Quote from: nicepumper on September 06, 2014, 05:49:11 PM

Incredible graph! but no matter how many graphs you have, all these fuds like fallling will not shut the fuck up Roll Eyes

Noblesse oblige... Grin

Quote from: oda.krell on September 06, 2014, 03:18:41 PM

Quote from: JorgeStolfi on September 06, 2014, 06:52:49 AM

[...]
Anyway, those blockchain.info plots are rather strange. They don't show increasing adoption. I don't know what they are showing.

Transactions cost money. There is a hypothetical incentive for a hypothetical manipulator to "pad" the growth (e.g. a miner intending to sell at inflated prices), but the cost of faking those transactions is substantial,

The total fees being paid now are only ~10 BTC/day (~5000 USD/day), which divided by 70000 tx/day gives only 0.00014 BTC/tx, or little more than the minimum fee. If one is tumbling 680'000 stolen MtGOX coins, a loss of 10 BTC/day is nothing. If one is expecting to sell 200'000'000 USD of fund shares, one may be willing to spend 1% of that to simulate a vigorous bitcoin economy for a year.

Quote from: oda.krell on September 06, 2014, 03:18:41 PM

The no. of transaction graph roughly matches total USD volume transferred roughly matches the no of addresses graph roughly matches the hash rate graph. If the data suggests (even though it is no guarantee for) exponential growth along several dimensions, the most likely conclusion, in the absence of contrary evidence, is that some underlying variable does undergo such growth in fact. [ ...] Looking at the blockchain data however and questioning whether there was exponential growth until now is a stretch. [ ... ] the most pertinent graph. From 100 transactions a day to almost 100k transactions a day, over the course of 5 years.

All bitcoin indicators grew enormously over those five years. I would question whether it is appropriate to call it 'exponential growth' and then extrapolate to the future, but that is not the point.

Quote from: oda.krell on September 06, 2014, 03:18:41 PM

Hence, the conclusion based on the least assumptions is that the data is not fake, until evidence to the contrary is presented. (and, "the data doesn't look like commercial adoption" is no such evidence).

Like the Metcalfe Law plot that others have posted, the above plot squeezes the data for the last year into a tiny corner region of 3 by 0.2 cm at normal resolution. One must look closely at that corner to notice that the number of transactions has not increased substantially over the last year. One must look even closer to notice that the tiny wiggles in that plot are swings by 40% or more.

Now look at this plot of the estimated USD transaction volume per day, for the last year:

"Estimated" here apparently means "excluding the return-change outputs" as identified by some heuristic.
Note that the daily transaction volume in USD mostly decreased from ~Jan/01 to ~Mar/07, and was approximately constant from ~Mar/23 to now, apart from a couple of peaks/bumps (~Mar/07--Mar/23, ~May/22--Jun/20) and a depression (~Apr/27--May/21). This is quite at odds with the "increasing adoption" that was supposed to be happening during this time, which should result in steadily increasing USD volume.

The plot of the estimated BTC transaction volume per day is perhaps more revealing:

Note that the daily transaction volume in BTC was even more constant, arount 100'000 BTC/day, from Jan/01 to this day. (While the March peak is more striking in this plot, the other bumps and dips are somewhat smaller.) It is puzzling, for example, that the BTC volume on Jan/01 was almost the same as on May/01, even though the price dropped by 50% in the meantime. I take this fact as evidence that most of the traffic is not payments, but activities for which the price is irrelevant -- such as tumbling.

The plot of number of transactions per day is also strange:

Note that this quantity too is fairly constant, between ~60'000 and ~70'000 tx/day, from ~Feb/10 to this day. (The vertical scale of this plot starts at 40'000 tx/day, so the relative variations are smaller than what they appear.) The March peak of the other plots is here an inconspicuous bump of about 8%. (Therfore, that peak was the result of larger transactions, rather than more transactions.) While the increase of ~20% in January could signify increased adoption, the stagnation since then again contradicts it. (Increased adoption should cause more transactions as well as more USD volume.)