I can see a huge difference in AVG% between these two methods. Not really sure which one is more accurate (reflecting true amount) If I sell some of my shares in coin A and reinvest on coin B.
Example:
Buy 1 BTC for 100 USD in 2014
Sell 1 BTC 500 USD in 2015
Buy 1 BTC for 1000 USD in 2017
"
Average of all Assets" will calculate the average cost basis from all your purchases.
In this example: (100 + 1000) / 2 = 550 USD
"
Average of unsold Assets" will calculate the average ONLY from all unsold assets.
In this example the cost basis will be 1000 USD, as the first BTC from 2014 has already been sold in 2015.
Because the gain from
all your sales is already
included in your "
realized gain", I would recommend the new "Average of unsold Assets" method.
The older method ignores all sales and is just taking an average from all your purchases.
But both methods are fine, just in a different way.
It depends on how you would like to see your gain calculated

Best regards,
Dario