cr WPA2 how realistic with wordlists ?

As I undersand, brutefrocing WPA even with several GPU's is not really a good idea..

How many percents of the attemps to crack WPA with wordlists are normally successful?

Any feedback is appreciated very much. Thank you!

Depends on how good your wordlists are. You cannot estimate what the chances are.
I am no expert but some things to consider...

Know your target, list nicknames, pets names hobbies etc.

The main thing I look for when testing is if the user has changed the BSSID. I find that if they have changed the network name they "usually" change the password also. In a crazy twist they usually change it to something less secure than the original manufacturer one !

If the BSSID is left standard then they have probably left the password standard. Look on the net for default password patterns relating to the network in question and tailor your brute force accordingly.
The question shall be : how many people change the default WPA password ? Maybe 10% ?
And the default password can be 12-chars hexa, hard to bruteforce :/
From WPA cracking sites they report 40% cracked of all caps they get. Sounds a bit high for me.