21 million cap

[kbriggs]

I just become aware of Bitcoin this week watching "Security Now" at Twit.tv. The main question I have about the system is why was 21 million bitcoins selected as the final cap?  Why not something more "round" like 100 million or 1 billion?  And why such a relatively small number?  I understand about the 8 decimal places giving you essentially 2.1 quadrillion tradeable units.  But the general public is not used to dealing with more than 2 decimal places in their currency. If the system took off to even the size of PayPal alone, a single bitcoin is going to be worth some huge number and we're all going to be buying things priced at inconvenient amounts like 0.00023, etc., right?

[1714146667]

1714146667

[1714146667]

1714146667

[kiba]

You could use scientific notation or perhaps specific decimal point as default.

[davout]

I just become aware of Bitcoin this week watching "Security Now" at Twit.tv. The main question I have about the system is why was 21 million bitcoins selected as the final cap?  Why not something more "round" like 100 million or 1 billion?  And why such a relatively small number?  I understand about the 8 decimal places giving you essentially 2.1 quadrillion tradeable units.  But the general public is not used to dealing with more than 2 decimal places in their currency. If the system took off to even the size of PayPal alone, a single bitcoin is going to be worth some huge number and we're all going to be buying things priced at inconvenient amounts like 0.00023, etc., right?

21 million, purely arbitrary.
Also, you don't track units, you track available balances which conceptually allows for indefinite division, there's no such thing as a conceptual atomic bitcoin unit.

[FatherMcGruder]

I just become aware of Bitcoin this week watching "Security Now" at Twit.tv. The main question I have about the system is why was 21 million bitcoins selected as the final cap?  Why not something more "round" like 100 million or 1 billion?  And why such a relatively small number?  I understand about the 8 decimal places giving you essentially 2.1 quadrillion tradeable units.  But the general public is not used to dealing with more than 2 decimal places in their currency. If the system took off to even the size of PayPal alone, a single bitcoin is going to be worth some huge number and we're all going to be buying things priced at inconvenient amounts like 0.00023, etc., right?

0.00023 BTC = 0.23 mBTC.

[kbriggs]

You could use scientific notation or perhaps specific decimal point as default.

The average guy on the street has no idea what scientific notation is.

[davout]

You could use scientific notation or perhaps specific decimal point as default.

The average guy on the street has no idea what scientific notation is.

It is just as easy to transport 2000 bitcoins than 0.00002 bitcoins.
When money is immaterial it all boils down to notation conventions.

[kbriggs]

Also, you don't track units, you track available balances which conceptually allows for indefinite division, there's no such thing as a conceptual atomic bitcoin unit.

But if you're a merchant, you gotta post a price next to your product.  I'd rather there were 10 trillion bitcoins in circulation and I could post a price mostly in integers that my customers would understand.

[kbriggs]

It is just as easy to transport 2000 bitcoins than 0.00002 bitcoins.
When money is immaterial it all boils down to notation conventions.
[/quote]

Transportation isn't the issue. I'm talking about the market dynamic between merchant and customer. Average people who think "fractions are hard". Not the engineers and cryptologists that hang around here.

[kiba]

The average guy on the street has no idea what scientific notation is.

Really? I learn that in schools like...several time!

[FatherMcGruder]

The average guy on the street has no idea what scientific notation is.

Nor does the average guy know that the calories on your foods' nutrition labels is actually kilocalories. If a pint should ever cost 0.005 BTC, your bartender will ask you for "five mills" or something like that. He won't say "five thousandths of a bitcoin".

But if you're a merchant, you gotta post a price next to your product.  I'd rather there were 10 trillion bitcoins in circulation and I could post a price mostly in integers that my customers would understand.

If you ever have to sell something for millibitcoins, other merchants will probably have to do the same and customers will just come to expect it.

<inner physics teacher>Always include units though.</inner physics teacher>

[kbriggs]

The average guy on the street has no idea what scientific notation is.

Really? I learn that in schools like...several time!

Yes, really.  We're not normal, we're geeks.  Like everyone else reading this.

[kiba]

Yes, really.  We're not normal, we're geeks.  Like everyone else reading this.

I swear that everyone in school learned this stuff!

[grondilu]

The average guy on the street has no idea what scientific notation is.

Come on.  He doesn't know what a millimeter is?  He doesn't ever use words such as kiloWatt, kilogram, MegaHertz, GigaOctet, and so on? ...

[kbriggs]

Yes, really.  We're not normal, we're geeks.  Like everyone else reading this.

I swear that everyone in school learned this stuff!

I learned a lot of crap in school and retained it only to the next test.  Ask your mother if she knows what scientific notation is.

[FatherMcGruder]

Americans used to describe their currency in cents. Then as it inflated, people came to refer to dollars, Benjamins, and grands. Same thing with Bitcoin, except the trends will follow deflation.

[grondilu]

I learned a lot of crap in school and retained it only to the next test.  Ask your mother if she knows what scientific notation is.

Well, I guess anyone who is incapable of learning a few prefixes does not have much economic value anyway.  So I don't mind if these people don't use bitcoin.

PS.  If I ask my mother what scientific notation is, she will obviously say she has no idea.  But if I ask her how many millimeters there are in on meter, I'm pretty sure she'll have to give it about 10 seconds thought, but hopefully she'll come out with the right answer.

[marcus_of_augustus]

Millibitcoins
Microbitcoins
Nanobitcoins
Picobitcoins

I don't see an issue ... maybe just a small advantage for the scientifically literate who understand what the those things actually mean.

Bitcoin cent is kind of a mouthful ... centibitcoins or bitcents?

then it could get messy but still decipherful ... millibitcents, microbitcents, etc ... trying to know what terminology people will use for their monetary divisions or numbering is futile. Look at the UK billion and the US billion ... and how many people on the street even understand what a billion is .... they hear it but  ....?

[LZ]

1 BTC ≈ 1 USD now. Does someone have any problems with the 0.001 USD now?
I think not, because you can always sell enough goods to get at least 0.01 USD.
And, by the way, if bitcoin would cost too much, we can always move the point.

[kbriggs]

1 BTC ≈ 1 USD now. Has someone any problem with the 0.001 USD now?

Did you mean 0.01 USD?  Because many would have a problem if they saw a single item priced at a tenth of a cent.  But as FatherMcGruder said, people will likely adapt to a common nomenclature over time. I still think 21 million was an odd choice, though.

[LZ]

No, I meant exactly 0.001 USD. And it is not a problem because you can sell 10 pieces for 1 cent.

[mndrix]

I still think 21 million was an odd choice, though.

I suspect that Satoshi started with an initial per-block reward of 50 BTC which cuts in half every 4 years (210,000 blocks).  The consequence was 21 million coins over time.  As mentioned above, the top limit doesn't really matter, but using psychologically convenient numbers when the system bootstraps probably does.

[bossweld]

Yes, really.  We're not normal, we're geeks.  Like everyone else reading this.

I swear that everyone in school learned this stuff!

I learned a lot of crap in school and retained it only to the next test.  Ask your mother if she knows what scientific notation is.

Exactly. It may be on the school curriculum, but most people have no interest in, or a need to use, scientific notation. Or any numbers outside of what they have to deal with in their daily life. And what you don't use, you lose.

If bitcoin wants to be taken seriously by a wider audience, it will need to address this. It is (or will be) one more hurdle to acceptance, and there are more than enough hurdles for an idea like bitcoin to deal with in the first place.

Americans used to describe their currency in cents. Then as it inflated, people came to refer to dollars, Benjamins, and grands. Same thing with Bitcoin, except the trends will follow deflation.

The rate of this change is fairly key to the discussion. If the change from bitcoins to milli-bitcoins happens over a couple of generations, fine. If it happens in the space of 10 years, that's a problem.

People need to be able to compare prices over time to be able to make useful decisions about purchases now compared to the past, and their relative value. Large-scale deflation raises the same problems with this, and with menu costs, as large-scale inflation does.

[RHorning]

No, I meant exactly 0.001 USD. And it is not a problem because you can sell 10 pieces for 1 cent.

0.001 USD is called a mill.  It did find some use early on in the history of America, and it still can be found in American gasoline refilling stations (where the gasoline price shows an additional 9 mills per gallon).  It also applies to taxes where sometimes taxation rates are set in terms of a number of mills per thousand dollars of value.

Considering that the penny is now threatened with extinction due to inflation, it is obviously a seldom used denomination.  If the dollar ever went deflationary again over a prolonged period of time it might be used again, but that wouldn't happen as long as the Federal Reserve Bank is in charge of monetary policy in relation to the U.S. Dollar.

[FatherMcGruder]

The rate of this change is fairly key to the discussion. If the change from bitcoins to milli-bitcoins happens over a couple of generations, fine. If it happens in the space of 10 years, that's a problem.

The rate of Bitcoin adoption is pretty key, too. I suspect that the rate of deflation will decrease with the rate of newcomers over time.

[kbriggs]

I suspect that Satoshi started with an initial per-block reward of 50 BTC which cuts in half every 4 years (210,000 blocks).  The consequence was 21 million coins over time.

210,000 is an equally strange number, other than being exactly 1% of 21 million.

[Local]

I suspect that Satoshi started with an initial per-block reward of 50 BTC which cuts in half every 4 years (210,000 blocks).  The consequence was 21 million coins over time.

210,000 is an equally strange number, other than being exactly 1% of 21 million.

210,000 comes from 10 minutes, 4 years, and 50BTC/block. Those are a little less strange imo. Only the block award is arbitrary, but 50 feels nice to me.

edit: arbitrary is the wrong word. I mean that 10 minutes can be chosen for a reason and so can 4 years.

[MoonShadow]

I suspect that Satoshi started with an initial per-block reward of 50 BTC which cuts in half every 4 years (210,000 blocks).  The consequence was 21 million coins over time.

210,000 is an equally strange number, other than being exactly 1% of 21 million.

No.  The number comes from the combinations of the initial block reward (50 coins), the target blocks per hour (6) and the halving period (4 years).  There will be 10.5 million created in the first four year period, and then the reward will be halved while all other metrics remain the same.  As the reward continues to half again with each four year period, the total number of coins issued will trend toward the mathmatical limit (as in a logrithmic) of 21 million.  The numbers that define the outcomes are the initial reward, the target block interval, and the halving term.  All of these were design decisions that resulted in the outcome of 21 million, not the other way around.  The interval could have been 9 minutes, or 12, or anything; same with the halving period or the initial reward.  The interval needs to occur quick enough to verify transactions within a rational amount of time, and be long enough for a huge future bitcoin network to propagate transactions and blocks without significant latency issues; but why 10 minutes and not 6 or 15?  Mostly it was an arbitrary design decision, and 6 or 15 or 525 seconds or anything else would have worked.  The same is true for the initial block reward, why a round decimal and not a round binary, such as 64 or 128?  Why 50 and not 100?  Why an even four years and not 3, 5 or a moving term?  (i.e. why half only the reward?  why not two-thirds of the reward and two-thirds of the halving term?  too complex?)  In the end, someone had to make decisions before this all began, and their merits are arguable, but in the end they are still arbitrary.

[theymos]

The end total was probably also given some consideration. If an exact number of blocks for 4 years was used, the result would be some odd number:
http://blockexplorer.com/q/changeparams?interval=204480&precision=8&subsidy=50
You'd have to say that the amount of BTC will never go beyond 20.5 million in that case, which doesn't have the same ring to it.

[kbriggs]

The number comes from the combinations of the initial block reward (50 coins), the target blocks per hour (6) and the halving period (4 years).

Ah, that makes more sense now.

[Tsudico]

The number comes from the combinations of the initial block reward (50 coins), the target blocks per hour (6) and the halving period (4 years).

Ah, that makes more sense now.

And here I thought it was because when you divide the starting number of coins generated by the number of blocks before the coins halve and then multiply by the current recommended transaction fee you get the answer to life, the universe, and everything:

(210,000 / 50) * 0.01 = 42

[MoonShadow]

The number comes from the combinations of the initial block reward (50 coins), the target blocks per hour (6) and the halving period (4 years).

Ah, that makes more sense now.

And here I thought it was because when you divide the starting number of coins generated by the number of blocks before the coins halve and then multiply by the current recommended transaction fee you get the answer to life, the universe, and everything:

(210,000 / 50) * 0.01 = 42

May you be touched by his noodlely appendage.

[Vasili Sviridov]

May you be touched by his noodlely appendage.

Ramen, brother.

[Reiku]

And here I thought it was because when you divide the starting number of coins generated by the number of blocks before the coins halve and then multiply by the current recommended transaction fee you get the answer to life, the universe, and everything:

(210,000 / 50) * 0.01 = 42

Screencapped and added to flashdrive.