They may not get identify all fake accounts, but there are still tons of accounts which can be clearly identified as fake.
Other platforms are of course more favourably disposed towards multiple accounts, automation assisted high volume posts, shifting identities and parody accounts
They do slightly, but only to the degree that their actual users get annoyed by all the bots. And people are surprisingly willing to put up with the bots, so.... yeah, I agree.
Edit: I need to point out that this should be optional. Anonymity should still be the default.
The other thing is - imagine I were a gay man living in Poland (a country that recently made europe-wide outrage headlines with "LGBT-free zones"), or in Russia where "advocating" for LGBT is a crime, or in Saudi-Arabia where you might end up literally decapitated. I could go and discuss topics regarding my sexuality on Twitter anonymously - but with a forced tie to a government ID?! I would have to fear authorities using my free speech to get me imprisoned or worse all the time!
Also, I never meant to suggest proof of ID would be mandatory to use the service! It would be an additional status you can get optionally, if you want to get extra credibility. If you desire anonymity instead, that should absolutely be your right.
At the moment, you can get a checkmark on your twitter but how you do it is a bit unclear. Technically you have to be somewhat noteworthy, but the bar seems to be really low. If this was an entirely user-initiated process, over time people would associate that status with higher credibility.