What surprised me about molecular's analysis though is the amount of coins that were never moved. I wonder if that is the usual trend, or it shows a different behavioral pattern for first comers. Has anybody made a "the number of coins that haven't moved since" graph?
Been thinking about that one or a similar analysis (coin age) and thanks to your post, finally came around

I did this by identifying the leaf nodes of the transaction chains and then determining the time of the block these were included in. That's the "last time these coins were moved".
CAVEAT: I know I'm making the mistake of including blocks from orphaned chains here. Still looking for an easy way to exclude these. The error should not be too huge, though.
select year(from_unixtime(block.block_nTime)) as year, month(from_unixtime(block.block_nTime)) as month, sum(txout_value)*1E-8
from block_tx
inner join block on block.block_id = block_tx.block_id
inner join tx on block_tx.tx_id = tx.tx_id
inner join txout on txout.tx_id = tx.tx_id
left join txin txin2 on txin2.txout_id = txout.txout_id
where txin2.txout_id is null
group by year, month;
Made a pretty histogram of the result:
like it? even 0.01 BTC help continue cool queries: 1SQL2R5ijCn2ZiBG8aDYpCiecNPuxzMJJ
Note: 2009 summed up = 1,205,454 - all summed up: 7,087,766
Now, what conclusions can we draw from this?