Quote from: rpietila on April 26, 2014, 02:44:16 PM

Quote from: wachtwoord on April 26, 2014, 02:24:01 PM

Scrolling back however (see: https://bitcointalk.org/index.php?topic=316297.msg6337742#msg6337742) I see that the 13k people that own 100-1000 BTC own a total of 3.0 BTC. This is entirely between the lower and upper limit if 1.3M and 13M. So I don't understand boumalo's problem either Tongue

Correct. This is my best estimate of the actual situation.

Quote from: wachtwoord on April 26, 2014, 02:24:01 PM

I don't understand what you mean either.

I try to explain my post below:

Quote

Quote

Quote from: rpietila on April 22, 2014, 12:24:00 PM

The new information is in the distribution parameters that were updated based on the Mt.Gox data.

The Mt.Gox leak was the data after accounts were frozen right? Considering there were many hints of Mt.Gox's impending failure well in advance (goxing was a verb a long time ago), I expect a large percentage of people withdrew their funds either completely or mostly (my case). Therefore, I am not sure how useful raw Mt.Gox numbers would be for finding the overall distribution.

20 pages back in this thread, when the methodology was created, we used the distribution data from a certain company called Silvervault as a basis for some assumptions of this model. Even impaired Mt.Gox data is an improvement. An example problem where we have better answer now, is:

1.0 million coins belong to people with 100-1000 coins
There are 45,000 people with 10-100 coins.

How many people own 100-1000 coins?

The bolded part is what the words "example problem" (above) refer to. It is a riddle, which is used as an example. I have given two statements which are assumed to be true, and one question, which the reader must answer (but cannot, as from his point of view, there is not enough hints). At face value it does not have anything to do with the real thing (it is an example, remember), but what it illustrates is the "new information" that we found in Mt.Gox leak.

I totally lack terminology to explain what this info is, but it can help us know the average of coins per holder in each of the brackets. Previously we had only 46 people with 10k+ coins, now this number is 70 due to this more accurate parameter. This is a great improvement in accuracy.

Thanks for this. Are you able to estimate how many people are in the 10-25 and 25-50 BTC ranges?