So the conclusion is that even if Carlsen beats all other players in long matches but fails to beat one of them, this shows that he isn't the best player. And if he wins 50% of the tournaments this shows that someone else that wins 10% of them may be better :-) The 50% outcome in the example above is slightly creative though, since tournaments Carlsen actually won on tiebreak are counted among the "lost" tournaments to make it a nice 50%.
First the 50% is based on when carlsen recieved the same exact score as someone else in the tournament I call the tournament a draw. Not a loss or a win.
My point is a chess match has long been held as a legitimate way to tell who is better. Ratings are an alternate way. But the problem with ratings is they really never lead to any big chess events to popularize chess. Its just more of the same one tournament after another.
Matches IMO are an equally valid way to tell who stronger between 2 players. At the end of the Kramnik Anand match if someone asked me who was playing better chess I wouldn't go look at their ratings. Anand just proved it over the board.
And note: I am likely one of Kramnik's best fans. It was hard medicine for me to take but I took it.
Well, that's sport. There's no such thing as a guarantee.