THANKS! I wonder about the data, perhapse there is a couple of bugs?
1) The sum of the first column (number of address ever in each bin) is 48,297,867 and that seem right. However the sum of the second column, the number of active addresses, is only 345,558. That seems low to me. That would indicate that all of the BTC in existence are stored on only 345,558 unique addresses?
2) This is a nit but there are only 195,111 rows. The last row for bin 1zzz is missing.
I did a quick sort on the number of currently active addresses in each bin and the top ten are:
Index Bin Ever Now
------ ---- ----- ---
32543 1Ag6 852 106
32542 1Ag5 842 82
9852 13vs 760 71
32541 1Ag4 879 56
23402 17xV 666 51
32593 1Agx 837 48
684 11Co 45 35
35159 1BTC 1551 34
0 1111 374 33
1548 11Th 50 32
There does seem to be a lot of them in the 1Ag range but this is easily explained by Casascius because the way he created his coins skews the addresses into certain bins. For example he created 1786 addresses starting with 1Ag in this batch of coins alone:
http://casascius.uberbills.com/?type=1&status=active I also did a quick sort on the number of addresses ever in each bin and the top 20 are:
Index Bin Ever Now
------ ---- ----- ---
116804 1bit 12043 6
35159 1BTC 1551 34
68892 1MUo 1148 5
36069 1Bit 892 27
36712 1Buy 883 1
32541 1Ag4 879 56
32571 1Aga 869 4
32572 1Agb 867 8
32539 1Ag2 866 3
32592 1Agw 863 7
32538 1Ag1 859 7
32595 1Agz 856 5
32543 1Ag6 852 106
32594 1Agy 846 25
32542 1Ag5 842 82
32593 1Agx 837 48
75469 1PSC 834 9
66967 1Luc 819 16
37550 1CAR 813 13
34187 1BAS 807 31
Again, vanity addresses obviously skew the distribution.