Bitcoin Forum

Just observed that in the past 24 hours, we have

7,149,570.84

Bitcoins sent.

Is this the highest ever? Are there individual big chucnks of BTC sent around? IS Satoshi active again :-)  ??

It is frequently brought up but it's pointless. Because bitcoin transactions have frequently two outputs, one of which is change, sending 1 BTC every hour from a wallet with 50000 BTC will result in 1.2 million BTC sent, while only 24 BTC was sent to a wallet owned by a different person.

Yesterday, there were several transactions with such a large wallet. Maybe it belongs to  mtgox?

Someone was sending the same ~95000 BTC repeatedly. He could have gotten the "daily sent BTC" to 22 million if he kept it up.

Thanks for clarifying+

Is there anyway a more accurate number could be calculated? One way would be to only count the smaller part of the transaction, assuming people are spending less than half of their total wallet.  This would at least be an underestimate instead of a nearly meaningless number.

I think you just found a inescapable vulnerability, what if someone sets up two accounts and they just send bitcoin back and forth.  It would crash the system.  I can see no fix. 

If you built up a map of the blocks from blockexplorer you could identify coins that were circulating through small networks, perhaps two wallet files, or a small group of friends

This has been discussed.  The fix is transaction fees.

Knew there had to be a fix somewhere thanks.  When do transaction fees kick in.

A better global metric of transaction volume would be the number of bitcoindays destroyed.

I believe that transactions are prioritized according to the value of the transaction multiplied by the amount of time since the coins were spent. A similar concept is useful in measuring the transaction volume.

If someone has 100BTC that they received a week ago and they spend it then 700 bitcoin days have been destroyed. If they take those 100BTC and send them to several addresses and then spend them then although the total transaction volume could be arbitrarily large the number of bitcoindays destroyed is still 700.

Currently, bitcoindays are replenished at around six million bitcoindays per day which reflects that there are about six million bitcoins in existence.

If there are days when there are few transactions then the total number of bitcoindays increases dramatically. If everyone exchanged all their bitcoins at once then the total number of bitcoindays would be reduced to zero. Some bitcoins will never be spent because the private keys have been lost (or never existed) so the number of bitcoindays will never actually fall to zero.

Note how transaction flooding and mix-network activity do not significantly influence the number of bitcoindays destroyed. I believe that the bitcoindays measure is a good indicator of market health and participation.

ByteCoin

This is a really good idea, and would be a great way to measure velocity.

+1
Excellent point, ByteCoin!

Of course this activity will not reflect activity that occurs within eWallet services such as MyBitcoin and Mt. Gox.  Those transactions made between accounts within the service are never seen by the Bitcoin network.

Very good idea.  Anybody want to implement it?  I've got a Python tool that walks the block chain gathering transaction statistics:
  https://github.com/gavinandresen/bitcointools/blob/master/statistics.py

Teaching it to compute 'bitcoindays destroyed' shouldn't be terribly hard.  I think.

This morning I taught it to add just the smallest or just the largest output in each TxOut and report the range to get an estimate of 'true' transaction value being exchanged without counting change TxOuts or mining pool payouts.

So, to be conservative, assume that the biggest-value TxOut for every transaction is change and the smallest is the actual bitcoins being transferred.  Taking the smallest TxOut of all the transactions last month, an average of about 35,000 BTC were sent per day.

This month the average is about 55,000 BTC per day.   Add in the MtGox trading volume to get a reasonable lower estimate of something like 70-80,000 BTC changing hands every day.

This is a great approach. thanks so much :)

http://forum.bitcoin.org/index.php?topic=9300

Two questions maybe someone can answer:

How could one determine the ACTUAL number of BTC transfered daily?

Where do transaction fees go?

www.bitcoinwatch.com


Transaction fees are added to the block reward given to the miner who finds the solution to the block your transaction is included within.

Transaction fees have already kicked in. Right now, we still get a pretty quick response even if we don't pay transaction fees. But if the network ever started getting clogged, you could pay a small transaction fee to be given higher priority. In the future, it is possible that transactions without fees take days, while transactions with fees go much quicker.  

As the volume of transactions gets higher and the incentive for finding blocks gets lower (bitcoins per block get divided by 2 every 4 years, and difficulty is rising quickly), isn't there a risk that in the future, fees become the main incentive, increase dramatically in price (and also become unavoidable)? Then doing a bitcoin transaction would be for rich people only. We certainly don't want Bitcoin to become a rich people tool – those who can afford the high taxes.

Yes, that is a possibility. But since the network is open, you can always assume that someone will be mining and taking free transaction, for the good of the network. It might take a while to get confirmed, but it'll go through sooner or later. And if transaction fees were really getting high, that is just incentive for more miners to get involved. If people are leaving the bitcoin network because of high fees, then miners have a reason to try accepting more low fees, to keep the network alive. It should all work itself out. The future structure of fees isn't yet completely clear, and everything is still tweak-able.

Eventually, mining stops, since the supply of potential coins is capped at a number about 4x the current number in existence.

Mining will not stop. people will pay to mine (by paying the electric bill) just to keep their own transactions going, also to keep their investment secure.

That isn't how is works. Even without the reward, miners will be doing it for the transaction fees. All transaction have to be stuck in a block, even after the reward is gone. The miners will no longer be mining for a pre-set reward, but for the transaction fees of the transaction in the block they make. Even now they get that reward, but it is usually pretty small compared to the 50 BTC reward. Although I have seen at least one block with 3 BTC of fees in it.

The upshot is that mining must continue as long as bitcoin lives.

In the future I see big commercial players of Bitcoin running their own mining operations not for Bitcoin but to ensure the sustainability of Bitcoin itself as it is their business.  For instance, a friend of mine that works for Google vaguely (confidential on the details) told me that if the Internet were to go away in terms of a disaster any one of their major datacenters could bring it back online.

Some thoughts and a confirmation of my understanding of 'BTCDD':

BTCDD is different than velocity in the traditional sense, because it uniquely accounts for how long an individual coin has been in the hands of the owner. 'Velocity' is a more generic measure which calculates how many times the same supply of money has passed through a given economy to support a given level of activity (GDP). If you applied this concept to Bitcoin, you’d be fooled by the fact that transactional volume does not represent ‘spending’, because of the fact nearly all BTC transactions return some fraction of coins to the spender (although you cannot determine with certainty whether they are returning to the spender’s wallet, or going someplace else as part of genuine additional spend; technically, it is impossible to see from any given naked transaction which is the ‘spent’ part, and which is the ‘change’ part). BTCDD deals with this by adjusting all transactions to account for the number of days since the public key associated with the coins spent (and therefore the coins themselves), last showed up in the block chain.

I also found: http://abe.john-edwin-tobey.org/chain/Bitcoin?count=2016&hi=143075 and cranked the last 2 weeks of data through Excel very quickly to get a short term graph of the trend:

https://i.imgur.com/PJ74j.png

The slope ‘down and to the right’ represents a hoarding trend as time goes by and the BTC days figure ticks up by 7M / day, and the ‘upticks’ represent some combination of medium sized transactions with ‘old’ coin, or larger transactions with somewhat newer coin. Theoretically, I think the conditions for the ‘reset’ of BTCDD to 100% would be that all coins issued transfer from all n outstanding funded wallets to a single destination wallet in one coordinated effort over a single 10 minute interval. This would necessarily involve at least n transactions. It is also well established that some BTC have been permanently destroyed, so this is technically impossible, not to mention that the block reward for that block would throw things off by 50 BTC.

---

Moderator: Please migrate to thread https://bitcointalk.org/index.php?topic=6172.msg90789

There is a question about BTCDD aka CoinDD at the Bitcoin StackExchange site:
 http://bitcoin.stackexchange.com/q/845/99

I note that there seems to be a discrepancy in the percentage calculations between the chart shown in the wiki and the value calculated by ABE.  E.g. the ABE result for block 131400, which was also around June 17 2011, was "Cumulative Coin-days Destroyed: 35.1672%", but the value from the wiki graph is about 28% on June 17. The corresponding value on the banana.mine.nu graph is about 900,000,000 bitcoin days.  More info and links to the details are on the StackExchange site.

Does anyone know exactly how the values are calculated by any of these three sites?

I note also that the percentage value seems relatively flat over the last quarter, at least according to ABE.  They
 report it as 36.1245% as of block 145677 at 2011-09-17, vs 35.1627% on June 17.

Does anyone know of a regularly updated graph by percent, rather than the graph of the raw numbers at http://banana.mine.nu/daysdest.html

Your friend is pulling your leg.  The Internet is two things - data and delivery.  They may have a good percentage of the data available, but they don't have the pipes running to every corner of the world.

Now that we have what amounts to high-frequency trading on Mt. Gox, transaction volume isn't that interesting. There seem to be a lot of bots moving the same money around in search of some small gain per transaction. Here's a recent seqblock of trades from Mt. Gox:

Sep 17, 2011, 18:24:00   4.79009   0.24   
Sep 17, 2011, 18:24:00   4.79019   0.01   
Sep 17, 2011, 18:24:00   4.79032   0.04   
Sep 17, 2011, 18:23:59   4.79049   0.01   
Sep 17, 2011, 18:23:59   4.79049   0.01   
Sep 17, 2011, 18:23:59   4.79070   0.03   
Sep 17, 2011, 18:23:59   4.79084   0.01   
Sep 17, 2011, 18:23:59   4.79100   0.01   
Sep 17, 2011, 18:23:58   4.79119   0.02   
Sep 17, 2011, 18:23:58   4.79139   0.01   
Sep 17, 2011, 18:23:58   4.79154   0.02   
Sep 17, 2011, 18:23:58   4.79157   0.04   
Sep 17, 2011, 18:23:58   4.79157   0.02   
Sep 17, 2011, 18:23:58   4.79157   0.04

BTCDD is proposed as a measure of monetary velocity for bitcoins: that is not the same as transaction volume so I am not sure to be posting on the right thread.

Arithmetically it seems to me that the percentage cannot exceed 50%: in fact it should reach 50% asymptotically.
If x is the average of the last block number seen for any bitcoin, the percentage is (x- N/2) divided by N (N being the current block number).
x should be close to N as velocity increases.

I'm not sure your math is right. Since a theoretical series of transactions which move 100% of the existing bitcoin to some previously-unknown wallet address would reduce the BTCDD figure to 0%, I don't believe the figure is asymptotic to 50%. Please explain further.

Even if you consider that there are lost coins, the theoretical BTCDD figure approaches zero as the number of days goes to infinity.

You meant "increase the BTCDD to 100%". Your previous post in this thread was correct in this regard.


If there are lost coins and a finite total number of coins then the BTCDD percentage cannot possibly reach 100% ever.
If there are no more transactions then the BTCDD tends to 0%.
In the long run, if there are any transactions then the BTCDD has a lower bound given by the proportion of the transacted coins out of the total coins.

Both statements are incorrect.

ByteCoin 

OK, so I reversed 100% and 0%, you're absolutely right. But it seems like you realized that!

Sticking with the substance of the discussion: if the total transactions per day average 50% of the total coins outstanding over some longish period of time, then sure, it's asymptotic to 50%.

Let's do some math on that: assuming the BTC stabilizes at $5, the long term implication of the 50% asymptote is about 22M * $5 = $11M of Bitcoin transacted commerce per day or about $4B / year. That is 1/1000 the size of the VISA network at $4T/year, or about 0.1%. Not great. To reach 10% of the VISA network at a 50% asymtote, the network would require a BTC valuation of $100. It's not a bad goal!

The good thing is that there is no central planning agency for bitcoin.

Fast forward to after last coin has been mined.   Some miners may want a high transaction fee and will ignore transactions with a low fee.  However no ever miner will.  some will accept a lower fee, some an even lower fee.  some will accept any fee at all (even 0.000000000000001%) and some will simply accept fees to be a "confirmer of last resort".  Why will they do that?  Maybe they have a bitcoin business, maybe they are running an exchange, maybe they have a lot of wealth in bitcoin.  All those entities are best served by keeping the network functional.

So the amount of transaction fee can only determine how fast a transaction in processed.  As long as a single node (even yourself) is willing to process transactions for free (or for very low fee) all transactions will eventually be confirmed.

I think the idea is that the transaction volume will increase so that even a very small fee will provide enough revenue for miners.

eg.
10 transactions per block @ 0.01 = 0.1 BTC per block - not incentive to mine the block
5000 transactions per block @ 0.01 = 50 BTC per block - where we are now

If we have both low transaction volume and low fees then it could be a problem.

Even then it isn't a problem due to difficulty change.

Say in 2030 transaction fees per block (the product of transactions per block & average fee per transaction) are so low at the then current difficulty that they don't cover electrical costs.

The network won't fail.  Lots of miners will turn off their equipment and the hashing power of the network will decline.  Eventually difficulty will reach equilibrium with transaction fee volume.  Obviously a larger network both in terms of transaction volume and hashing power is preferable because it makes attacks more difficult but the network can reach equilibrium at any point.

Today all bitcoins have been mined at a steady rate of 50 BTC per block.
If none have been spent to date and all are spent today we would have the maximum of BTCDD (i.e minimum monetary velocity).
The percentage of DD for each bitcoin amount would be today's block number (N) minus the block number of the block generating it which we average by halving today's block number (because we have a linear growth rate), divided by N.
Hence the maximum percentage is currently 50%.
When the reward is halved next year from 50 to 25 BTC, then there will be fewer "young" bitcoins: the weighted average will get tilted towards the first bitcoins generated, increasing the maximum BTCDD percentage to 66% at the end of the next period (2016).

Regarding days destroyed, I have come to the conclusion that it could be useful to have a days destroyed chart that multiplies the days destroyed by the historic bitcoin price at the time of that block.  This would show a measure of how much relative value is actually moving on the blockchain, rather then just showing it by number of coins which have a rapidly changing value at this early time. This is essentially similar to how bitcoincharts allows us to look at volume in bitcoins or USD.

What differentiates bitcoins sent from one address to another vs. bitcoins "spent"? Isn't transfering coins from one address to another considered "spent"? If not how do you make a determination of the difference? Can you define what you consider "spent"?

That's pretty well the definition of "spent" for the purposes of this discussion.

Actually, it's more technical than that as you could have a transaction that spends coins from an address and sends them to the same address and it would (probably) still destroy some number of bitcoindays. What you're actually spending is one or more of what is commonly called a "TxOut" which is worth some number of bitcoins. An address can have zero or more spendable TxOuts. In order to spend a TxOut you need to supply an appropriate signature generated using the private key associated with that address.

ByteCoin

I came over this old discussion about BTCDD. Actually BTCDD is the most confusing chart on blockchain.info, I never had a good understanding of its meaning. What does this statistic really measure?

If someone had 100BTC for 7 days and send 1 BTC and 99 BTC returned to his wallet as change, how many bitcoindays were destroyed? 7 or 700?

I guess in this case it will be 700, because both coins were moved. But then this indicator still don't count the real meaningful transaction

It is 700.  With Bitcoin there is no way to know the "meaningful" tx.  Both the "real" tx and the change are simply outputs.

Nobody is saying BTCDD is a perfect measure it is just one insight.