The curious thing is that Carlsen’s ELO didnt’t get nibbled even after all theese consequent draws. They’re too close.
For example the requirements for the most common way to earn the GM title include a requirement that your rating has reached 2500. The live rating is used for that.
I always believed it is because either FIDE ratings are only updated once a month (see 7.1[1]) or because their ELO is too close, therefore the prob. of winning is 0.5 for both anyways.
[1] http://www.fide.com/component/handbook/?id=172&view=article
I heard (possibly from the Chess24 commentators) that if their difference in rating was 4 or more, each draw would bring them closer together in live rating - Magnus would lose rating with each draw; but as it's 3, their ratings don't change. i.e. anacleto was right, thomasahle wrong.
Also, Elo isn't an acronym (ELO) as one might think - it's named after its inventor Arpad Elo. https://en.wikipedia.org/wiki/Elo_rating_system
Interestingly, when I first looked at this page, most of the good/correct comments were voted down, most of the bad ones not, it was weird. (It made me suspect that other HN pages are like that but I don't know it! - I know chess better than most subjects discussed on here) But that has now almost totally been corrected.