rpietila Wall Observer - the Quality TA Thread ;)

[SlipperySlope]

Using a two year version of the chart, I read 3.2x per year transaction growth

Interesting.  You are claiming that usage statistics (like transactions) are increasing at about 3.2X per year, yet price is increasing at roughly 10X per year.  If we assume that usage stastics are linearly proportional to the number of users, N, perhaps what we are seeing is Metcalfe's Law play out in front of our eyes:

Metcalfe's law states that the value, V, of a telecommunications network is proportional to the square of the number of connected users of the system, N:    V ~ N²  .

According to Metcalfe's Law, the value of the bitcoin network increases by a factor of 10X each time the number of users increases by sqrt(10) = 3.16.  If you remove the volatility in the price action, we get a pretty good fit!

Apparently there is some controversy whether Metcalfe is right with regard to N squared. N * log (N) has been proposed as an alternative. However I agree that your observation is compelling.

[1715067630]

1715067630

[1715067630]

1715067630

[1715067630]

1715067630

[Peter R]

Here is the square of the number of TXs per day (excluding popular addresses) overlaying the price history and shown on a log scale.  I've shifted the N²_TX curve for easier comparisons.  Indeed, Metcalfe's law appears to hold (the two curves tend to have the same slope on the log curve at any given point in time).    

[SlipperySlope]

Here is the square of the number of TXs per day (excluding popular addresses) overlaying the price history and shown on a log scale.  I've shifted the N²_TX curve for easier comparisons.  Indeed, Metcalfe's law appears to hold (the two curves tend to have the same slope on the log curve at any given point in time).    

Awesome chart that well illustrates your discovery. May I refer to it in my logistics model thread?

[Peter R]

Apparently there is some controversy whether Metcalfe is right with regard to N squared. N * log (N) has been proposed as an alternative. However I agree that your observation is compelling.

If it's N log N then we are presently overvalued:

[Peter R]

Awesome chart that well illustrates your discovery. May I refer to it in my logistics model thread?

Thanks for the compliment!  I've been meaning to check this relationship out for months.  Only took me 15 min lol.  Yes, please feel free to use it however you like. 

[Pruden]

Y'all are using blockchain.info charts, which are broken! Compare the thumbnail in the charts menu to the actual chart. They changed, some of them still lack data and in many cases they paint a very different picture to before their downtime last week. I already asked about it on their thread but got no reply yet.

"Number of transactions excluding popular addresses" still gives very bullish overview, because whereas during the 2013 bear market it stayed flat after the quick drop from the peak, during this bear market it kept growing after the peak and is close already to surpassing it.

And the transaction/trade ratio was clearly based on MtGox exclusively, which makes it look like there is much less speculation in relation to commercial transactions than it should.

[Peter R]

Y'all are using blockchain.info charts, which are broken! Compare the thumbnail in the charts menu to the actual chart. They changed, some of them still lack data and in many cases they paint a very different picture to before their downtime last week. I already asked about it on their thread but got no reply yet.

"Number of transactions excluding popular addresses" still gives very bullish overview, because whereas during the 2013 bear market it stayed flat after the quick drop from the peak, during this bear market it kept growing after the peak and is close already to surpassing it.

The new (broken?) versions of the TX per day chart appears to be a noisy version of the previous (non broken?) chart.  The underlying pattern is the same (minus the noise) I think.   

Does anyone know a secondary source from which I could download the number of transactions per day excluding popular addresses?

[Queeq]

I don't think that number of transaction depicts adoption rate. It is a boom of altcoins now and they are mostly traded with BTC. So there are lots of them moving back and forth inside and between the exchanges.

[Peter R]

I don't think that number of transaction depicts adoption rate. It is a boom of altcoins now and they are mostly traded with BTC. So there are lots of them moving back and forth inside and between the exchanges.

I'd like to repeat this analysis with more observable variables that could be proxies for the N in Metcalfe's Law.  I think number of bitcoin addresses used per day, and number of blockchain.info wallets are both decent proxies as well.    

In your opinion, what is the most accurate way to estimate N?

[Queeq]

I don't think that number of transaction depicts adoption rate. It is a boom of altcoins now and they are mostly traded with BTC. So there are lots of them moving back and forth inside and between the exchanges.

I'd like to repeat this analysis with more observable variables that could be proxies for the N in Metcalfe's Law.  I think number of bitcoin addresses used per day, and number of blockchain.info wallets are both decent proxies as well.    

In your opinion, what is the most accurate way to estimate N?

I'm not sure it exists. It should be something that is connected only to the amount of new fiat flowing into cryptocoins. And even that does not necessarily correlate with adoption rate.

[chessnut]

....Last night bitcoin went up..... while the stock market went down.

 

[SlipperySlope]

....Last night bitcoin went up..... while the stock market went down.

 

There does appear to be a rally developing which can be clearly seen on the Huobi and Bitstamp one hour charts.  Support held at $550.

[Biodom]

Awesome chart that well illustrates your discovery. May I refer to it in my logistics model thread?

Thanks for the compliment!  I've been meaning to check this relationship out for months.  Only took me 15 min lol.  Yes, please feel free to use it however you like.  

Very nice charts, indeed, but they suggest current fair value of $70-300  

[Peter R]

Awesome chart that well illustrates your discovery. May I refer to it in my logistics model thread?

Thanks for the compliment!  I've been meaning to check this relationship out for months.  Only took me 15 min lol.  Yes, please feel free to use it however you like.  

Very nice charts, indeed, but they suggest current fair value of $70-300  

The N log N chart suggests a low value, but the Metcalfe Value chart (N^2) would be near $500 now.  Also note this is MtGox price data and only includes data to February 5, 2014.  The Metcalfe value has been growing for another month and a half. 

I should do some nicer plots where I merge the MtGox (past) price with the BitStamp (current) price so that we can track the Metcalfe Value in real time.  I should also include the calendar date so that it is less confusing! 

[windjc]

Right, TA is better than FA because bitcoin is a stock. Or a commodity. Or a stock. Or a commodity. Not a technology. Got it.

BTC is a currency. Bitcoin is a technology.

Charts show you what is going on regardless of what you trade, currency/commodity/stock/etc makes no difference.

What was your point again (if you have one)?

My point is obvious. Stocks and commodities do not grow exponentially.

[N12]

Right, TA is better than FA because bitcoin is a stock. Or a commodity. Or a stock. Or a commodity. Not a technology. Got it.

BTC is a currency. Bitcoin is a technology.

Charts show you what is going on regardless of what you trade, currency/commodity/stock/etc makes no difference.

What was your point again (if you have one)?

My point is obvious. Stocks and commodities do not grow exponentially.

So you think Bitcoin is the only thing in the world where you can draw straight lines on a log chart?

[windjc]

Right, TA is better than FA because bitcoin is a stock. Or a commodity. Or a stock. Or a commodity. Not a technology. Got it.

BTC is a currency. Bitcoin is a technology.

Charts show you what is going on regardless of what you trade, currency/commodity/stock/etc makes no difference.

What was your point again (if you have one)?

My point is obvious. Stocks and commodities do not grow exponentially.

So you think Bitcoin is the only thing in the world where you can draw straight lines on a log chart?

Nope. Not at all. But I do think that Bitcoin continues to surprise typical traders with its violent upswings.  Going down, it all seems to make sense to them. Bottoms and tops, not so much.

[chessnut]

Why is it so hard to understand that fundamentals and TA are both significant?

so we know Bitcoin is fantatic, love it, great fundamentals. but have we been beaten to the scene? is it in a bubble? is there a large unknown seller in the market? what is the subjective value of a bitcoin? what is the rate of growth? those are questions that Fundamental analysis cannot answer easily on its own.

[Peter R]

Here is an improved graph of the bitcoin price history versus the Metcalfe value.  I am now using the price history from bitcoinaverage.com (giving me post-Gox prices as well), and I am using both the number of transactions per day (excluding popular addresses) and the number of unique bitcoin addresses used per day as the N in Metcalfe's Law : V ~ N^{2 .}

I would like to do a proper regression to get a numerical model, but the blockchain.info charts currently have holes in the data (the large spikes visible on the curves) that render such a processes useless.  For the time being then, I've just eyeballed the fits as you see below.  

The current price of bitcoin is slightly higher than its Metcalfe value.  However, should the network continue to grow at its historical rate of 3.2X per year, the Metcalfe value should exceed the current bitcoin price later this spring.  

[dnaleor]

Here is an improved graph of the bitcoin price history versus the Metcalfe value.  I am now using the price history from bitcoinaverage.com (giving me post-Gox prices as well), and I am using both the number of transactions per day (excluding popular addresses) and the number of unique bitcoin addresses used per day as the N in Metcalfe's Law : V ~ N^{2 .}

I would like to do a proper regression to get a numerical model, but the blockchain.info charts currently have holes in the data (the large spikes visible on the curves) that render such a processes useless.  For the time being then, I've just eyeballed the fits as you see below.  

The current price of bitcoin is slightly higher than its Metcalfe value.  However, should the network continue to grow at its historical rate of 3.2X per year, the Metcalfe value should exceed the current bitcoin price later this spring.  

just a small thing: I guess you need to multiply the price of a bitcoin with the number of coins.
The value of a network is the market cap, I suppose...

If you do that, I guess the large "spread" at the left side in the chart will become smaller.