I rotate between three different characters depending on the day, situations, match ups. But it's not so much of the fact that I think top players counter pick to win, it's because I'm used to counterpick info to win/do better in other aspects of my life. It's been proven to help.
I've mentioned this before in this thread but I play baseball. I'm a catcher and I also pitch. A pitcher doesn't toss one pitch again and again in hopes to overpower the batter (unless their really good, Mariano Rivera good). It's an effort in futility especially when you can learn another pitch which I'd
equate to learning another character. The dedication and time in practice to toss a pitch consistently is equivalent in my experience. Counterpicking is a viable strategy. Proven to be effective, it just takes longer to see identical success but the depth of success is larger. Why on earth am I gonna be road blocked by a (or many) specific match up(s) or more, when I can just learn another character and reap those rewards.
I'm not gonna toss a fastball against a fastball hitting team, I'm gonna need to toss a change up or slider every once in a while.
Just to be clear, I think counterpicking is a valid strategy, and that if you can learn multiple characters at a high level it's only to your advantage. My issue is rather with the way that lesser players potentially learn the wrong lessons from counterpicking, where they treat it as a kind of rock-paper-scissors scenario and stifle their own development and knowledge in the process.
The classic example is FG Link, throwing projectiles all day long. Someone loses to Link's projectile spam, and they run to a character with a reflector, another projectile-heavy character, etc. However, there's probably more that they could have done with their own character, had they not given up and ran to the obvious counterpick. Instead of exploring all of the options available, both general and character-specific, and learning fundamentally how to fight projectiles, they go for the easy solution. That player who lost to Link ends up only getting surface-level knowledge, and doesn't actually learn how to play a matchup, or indeed what could be changed about their existing play that would help them win better.
While results matter, I think that character specialists (not necessarily loyalists) have an important place in the metas of games because they're the ones most willing to develop individual characters and to push them to their limits. While Armada has made the switch to Fox over the past year, where would Peach have been without his dedication to the character? People were crying about the Luigipocalypse, only for Boss to come in and show just what it means to play Luigi and not rely on easy damage. DKWill takes his DK knowledge from across various games and implements it in DK's horrifying new ding dong form.
Notice the examples of players I gave all vary their character use, so it's not like I'm decrying people who use multiple characters. However, they pushed their characters to the utmost, and their decisions to switch to Fox or Sheik or whatever comes from intimate knowledge of themselves and their character. There's a substantial process that leads to that decision that players learning might not understand, which leads them to thinking, "RPS."