Stephen Reed's Million Dollar Logistic Model

[2586]

And my estimates of Bitcoin network energy consumption are for the period after the halving.  I don't know how could you miss this, since I put all three periods in:

--->
If miners decide to sell all their mined Bitcoins, (or 50%), this is how much daily fresh money is needed on the exchanges for the price to be stable:

2014-2016 :   $3.6 billion ($1.8 billion if 50% sold)
2017-2020:   $1.8 billion ($0.9 billion)
2021-2024:   $0.9 billion ($.45 billion)
etc.
<---

Better double check your math. We've already had one reward halving, taking the block reward from 50 BTC to 25. The next one will take it from 25 to 12.5. If the price were $1M per coin right now, there would be $3.6B worth of coins generated per day.

Bitcoin miners would have incentives to spend up to $1.8 billion dollars, daily, on their non-free running costs, of which the major part would be electricity.

If you were to purchase and run one of those for a year, the power cost for the Blade would be 17.2% of your total cost. For the Cube, it would be 24.16%.

And that is exactly why we see such huge increases in difficulty.  

Want to know when the increases in difficulty will stop?  When power costs of generating one BTC get close to one BTC.  In other words, in the situation when most of the block awards get used on paying electricity bills.

You might want to present some evidence for that. The total cost of mining 1 BTC will approach 1 BTC, but you can't just expect us to assume that the majority of that cost will be for electricity.

[1714016816]

1714016816

[1714016816]

1714016816

[1714016816]

1714016816

[1714016816]

1714016816

[mestar]

And my estimates of Bitcoin network energy consumption are for the period after the halving.  I don't know how could you miss this, since I put all three periods in:

--->
If miners decide to sell all their mined Bitcoins, (or 50%), this is how much daily fresh money is needed on the exchanges for the price to be stable:

2014-2016 :   $3.6 billion ($1.8 billion if 50% sold)
2017-2020:   $1.8 billion ($0.9 billion)
2021-2024:   $0.9 billion ($.45 billion)
etc.
<---

Better double check your math. We've already had one reward halving, taking the block reward from 50 BTC to 25. The next one will take it from 25 to 12.5. If the price were $1M per coin right now, there would be $3.6B worth of coins generated per day.

Ok, let me double check that.  Checking... Checking... Yup, the line that starts with 2014 still shows $3.6 billion. 

[aminorex]

I like this thread because it suggests a good way to estimate when the logistic curve tips over, i.e. a parametric model.

My impression is that roughly 30% of mined coin get sold.  That number will go up as the price curve inflection becomes obvious, and tend to 100% in the limit.  Suppose it is 50% at the inflection.
The long-term value will be stable when the issuance equals the organic growth of the economy.   The bitcoin economy is some percentage of the total human economy.

G = global GDP, ~42tln USD2012 / an.

E_btc = P_btc * Q_btc = P_btc/tot * E_tot = G * r_growth^n_years  for an exponentially growing global economy.

M_btc = P_btc * Q_btc / V_btc

E_btc(y_i+1) - E_btc(y_i) = I_btc * M_btc * N_btc =  P_btc/total * G * (exp((y_i+1-y_0)*ln(r_growth))  - exp((y_i-y_0)*ln(r_growth)))

I_btc = delta(E_btc,i,i+1)/(M_btc*N_btc)

THEREFORE

Price stability occurs when the annual block reward reaches P_btc:total * G * r * (r - 1) / M_btc * N_btc

Now we have enough data to fit a logistic, except for one thing:  We don't have a good estimator of P_btc:total, the proportion of the human economy denominated in BTC, at the equilibrium time.

Well, that and the exponential growth of the underlying economy is probably not a good axiom going forward.  But at least it is enough to model the scenario topology parametrically, and find the catastrophe lines.


 

[mestar]

You might want to present some evidence for that. The total cost of mining 1 BTC will approach 1 BTC, but you can't just expect us to assume that the majority of that cost will be for electricity.

Sure, the electricity will be the limiting factor, or better to say, total running costs, which include hosting costs, your time, whatever.  As opposed to your hardware costs which you have to pay up-front.

Now the question is, once you have it running, at what point do you turn your mining operation off?  You turn it off when your running costs become higher than your rewards.  This is true no matter what you paid for your hardware.

Then you can argue that people will simply not buy more hardware if your running costs are high.  Ok, we have to look at the endgame.  Say, we got to the point where difficulty is not increasing anymore.  Let say that at this point your running costs are at 51%.  Do you buy your money printing machine that takes $51 of electricity and turns it into a $100 worth of Bitcoins?  Sure, if you can pay your investment in some reasonable time, like 1 year, or 3 years.  Business that can turn profit after only one year is still a incredibly profitable one. 

After initial research and development costs, mass producing chips is fcking cheap.  Today's mining hardware is expensive because it can be.  From this thread https://bitcointalk.org/index.php?topic=295270.0 Puppet estimated that a $36 chip uses $270 of power per year (at 0.12 $/kWH).  So, at 51%, this chips can pay for itself in less than two months.

If a $36 chip uses almost 10 times its cost in electricity per year, it is kind of obvious that the cost of electricity, and not your hardware costs will be the limiting factor.  And even if it costs $270, we are exactly at break-even point after one year, at 50% profitability.  So, it is safe to say that, yes, electricity will be the majority of your costs.

[BitDreams]

Wiki'd it so I thought I'd share 

"parametric model"

can be described using a finite number of parameters. These parameters are usually collected together to form a single k-dimensional parameter vector θ = (θ1, θ2, …, θk).

The big picture using what we best know.

[marcus_of_augustus]

I like this thread because it suggests a good way to estimate when the logistic curve tips over, i.e. a parametric model.

My impression is that roughly 30% of mined coin get sold.  That number will go up as the price curve inflection becomes obvious, and tend to 100% in the limit.  Suppose it is 50% at the inflection.
The long-term value will be stable when the issuance equals the organic growth of the economy.   The bitcoin economy is some percentage of the total human economy.

G = global GDP, ~42tln USD2012 / an.

E_btc = P_btc * Q_btc = P_btc/tot * E_tot = G * r_growth^n_years  for an exponentially growing global economy.

M_btc = P_btc * Q_btc / V_btc

E_btc(y_i+1) - E_btc(y_i) = I_btc * M_btc * N_btc =  P_btc/total * G * (exp((y_i+1-y_0)*ln(r_growth))  - exp((y_i-y_0)*ln(r_growth)))

I_btc = delta(E_btc,i,i+1)/(M_btc*N_btc)

THEREFORE

Price stability occurs when the annual block reward reaches P_btc:total * G * r * (r - 1) / M_btc * N_btc

Now we have enough data to fit a logistic, except for one thing:  We don't have a good estimator of P_btc:total, the proportion of the human economy denominated in BTC, at the equilibrium time.

Well, that and the exponential growth of the underlying economy is probably not a good axiom going forward.  But at least it is enough to model the scenario topology parametrically, and find the catastrophe lines.

Nice work, I like your reasoning.

[Andrew717]

Not included are all the efficiencies Bitcoin will introduce into the economy if it goes mainstream. A fully internationalized, frictionless division of labor will be an incredible boon to the economy. And there's so much more. I think a price target of $100 million, conservatively, is more realistic.

If Bitcoin goes big, it pretty much has to go whole hog. We are witnessing nothing less than a global transition from physical goods and trust-based money to an unimpeachable universal asset ledger. Anything else will look archaic by comparison, and the value added to the economy in the next 10 years will blow even these first 20 years of the modern Internet out of the water.

100 million dollars per bitcoin is ridiculous, in my opinion. Given that there are 10,000 people with over 100 bitcoins, there would be 10,000 people worth over 10 billion dollars. Today, there are only 1,500 people in the world worth one billion dollars or more.

In 1982 there were only 12 billionaires.   In 2000 there were about 300.   And now 1,500.   So in 2025 maybe 10,000+ billionaires. 

[bitrider]

I like this thread because it suggests a good way to estimate when the logistic curve tips over, i.e. a parametric model.

My impression is that roughly 30% of mined coin get sold.  That number will go up as the price curve inflection becomes obvious, and tend to 100% in the limit.  Suppose it is 50% at the inflection.
The long-term value will be stable when the issuance equals the organic growth of the economy.   The bitcoin economy is some percentage of the total human economy.

G = global GDP, ~42tln USD2012 / an.

E_btc = P_btc * Q_btc = P_btc/tot * E_tot = G * r_growth^n_years  for an exponentially growing global economy.

M_btc = P_btc * Q_btc / V_btc

E_btc(y_i+1) - E_btc(y_i) = I_btc * M_btc * N_btc =  P_btc/total * G * (exp((y_i+1-y_0)*ln(r_growth))  - exp((y_i-y_0)*ln(r_growth)))

I_btc = delta(E_btc,i,i+1)/(M_btc*N_btc)

THEREFORE

Price stability occurs when the annual block reward reaches P_btc:total * G * r * (r - 1) / M_btc * N_btc

Now we have enough data to fit a logistic, except for one thing:  We don't have a good estimator of P_btc:total, the proportion of the human economy denominated in BTC, at the equilibrium time.

Well, that and the exponential growth of the underlying economy is probably not a good axiom going forward.  But at least it is enough to model the scenario topology parametrically, and find the catastrophe lines.

Nice work, I like your reasoning.

I love this forum.

+1

[aminorex]

I like your reasoning.

Hence the new sig

[SlipperySlope]

The Joy of Mining

The emergence of scrypt-based coins, popularized by Litecoin, has made it profitable to again earn bitcoin on my dusted-off graphic card mining rigs. I am running three rigs with a total of six 5770 cards. The operating system is Ubuntu, I run cgminer 3.7.2 and achieve about 144 KHs per card. The conversion to bitcoin is handled automatically by the pool software which heuristically chooses among various exchange-traded crypto currency coins and mines a particular network until it becomes unprofitable to mine when compared to other candidate crypto currency coins. I use TradeMyBit as its operator posts frequently on the pool forum. It is an easy upgrade from the obsolete 5770 cards to expensive but currently in stock R9 280x cards. My first one is on its way - yay!

The coinwarz web site gives a great profit calculation of daily dollar profit on each ranked crypto currency coin given hashing power and price of power. When I input the figures for a single R9 280x card, 700 MHs and 250 watts respectively, the projected Revenue / Profit (per day) for my local power usage is projected to be $15.24 / $14.54 for Tagcoin at rank number one, and  $8.70 / $8.00 for Neocoin ranked five. The actual profit is lower, chiefly due to fees and the effects of periodic coin network switching that result in stale shares and rejects.

When bitcoin prices eventually explosively rally again as suggested by the logistic model trendline, it is very likely that the most-popular-at-the-time scrypt-based coins will exceed the price acceleration of bitcoin, i.e. have a higher beta. I look forward to manually selecting coins to mine from TMB at that time for postponed exchange into bitcoin.

[BitDreams]

Posted by SlipperySlope: When bitcoin prices eventually explosively rally again as suggested by the logistic model trendline, it is very likely that the most-popular-at-the-time scrypt-based coins will exceed the price acceleration of bitcoin, i.e. have a higher beta.

These younger coins are at an earlier point on their own unique S curve, mirroring Bitcoin from perhaps 2010. Their price movements come as less of a surprise as Bitcoin blazes the pricing trail. Looking at Litecoin which has been around for as long as any other alt coin, price has retreated (23 todays avg) from the high (42 apx high) almost 50%. Bitcoin appears to be settling about 25% down from the high near 1200's.

I believe mining the alt coins is beneficial to Bitcoin and welcome the alt coins in fact. DOGE coin has introduced thousands of new users who can learn and experience cryptocurrency with little to no risk.

Leadership from Bitcoin, innovation from the Altcoins. I wonder what consolidation among coins will look like? I hope enthusiasts discover methods of gracefully merging those deserving more than abandonment. Nothing sadder than pulling up a QT wallet and finding no peers.

[SlipperySlope]

Leadership from Bitcoin, innovation from the Altcoins. I wonder what consolidation among coins will look like? I hope enthusiasts discover methods of gracefully merging those deserving more than abandonment. Nothing sadder than pulling up a QT wallet and finding no peers.

Interesting points indeed. New altcoin promoters should give some assurances that sufficient long term mining rigs will be dedicated to at least one pool mining that coin,

[kdrop22]

The Joy of Mining

The emergence of scrypt-based coins, popularized by Litecoin, has made it profitable to again earn bitcoin on my dusted-off graphic card mining rigs. I am running three rigs with a total of six 5770 cards. The operating system is Ubuntu, I run cgminer 3.7.2 and achieve about 144 KHs per card. The conversion to bitcoin is handled automatically by the pool software which heuristically chooses among various exchange-traded crypto currency coins and mines a particular network until it becomes unprofitable to mine when compared to other candidate crypto currency coins. I use TradeMyBit as its operator posts frequently on the pool forum. It is an easy upgrade from the obsolete 5770 cards to expensive but currently in stock R9 280x cards. My first one is on its way - yay!

The coinwarz web site gives a great profit calculation of daily dollar profit on each ranked crypto currency coin given hashing power and price of power. When I input the figures for a single R9 280x card, 700 MHs and 250 watts respectively, the projected Revenue / Profit (per day) for my local power usage is projected to be $15.24 / $14.54 for Tagcoin at rank number one, and  $8.70 / $8.00 for Neocoin ranked five. The actual profit is lower, chiefly due to fees and the effects of periodic coin network switching that result in stale shares and rejects.

When bitcoin prices eventually explosively rally again as suggested by the logistic model trendline, it is very likely that the most-popular-at-the-time scrypt-based coins will exceed the price acceleration of bitcoin, i.e. have a higher beta. I look forward to manually selecting coins to mine from TMB at that time for postponed exchange into bitcoin.

Are you really getting an ROI, when you figure in the time you spend setting it up and dealing with issues to etc... Just Curious to know.

[aminorex]

Are you really getting an ROI

Mining is a crap shoot, very bimodal.  Either you win or lose money.  If you do it well, persistently, you come out ahead.  Is it worth it?  Depends on how desperate you are for more coins, I guess.  I'm pretty desperate.  My day job only pays 300k/an, and I already took out a home loan and a business loan, so to get more coins, I must mine.

[skivrmt]

Are you really getting an ROI

Mining is a crap shoot, very bimodal.  Either you win or lose money.  If you do it well, persistently, you come out ahead.  Is it worth it?  Depends on how desperate you are for more coins, I guess.  I'm pretty desperate.  My day job only pays 300k/an, and I already took out a home loan and a business loan, so to get more coins, I must mine.

Your day job is 300k/an.  An = annually?  Is that in Lira?  Pesos?  Hungarian Floret? 

When you save you "must" mine, I say you "must" get a better financial planner. 

[ZephramC]

Are you really getting an ROI

Mining is a crap shoot, very bimodal.  Either you win or lose money.  If you do it well, persistently, you come out ahead.  Is it worth it?  Depends on how desperate you are for more coins, I guess.  I'm pretty desperate.  My day job only pays 300k/an, and I already took out a home loan and a business loan, so to get more coins, I must mine.

Your day job is 300k/an.  An = annually?  Is that in Lira?  Pesos?  Hungarian Floret? 

When you save you "must" mine, I say you "must" get a better financial planner. 

Here is something for some comparison :-)
http://www.globalrichlist.com/

[aminorex]

When you save you "must" mine, I say you "must" get a better financial planner. 

No financial planner would have my goals.  If they did, they would be institutionalized.  Also, I can't afford to hire anyone smarter than me, and I won't hire anyone dumber than me. 

[rpietila]

When you save you "must" mine, I say you "must" get a better financial planner. 

No financial planner would have my goals.  If they did, they would be institutionalized.  Also, I can't afford to hire anyone smarter than me, and I won't hire anyone dumber than me. 

Wisdom. (Except I never had a job that would pay $300k)

[SlipperySlope]

Are you really getting an ROI, when you figure in the time you spend setting it up and dealing with issues to etc... Just Curious to know.

I am a computer hobbyist, so it is better to have an engaging hobby that pays even a little as compared to a hobby that simply costs.

Furthermore, the setup of a typical mining rig is quite scalable insofar as additional rigs are hardware and software copies of the reference working rig. But suppose that I replace the six obsolete 5770 cards with modern R9 280x cards. One could argue that with bitcoin increasing in price 10x per year, the funds spent on a mining farm scale-out could be better invested in simply buying coin. I expect that I will incrementally upgrade my rigs.

Another concern is that ASIC rigs will be available sometime this year for scrypt-altcoin, e.g. Litecoin, mining. I am not too worried about the sudden obsolescence of my graphic card rigs because a new proof-of-work algorithm could be designed to appeal to the large number of graphic-card miners. It is the population of miners that provide the initial enthusiasm for a coin.

[TooDumbForBitcoin]

Also, I can't afford to hire anyone smarter than me, and I won't hire anyone dumber than me. 

I feel discriminated against.  Helen, get my lawyer on the phone!