The first count can be wrong, hence the "X
possible keys"
So it's normal than it's above the real number of private keys it imports
For instance with a new wallet (1 shown address + 100 pool addresses) I get that:
Starting recovery.
Read 0.0 Go in 0.0 minutes
Found 0 possible wallets
Found 0 possible encrypted keys
Found 159 possible unencrypted keys
The wallet is encrypted and the passphrase is correct
Importing:
Importing key 1/101:
You could compare a dump and a recovered wallet.
I'll check why it's so above though.
As for the second matter, please grep this instead:
"sec":
Because there is an "hexsec" entry (which should make the count 510)
I don't know where the 20 remaining occurences can come from