Insanely well. I've been posting publicly as dmd@3e.org for 16 years now, and at a few other addresses that forward to that for 22. All of that mail goes to gmail. My spam folder gets ~600 messages per day. I get a false-negative (spam in inbox) about once a month. I check for false-positives (ham in spambox) pretty regularly, but haven't ever found one.
I use Gmail, but I definitely get spam. I get false-positive spam all of the time (mailing lists I'm on have a decent chance of going to the spam box), and whenever a spammer has 'Google' in the email title, it has about a 60% chance of being a false-negative (i.e. showing up in the inbox). I also get false-negatives on spam when there is just a title and message body.
Perhaps Gmail's spam filters are mostly based on how you train them, so people who receive more email get a better filter?
Who cares about how much it filters out? The question is how much false positives does it have? And google most definitely are not good enough to even be usable in that department. If you never know if you've missed something you have to constantly check the filtered mails as well, making the spam filtering much worse than nothing.
If you rely on something as blunt as gmail spam filtering you don't take emailing seriously.
Very interesting! I wonder if spammers steer clear of @gmail.com addresses to prevent Google's algorithms from learning about their spam. (If so, they probably want to steer clear of GAFYD addresses, too, but maybe they don't know how to identify them).
At my last job (large, successful startup) I got the wildcard mail and got close to zero spam. Gmail spam filtering definitely works very well.