.

It's a 'pro' strategy in gambling.  So... FWIW.

[Puppet]

f BTC/USD price collapses, so will the network hash rate (as it did during the bubble pop)

Wrong. People shut down $150 GPUs that cost $15 per month in electricity.
Who is going to shut down a $30.000 minirig that costs $100 to run?
No one. So dont expect hash rate to scale down with BTC price until the main cost of asics is electricity again and network hashrate is measured in petahash.

[Beepbop]

Also, don't forget opportunity cost. While it's sort of neat to have "a computer that pays for itself, including upgrades", if there's anything else that has a higher ROI (and perhaps lower risk) it's better to treat mining as a gamble and just divest yourself once electricity cost catches up with payout. If it suddenly becomes more profitable again, you can just buy a whole new rig with fiat currency. Having an existing unprofitable mining rig and a hoard of BTC is not going to give you a leg up. Even if the vendor takes payment in BTC you'd just be better served buying those BTC on the open market.
And if you're sure that BTC will go up, just buy them and speculate with them. No need for a mining rig.

Buying ASICs for mining is a bet that you'll get yours before most others do, so you can earn it back fast, and after that you're doing a public service by securing the network and handling transactions. It's an arms race that's mainly pointless except as to make sure that no individual group controls 51% of the network. And fills up the order books of manufacturing facilitities like some kind of New Deal project

[fnordfnordfnord]

...  (depending ultimately on power costs) but rest assured the big spenders are in the same boat unless they have sourced cheap power.

Edit:  Just saw Puppet's post which more-or-less says the same.  No harm in repeating though

ASICs (Avalon) are something like 100x more power efficient than video card rigs.

[thoughtfan]

...  (depending ultimately on power costs) but rest assured the big spenders are in the same boat unless they have sourced cheap power.

Edit:  Just saw Puppet's post which more-or-less says the same.  No harm in repeating though

ASICs (Avalon) are something like 100x more power efficient than video card rigs.

And right now this means power efficiency of the various ASIC options and variations in power prices counts for very little.  But for how long?  I was referring to a time after the initial mayhem has settled.

[Keninishna]

Something to think about is that the network generates 25 btc every 10 min more or less depending on the difficulty. When you use a newer more efficient tech you are essentially taking profits from those in the network who are using the lesser tech. Mining is a competition not just a race for finite resource.