Bitcoin days destroyed - explained | Hacker News Reader

Bitcoin days destroyed - explained

bitcoin.stackexchange.com

Absolute rubbish metric, if anything it measures the absolute opposite of monetary velocity. The value is basically guaranteed to grow and grow even in the face of flat or declining economic activity

Monetary velocity is the sum of all transactions in a period divided by money supply Vt = nT / M

Imagine a model where the number of bitcoins is increased by one each period, and one random bitcoin is transfered to another party as economic activity

                         day 1 day2                                   day 10
  Total Money Supply:      1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.0
  Actual Transactions:     1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
  Mean BTC days destroyed: 0.00 1.00 1.33 1.66 2.00 2.33 2.66 3.00 3.33 3.67
  Monetary Velocity:       1.00 0.50 0.33 0.25 0.20 0.17 0.14 0.13 0.11 0.10

As you can see monetary velocity goes in the exact opposite direction of bitcoin days destroyed and doesn't follow any similar curve.

Whats worse is it doesn't have a consistent relationship with any of the real measures as you change your assumptions. Imagine instead you model a situation where every coin turns over every period. In this case Total money supply, actual transactions and BTC days destroyed will all be the exact same thing, instead of the earlier wide divergence. Monetary velocity will be exactly 1 each time because total transactions will equal total money supply.

BTC days destroyed also has a hard upper limit at total days created which means it can't measure very quick activity accurately (all money turning over once a day will be identical to it turning over 10x a day) and will over weight economic growth after a period of low activity.

If you look at the graph pictured on the post, copied from the bitcoin wiki, it does not always "grow and grow", because it is in fact graphing the percentage of the total bitcoin days, not the absolute number. This is further described on the stackexchange post, and you'll note that my answer there explains that it has been flat over the last 3 months.

I agree that it is quite different from a velocity measure, simply because it was not intended to be a raw velocity measure. Bitcoin has some unusual features which require unusual metrics. This is just one that is being explored, as an alternative to raw transaction volume, as the posts explain.

here is the graph that I saw linked, that I used: http://banana.mine.nu/daysdest.html

the scale appears to exceed 110,000,000 - I don't believe that it is a percentage.

I understand the desire to remove the noise of transfers that are not actual economic activity, but I think this is a poor way to go about it. With bitcoin it is almost as if you had the ability to record every time a dollar bill went into a pocket. Sure, some of that is economic activity but it can also just be someone changing their pants. Even if economists had that data available they'd likely still rely on the business and financial institution reports to gauge economic activity. In this way raw BTC volume is a bit of a red herring, at least until you can produce a good view of economic activity and see how it relates to BTC volume or BTCdd.

As BTC continues to have an unstable price, a large percentage of the real economic activity will include exchanges for other currencies or value stores. Since exchange volume is quite centralized, you'd only need to have the cooperation of a few exchanges (you'd want to filter out traders) to gain what would probably be a rather accurate picture of real economic activity in bitcoins.

This metric seems to be just as flawed as total bitcoins transferred. Why should a bitcoin that hasn't been spend in a while be worth more than one that was spent yesterday? If someone sells something for bitcoins and then buys something else the same day, that seems like a great indicator of bitcoin activity but the second transaction is completely ignored by this metric.

It's also interesting to note that if you were to purchase 100 btc per day of stock using day-old coins, it will appear to be the same as 10 btc per day using 10-day old coins.

Sure, but if you ignore the trickle of new coins you have to get those 100 from someone else, and that requires burning significant amounts of bitcoin days.

At around a 40% increase in the money supply in the coming year, it isn't so much an insignificant 'trickle'.

I agree. I don't want old money / new money culture in Bitcoins, I hate that attitude among people when dealing with conventional currency.

But that's the reality of bitcoin. A tiny group of early adopters are supposedly sitting on a huge percentage of the total coins. Meanwhile, a much larger group of late-comers are mining small amounts and trading in small volumes.

I'm not sure if this metric is useful, but it is interesting to see if the early users start to 'cash out'.

If you were to monitor velocity of money http://en.wikipedia.org/wiki/Velocity_of_money naively, by calculating total bitcoins transferred, then it would be trivial for any party to increase the velocity arbitrarily high, by just sending their 10 bitcoins through a very long cycle of addresses. And there have been examples in the block chain, where parties unknown have moved large amounts of bitcoin through many accounts, for no obvious reason. Bitcoin-days-destroyed at least seems to avoid that problem.

I agree its still a flawed metric. Your example of the same bitcoins being used in multiple transfers, in the one day, seems completely valid to me.

But I don't think there's a perfect way of measuring the velocity, unless you've a complete map of addresses-to-identities, which isn't available (and even that is neglecting transactions done within exchanges, that aren't backended onto the bitcoin system - which is probably a fair enough limitation.)

The same is true, if you are trying to calcuate velocity of money in any traditional economy, where cash transactions are permitted. The cash could be used in several transactions between observations - theres no way of telling - and theres nothing that can be done to solve that. So, such measures are always going to be approximate.

On this topic, something that surprised me recently, was the assertion in Paul Krugmans blog http://krugman.blogs.nytimes.com/2011/09/07/golden-cyberfett... "The actual value of transactions in Bitcoins has fallen rather than rising. In effect, real gross Bitcoin product has fallen sharply. "

I'd love to know how he came to that conclusion, or what sort of data he used, and whether its accurate.

The data isn't from Krugman but was in the article he linked to.

Ok - it looks like they reached that conclusion by multiplying the number of Bitcoins transacted, by the dollar value of a Bitcoin.

Fair enough, but its a measure vulnerable to the problem that days-destroyed is trying to solve - it'd be easy for people currently sitting on large volumes of Bitcoin to 'solve the problem' that is highlighted by Krugman, by just shuffling their coins around.

Hanging around certain bitcoin corners of the internet, i get the distinct impression that a good section of users (or, speculators) are keen to spin reality to distort the image of bitcoin prospects.. either through hopeful delusion, or (more likely IMO) as attempts to hype up the market.

This metric may deal with one issue, but it is still totally useless as an objective metric of 'the bitcoin economy'. A much more reality-based metric would be the number of retailers (etc) accepting payments in BTC (and their volumes if they were available). But this would be a much more sobering picture.

As a very early adopter (probably in the first 100) and author of DiabloMiner, and forum admin... I have exactly no idea what this is supposed to prove.

The only metric that seems to be valid is "are new people using Bitcion?" The answer is yes, thus Bitcoin is still here.

Can someone explain to me the Y-Axis here? I don't can't figure out the denominator of that percentage. I assume the numerator is "days destroyed". Is the denominator the total number coin "age days" at the start of a given day?

The graph is from this wiki page, which explains the axes slightly better: https://en.bitcoin.it/wiki/Bitcoin_Days_Destroyed#Graph_of_P... : the x-axis is time measured in blocks (created on average once every 10 minutes).

The percentage is (bitcoin_days_ever_destroyed / bitcoin_days_ever_created). Today there will be ~7.2 million bitcoin days created, because that is how many BTC are in circulation.

I am reminded of the weird financial metric that Groupon invented...

Reply on news.ycombinator.com