How many Bitcoiners are there?

[dree12]

I have tried to produce estimates for the number of Bitcoiners using three means of estimation. However, these means disagree significantly. Hence I am asking fellow BitcoinTalk users for feedback.

Methodology 1: Naive Count

This method involves counting the number of users who use each popular form of wallet.

• From https://getaddr.bitnodes.io/, we learn that 7533 run full nodes.
• From https://blockchain.info/charts/my-wallet-n-users, we learn that there are 1600766 MyWallet users.
• From http://blog.coinbase.com/post/78016535692/a-major-coinbase-milestone-1-million-consumer-wallets, we learn there are 1000000+ consumer wallets on CoinBase.

Total: 2608299 Bitcoiners

Strengths:

• This approach counts directly instead of extrapolating from a correlated metric.

Weaknesses:

• Many smaller wallet providers are not enumerated. (Net undercoverage)
• Partial nodes are not counted. (Net undercoverage)
• The number of full nodes is a lower bound; many full nodes cannot be found and enumerated. (Net undercoverage)
• Many Bitcoiners have more than one wallet provider, so are double or even triple-counted. (Net overcoverage)
• Many accounts on wallet providers belong to former Bitcoiner users who are now disinterested or deceased. (Net overcoverage)

Methodology 2: Coinmap.org Extrapolation

This method uses http://www.just-style.com/store/samples/wrdas_samples.pdf and http://coinmap.org as sources. From the first source, there are approximately 5000 retail stores per million inhabitants (as an average of the developed world). From the second source, there are 4333 retail outlets that accept Bitcoin. A trivial calculation yields:

Total: 866600 Bitcoiners

Strengths:

• This approach successfully discounts inactive Bitcoiners, as they do not generate retail demand.

Weaknesses:

• Many stores are not listed on CoinMap. (Net undercoverage)
• Many stores listed on CoinMap are not retail outlets. (Net overcoverage)
• Many stores listed on CoinMap have since ceased to accept Bitcoin. (Net overcoverage)
• There is no evidence that the retail-to-population ratio will hold for Bitcoin. (Unknown effect)

Methodology 3: Address counting

The method employed here is simply to count the number of non-empty or non-trivial addresses. These are addresses with over 1 mBTC. http://bitcoinrichlist.com/charts/bitcoin-distribution-by-address?atblock=300000 provides such a list, giving:

Total: 1456865 Bitcoiners

Strengths:

• This approach is relatively complete: even MyWallet and CoinBase users often have at least one non-empty address (though with shared wallets it is sometimes complicated.)

Weaknesses:

• Some shared wallets may pool addresses of clients together, making most empty. (Net undercoverage)
• Many people have more than one non-trivial address; some may have hundreds. (Net overcoverage)
• Many addresses probably hold coins whose keys have since been destroyed. (Net overcoverage)

[Malin Keshar]

Do you count long time holders as bitcoiners too?
Maybe see the transactions on blockchain could give a hint about how many active btc users(not hodlers) there are, but i'm not sure about how to implement that

[dree12]

Do you count long time holders as bitcoiners too?
Maybe see the transactions on blockchain could give a hint about how many active btc users(not hodlers) there are, but i'm not sure about how to implement that

I should be clear on my definition of Bitcoiner:

One who actively engages in the possession or trade of Bitcoin.

To "actively possess" means to knowingly have a non-trivial amount of bitcoins, so long-term holders are included, but 2010 miners who have since forgotten about their holdings are not.

[Malin Keshar]

Do you count long time holders as bitcoiners too?
Maybe see the transactions on blockchain could give a hint about how many active btc users(not hodlers) there are, but i'm not sure about how to implement that

I should be clear on my definition of Bitcoiner:

One who actively engages in the possession or trade of Bitcoin.

To "actively possess" means to knowingly have a non-trivial amount of bitcoins, so long-term holders are included, but 2010 miners who have since forgotten about their holdings are not.

Would be difficult to see the difference between someone is is saving btc for his retirement from someone that forgot about their btc or lost their privates keys, if not impossible. And someone that recover the key or find the forgotten btc could instantly become a bitcoiner. But not sure if there are enough people in this situation to affect the results in a significant way.
Maybe you should study first who actively do bitcoin transations and trading(including black market uses like Silk Road), then think about how estimate long time holders and people that lost private keys.

[Beliathon]

According to my sources.

[Sindelar1938]

I have heard the 3mn number bandied about but that doesnt differentiate between active and inactive

Around 1mn actives sounds about right

[Nagle]

How many wallets have at least 0.01 BTC? It's probably reasonable to only count those.

[Sindelar1938]

0.01btc is a reasonable minimum bar as that would likely exclude most of the faucet free riders...

[coindozer7]

I assume that little less than 1 million active Bitcoin users is the correct number as of today.
There are many users with multiple Bitcoin addresses and many of them just hold an address or wallet and not active at all.

[Boris-The-Blade]

I have about 20 different wallets and i imagine alot of other people use many multiple wallets.

[master-P]

I have a bunch of different addresses too. I'd estimate that the total number of bitcoiners is below 1 million.

[makebitcoin]

120.119 readers on Reddit Bitcoin at the moment. I guess that says something and you could use those numbers

[franky1]

How many wallets have at least 0.01 BTC? It's probably reasonable to only count those.

i think there are only 800k people MAX...  though i think its alot lower number of actual people than this.

although there are over 1.4 addresses with balances. remember some people have more then one address.

EG i have 5 addresses in my client, i also have 6 addresses linked to me, as they are exchange deposit addresses. so i have 11 addresses that are used and can have fnds on them on any given day for 'my' use.

i know that exchanges sweep thire addresses daily so they are not always full to be shown as addresses with balance. but on any given day (the time the stats was grabbed by the op) the balance in my exchange deposit addresses could have been included in the 1.4mill stats.

also i know that many people make new addresses like confetti, for personal transactions (non re-use address methodology) and hoard coins in umpteen addresses long term. and also merchants make new addresses, and yet again some of those may have appeared in the OP's stats as on any given day they had not been sweeped into a central pot.

EG imagine how many addresses bitpay has displayed to customers of 30,000 merchants in the last hour alone (before having chance to sweep)

[Polycoin]

Do you count non humanoid creatures too?

My pirate friend, TalonBeard, owns 3 bitcoins.

[waldox]

i would say there are about 3 million bitcoiners worldwide

[jc01480]

3 million and one here!  Bring on the truth!

[bitcoiner49er]

Everyone knows there are 42 of us.

[klmist]

How many wallets have at least 0.01 BTC? It's probably reasonable to only count those.

About 1.09 million - you can see how it changes over time here; http://klmist.com/keysreport.html

IMO the actual number of users is probably lower than this.. most users have multiple keys. There must be enormous overlap with all the different bitcoin wallet services and people create multiple coinbase and blockchain.info/mywallet accounts for different phones/tablets.. in the case of blockchain.info, perhaps installing/uninstalling on 3 or 4 devices could end up creating more than 10 accounts for a single user.

Also people sign up with every new service just to check it out, even if they can't use it because they live in another country.. (eg, coinbase and chinese exchanges). It's possible that there are a lot of users hiding behind large shared wallet keys, but I'm not really convinced of that.. surely anyone putting money into bitcoin will at least have one 0.01BTC key of their own too?

[franky1]

i would say there are about 3 million bitcoiners worldwide

there are not 3 million bitcoin addresses holding bitcoins.. thus there are not 3million bitcoiners.. id say as before maybe 800k at most and maybe multiple millions watching from the outside but not yet stuck their toe into bitcoin.

[dree12]

It's become clear to me that this problem is a lot harder than it first seemed. I've identified three general classes problems that plague the approaches.

• They tend to double count certain Bitcoiners (e.g. counting addresses/accounts of which one person can have many).
• They tend to omit certain Bitcoiners (e.g. those using hosted/shared wallets, those who don't trade, etc.).
• There is sometimes measurement uncertainty (e.g. counting full nodes).

I feel that there probably isn't a methodology immune to all three problems. However, if methods that suffer from only one problem can be developed, it should be possible to come up with bounds and uncertainty estimates.