:

[chodpaba]

:

[marcus_of_augustus]

Okay, I have a few questions;

1) your graph seems to suggest that by Jan 2013 (mining block reward halving date) BTC will be trading for $1500 to $2000 each, is that correct?

2) does the exponential growth rate in BTC/U$D have the same exponential power as the current rate of increase or something different?

[Cryptoman]

The blue markers represent the Bitcoin block bounty going to every mining rig currently running and what happens to that reward as new capacity is added to the network.

Shouldn't the total network bounty remain the same, only divided into smaller pieces?

One caveat, I haven't even touched fees with this. Frankly, I don't quite know how to deal with fees, but one possibility stands out. It may be better to buy Bitcoins now instead of investing in rigs... Then take a portion of the fortunes you earned through Bitcoin appreciation to get back into mining in earnest once things start to turn around mid 2013...

I think the whole fees thing will work itself out to keep the most efficient miners marginally profitable.  The two big questions in my mind, however, are 1) what kind of discontinuity will happen at block 210000, and 2) will the hash rate always remain above what is necessary to prevent hostile takeover of the network?

I'm more interested in seeing Bitcoin succeed as an alternative currency than I am in making obscene profits, so I will continue to mine and then sell those coins at the prevailing market rate rather than hoard them.

[cypherdoc]

The blue markers represent the Bitcoin block bounty going to every mining rig currently running and what happens to that reward as new capacity is added to the network.

Shouldn't the total network bounty remain the same, only divided into smaller pieces?

One caveat, I haven't even touched fees with this. Frankly, I don't quite know how to deal with fees, but one possibility stands out. It may be better to buy Bitcoins now instead of investing in rigs... Then take a portion of the fortunes you earned through Bitcoin appreciation to get back into mining in earnest once things start to turn around mid 2013...

I think the whole fees thing will work itself out to keep the most efficient miners marginally profitable.  The two big questions in my mind, however, are 1) what kind of discontinuity will happen at block 210000, and 2) will the hash rate always remain above what is necessary to prevent hostile takeover of the network?

I'm more interested in seeing Bitcoin succeed as an alternative currency than I am in making obscene profits, so I will continue to mine and then sell those coins at the prevailing market rate rather than hoard them.

Yes, the total network bounty remains the same. The graph only shows the portion of that bounty earned by mining nodes currently running. It is intended to show how that portion declines as a result of network growth.

I would rather characterize what happens at block 210000 as a disturbance rather than a discontinuity. What I expect to happen is an ultimate shakeout of less profitable miners. This is not necessarily a bad thing, I suspect it may result in a seachange form GPU mining to FPGA/ASIC mining and a re-adjustment of the market. As far as network security goes—this example, taken to it's extreme, shows a network hashing power of about 3,000 times the current capacity by January 2013.

The important question I was really trying to answer though is, what happens to the relationship between hashing power and price once the block bounty is halved? That answer seems to be that price begins to overtake network growth until a new equilibrium is reached. This is not a bad thing as far as long-term network security goes... Relative growth may stall temporarily as the market re-adjusts, but as soon as people start to believe in a new, faster rate of $/BTC growth (mid 2013?) then a new mining gold-rush will take off...

how in the world can u assume that rate of increase in $/BTC?

[SgtSpike]

You're right.

Everyone should stop mining now.  I shall continue.

[FreeMoney]

Mining is like a hedged position imo. If you buy coins and they go to nothing you have nothing, but if you buy equipment you can get most of your investment back. If they go to 10% current value and you bought coins that sucks, but if you bought equipment then when difficulty drops you will get 10x more coins (not exactly, I know there is lag and other factors) to compensate.

Really mining and immediately selling the coins isn't even a bet on Bitcoin, it's profiting (or trying to) from your extra efficiency over the average miner.

[bitcoinBull]

I don't understand the blue markers (bounty from mining).  They are priced in dollars like the pink markers (dollar price of coins), but they dont seem to factor in the price increase showed by the pink markers.  If they did, these blue markers would stay flat across.

EDIT: s/yellow/pink

EDIT2:  I se now.  The blue markers represent network difficulty.  Yellow marker is blue multiplied by price (pink) and is almost flat.

[JJG]

Mining is like a hedged position imo. If you buy coins and they go to nothing you have nothing, but if you buy equipment you can get most of your investment back. If they go to 10% current value and you bought coins that sucks, but if you bought equipment then when difficulty drops you will get 10x more coins (not exactly, I know there is lag and other factors) to compensate.

Really mining and immediately selling the coins isn't even a bet on Bitcoin, it's profiting (or trying to) from your extra efficiency over the average miner.

You're not looking at the whole picture.

When you buy a mining rig, you also have to pay for electricity, cooling, maintenance as parts fail, etc. Your hardware loses value rapidly as the next generation of gaming cards and supporting hardware comes out, and especially as your warranty expires.

Try this: Start with an imaginary balance of zero. Subtract the cost of your mining hardware. Add back the resale value of your hardware in one year (Usually 60% of what you paid if you get a good deal). Subtract the cost of 1 year of electricity to run that hardware 24/7.

For many rigs and locations, this value will be small, zero, or often negative.

Now take the value of bitcoins you expect to earn. This is a bit tougher, but the important part here is to be honest with yourself about difficulty increases. 30% difficulty increases every 10 to 11 days will follow the current trend. This means in 2 months, you will only be making 20% of the bitcoins you're making today. A year from now, about 1% of what you're making today.

Obviously, mining isn't going to be seriously profitable for very long unless the value of BTC continues to go up with difficulty. So now you're betting on BTC going up indefinitely. But if you're really going to bet on BTC going up indefinitely, why wouldn't you just buy some BTC right now?

Personally, I wouldn't invest anything significant in BTC unless you like to gamble. That includes buying mining hardware. Don't underestimate the number of cash-strapped college students who already have gaming GPUs and free electricity.

[AaronM]

If no one is willing to part with their Bitcoins (everyone holds on to them), then Bitcoin is definitely a bubble and the price is only staying high because of new users coming in and speculators buying bitcoins with local currencies.

If you want Bitcoin to have a future, buy real goods and services (not cash, not Ponzi schemes or lotteries) with Bitcoins.

[AaronM]

For continuous deflation, like is expected with Bitcoin, it would be a nice feature to be able to multiply the prices of all goods/services you have for sale by a number of your choosing (probably between 0.9 and 1.1). Cross-site is ideal but hard to implement, but it's still very useful on an auction site like BiddingPond.com.

The cure for rapid price changes is low-latency BTC value information and agile prices.

[SgtSpike]

Mining is like a hedged position imo. If you buy coins and they go to nothing you have nothing, but if you buy equipment you can get most of your investment back. If they go to 10% current value and you bought coins that sucks, but if you bought equipment then when difficulty drops you will get 10x more coins (not exactly, I know there is lag and other factors) to compensate.

Really mining and immediately selling the coins isn't even a bet on Bitcoin, it's profiting (or trying to) from your extra efficiency over the average miner.

You're not looking at the whole picture.

When you buy a mining rig, you also have to pay for electricity, cooling, maintenance as parts fail, etc. Your hardware loses value rapidly as the next generation of gaming cards and supporting hardware comes out, and especially as your warranty expires.

Try this: Start with an imaginary balance of zero. Subtract the cost of your mining hardware. Add back the resale value of your hardware in one year (Usually 60% of what you paid if you get a good deal). Subtract the cost of 1 year of electricity to run that hardware 24/7.

For many rigs and locations, this value will be small, zero, or often negative.

Now take the value of bitcoins you expect to earn. This is a bit tougher, but the important part here is to be honest with yourself about difficulty increases. 30% difficulty increases every 10 to 11 days will follow the current trend. This means in 2 months, you will only be making 20% of the bitcoins you're making today. A year from now, about 1% of what you're making today.

Obviously, mining isn't going to be seriously profitable for very long unless the value of BTC continues to go up with difficulty. So now you're betting on BTC going up indefinitely. But if you're really going to bet on BTC going up indefinitely, why wouldn't you just buy some BTC right now?

Personally, I wouldn't invest anything significant in BTC unless you like to gamble. That includes buying mining hardware. Don't underestimate the number of cash-strapped college students who already have gaming GPUs and free electricity.

I don't understand where people are coming up with their calculations on this, where you actually NEED to calculate it out for a year or better.

I bought 3 5850's.  At current market rates, and assuming MH/s of 300 for each one, they'll pay for themselves in about 17 days.  This is, of course, not taking in to account difficulty increases, but even with the upcoming estimated difficulty increases, it'll still take less than a month for me to get my money back on them.

Just about the lowest risk investment I've ever made in my life.  As long as I get enough BTC to pay for a restocking fee (15%), (which will take all of 2.5 days of mining), I am afterward making a profit.

[marcus_of_augustus]

Chodpaba:

How hard would it be to implement some kind of physical limit on growth factor into your model?

By my reckoning, the total number of GPUs on the planet that will ever be made available for mining is not infinite. The factories that make them can only bring them online at finite rate also. Exponential growth is monster that can hit physical limits quite quickly, without rapid innovation.

I suspect that the limiting factor for network growth will be the factories ability to supply GPU's in the near term if the exponential growth on price is sustained.

How would we go about determining the total number of GPU chips ever produced? (As a starting point for an upper bound on network strength calc.)

[LMGTFY]

I don't understand where people are coming up with their calculations on this, where you actually NEED to calculate it out for a year or better.

I bought 3 5850's.  At current market rates, and assuming MH/s of 300 for each one, they'll pay for themselves in about 17 days.  This is, of course, not taking in to account difficulty increases, but even with the upcoming estimated difficulty increases, it'll still take less than a month for me to get my money back on them.

Just about the lowest risk investment I've ever made in my life.  As long as I get enough BTC to pay for a restocking fee (15%), (which will take all of 2.5 days of mining), I am afterward making a profit.

That is the thing, it is such a good deal that many, many more potential block miners are piling on. Right now it might seem that your 3x5850 rig will pay for itself in 17 days... If you are in in it long term though you do need to think about where you will be in a year or more, including increases in Difficulty. Which is what I have been trying to do.

By all means, mine away. I am not out to discourage mining. But if you intend to make a go of it long term there are many costs which reveal themselves that may have been hidden by stored value, or subsidy for a one-off. Plugging a rig into an outlet in the garage is one thing, but plugging 100 in might take a bit more investment than anticipated... Unless you anticipated it.

The issue for me is "at current market rates". Back in mid February mining seemed like a great investment, and then we entered a six-week period of declining rates, with some eye-watering difficulty increases.

Like Chodpaba, I'm also not discouraging mining. I've done well out of it - I paid off my investment fairly quickly, and I'm into profit. Just be aware that you need to factor "below current market rates" into your calculations. If the price holds, you're profitable sooner than you feared. If the price falls, you should still be within your budget.

[SgtSpike]

I don't understand where people are coming up with their calculations on this, where you actually NEED to calculate it out for a year or better.

I bought 3 5850's.  At current market rates, and assuming MH/s of 300 for each one, they'll pay for themselves in about 17 days.  This is, of course, not taking in to account difficulty increases, but even with the upcoming estimated difficulty increases, it'll still take less than a month for me to get my money back on them.

Just about the lowest risk investment I've ever made in my life.  As long as I get enough BTC to pay for a restocking fee (15%), (which will take all of 2.5 days of mining), I am afterward making a profit.

That is the thing, it is such a good deal that many, many more potential block miners are piling on. Right now it might seem that your 3x5850 rig will pay for itself in 17 days... If you are in in it long term though you do need to think about where you will be in a year or more, including increases in Difficulty. Which is what I have been trying to do.

By all means, mine away. I am not out to discourage mining. But if you intend to make a go of it long term there are many costs which reveal themselves that may have been hidden by stored value, or subsidy for a one-off. Plugging a rig into an outlet in the garage is one thing, but plugging 100 in might take a bit more investment than anticipated... Unless you anticipated it.

I guess I still don't see your point.  Whether I buy 3 5850's or 100, they'll still pay for themselves in about a month.  Though I suppose that in order to run 100 of them, I would likely need an electrical panel upgrade, and some new circuits run.  Rewiring aside though, there's nothing "long-haul" about it.  Anything beyond the cards paying for themselves is just pure profit.  And if it ever reaches a point that revenue from BTC is reduced below the electricity it costs me to mine, then I'll simply stop mining until/if it becomes profitable to do so again.

I'm in it for the long haul (will continue mining as long as it is profitable), but I have virtually zero risk.  I'm not trying to say that my 5850's will continue to bring in as much profit in the future as they will right now, I'm just saying that I shouldn't have any fear about the future, where you indicate that for some strange reason, I should.

[BitterTea]

The more cards you (and others) buy, the faster we find the next 2016 blocks, the higher the difficulty becomes, the more difficult it is to find new blocks, the less money you will make (compared to projections), and the longer it will take to pay off your initial investment.

[JJG]

I guess I still don't see your point.  Whether I buy 3 5850's or 100, they'll still pay for themselves in about a month.  Though I suppose that in order to run 100 of them, I would likely need an electrical panel upgrade, and some new circuits run.  Rewiring aside though, there's nothing "long-haul" about it.  Anything beyond the cards paying for themselves is just pure profit.  And if it ever reaches a point that revenue from BTC is reduced below the electricity it costs me to mine, then I'll simply stop mining until/if it becomes profitable to do so again.

I'm in it for the long haul (will continue mining as long as it is profitable), but I have virtually zero risk.  I'm not trying to say that my 5850's will continue to bring in as much profit in the future as they will right now, I'm just saying that I shouldn't have any fear about the future, where you indicate that for some strange reason, I should.

Unless you have spare computers lying around, you also need a motherboard, processor, RAM, case, PSU, etc. By the time all of that shows up and is configured, the next difficulty will have arrived. And now you're pretty invested.

A few weeks ago when you bought your 5850s, it made perfect sense to do so. Now, it's much more of a gamble. Whether you buy bitcoins directly or buy mining hardware, you're making a bet on bitcoin value going up.

[SgtSpike]

Factor in $125 for each 5850 for a new computer build ($250 rig that would hold two cards).  It'd still be under 2 months for payback.  Admittedly, that's getting into risky territory as far as guaranteed payback, but I digress.

If others are more cautious about investing in mining hardware, all the better for me anyway! 

I actually bought two 5850's near the end of last week, and one just last night.  They'll be here before this weekend though.

[AaronM]

[...]One-off miners will come and go, but those that build out will need to trade at least some of their Bitcoins for their operations to sustain themselves, this provides a floor to BTC velocity that we will never reach.[...]

Okay, so this guarantees a demand for buying local currency with Bitcoins, but the reverse is what's in question: will there be a demand for buying Bitcoins with local currency?

I can think of a situation where nobody wants to buy the miners' Bitcoins with local currency.

If you want Bitcoin to have a future beyond this, then offer real goods and services.

Yeah, I agree. I forgot to mention that other aspect.

[phelix]

If no one is willing to part with their Bitcoins (everyone holds on to them), then Bitcoin is definitely a bubble and the price is only staying high because of new users coming in and speculators buying bitcoins with local currencies.

If you want Bitcoin to have a future, buy real goods and services (not cash, not Ponzi schemes or lotteries) with Bitcoins.

Yeah, but that's not going to happen. One-off miners will come and go, but those that build out will need to trade at least some of their Bitcoins for their operations to sustain themselves, this provides a floor to BTC velocity that we will never reach. Even those who are merely speculating provide needed liquidity to the market, which is not presently big enough to satisfy even a small-sized Wall Street player.

If you want Bitcoin to have a future beyond this, then offer real goods and services.

Why would anyone want to spend something that increases 10% in value per week? I think bitcoin will become something like gold, a stash for your money to protect it from inflation.

[phelix]

some far fetched thoughts about the limits of growth:
(please verify my calculations)

BTC vs. gold
from the numbers on gold on wikipedia I estimate the amount of gold ever mined to 180000 metric tons.
at 1400$/ounce that would theoretically be worth arround  8*10^12$. Dividing by 21 million, the maximum amount of btc there ever can be, results in 380000$/1BTC.

At that price the total value of all btc would play in the same league as gold. I would consider this an upper limit. At a high rate of 12% increase in price per week it would take 114 weeks to get there (similiar to your chart?; what increase rates did you use there?)

network hashing power
With a typical card used for mining having 300MHash/s (HD5870) the current 1,6THash/s could be reached with ~5300 cards.
From this I estimate AMD has shipped 45 million gpus. As a wild guess I say they have 100mhash on average. That makes a total hashing power of 4500THash/s.

Maybe they would be able to ramp up production or the ASICs come but I think if all these GPUs where mining some limit would be reached. It would mean ~2800 times as many cards mining as now. At difficulty increases of 33% every two weeks it would only take 56 weeks to get there. I calculate this hashing power would take 1% of world electricity generation at current efficencies.


Exponential growth is hard to grasp. Hopefully it will continue and bitcoin will go on to world domination.