Chess will never be solved, here's why

Some endgames can be 'perfect'.  And 'perfectly solved'.
Some of them are.  Both 'perfect' and perfectly 'solved'.
Is there any 'opening' that's 'Perfect' ?
Yes.
g4 e5 f4 - now how about Qh4# ...   Checkmate.
Looks kind of 'Perfect' to me. 

#990
There is some misunderstanding:
no improvement for white means: no improvement at any of the white moves. 
In the game Carlsen-Nepo, they both followed their engine preparation and then landed in a forced 3-fold repetition. I did not find any white move to blame.

Hypothetically - if somebody ever proved that chess is always a draw with best play by both sides - would that mean that chess is then 'solved'?
Related:  If chess is thoroughly 'solved' would it then be known as to whether its 'always a draw' with best play?
No.  Doesn't follow.  Some white first moves might force a win for white - or allow Black to win.  Some white first moves might allow a forced draw for either side.

@tygxc

 Oh, so you actually mean the best possible move played by the player at each position in the game? But then again, you don't know if every move has been perfect until chess has been solved and we have all the positions that will force a draw.

 I mean, you an engine may give you a line at depth X but unless you have all the positions from move 1 to move N for each player with best play for each move, you can't actually confirm it's been "best possible play" and "no improvement for white" in each move from move 1 to move N. 

 Just 1 small deviation from the correct solutions would not make it a perfect game, so we can't know it's a perfect game. 

 As I said, Im leaning towards NO, that game between Nepo Carlsen is probably not a perfect game.

#994
I wrote: "This is probably also a perfect game"
The probably means: I do not know, but I think so.
They both followed their analysis and played quickly.
Then they reached a forced 3-fold repetition.
What white move would you think can be improved on?
It is not necessary to first solve chess and then verify a game is perfect.
The other way around: a set of perfect games solves chess.

People who do a lot of chess tactics puzzles might better understand that things aren't what they seem to be.
'Best play' is often arbitrarily and invalidly decided on.  
'Shortcuts' for same are so often invalid - whether its a computer doing it - or a person.
Having a computer do invalid shortcuts partly defeats the purpose of the computer.

 Let me disagree on "It is not necessary to first solve chess and then verify a game is perfect."

 The only way to know if any given move can be improved on is by solving chess so that we don't have to rely on an arbitrary "depth N" engine analysis to judge whether the move is the best or not. Who's to say that "depth N+1" is not going to yield a BETTER move for white (or black) and change the line in its entirety? And then "Depth N+2" may change it one more time? You can easily notice this analysing a game at very low depths with stockfish, how it rapidly discards lines as the depth number goes up. 

 The only way to actually KNOW every move from move 1 till the last were the very best by both players is to first solve it, have a position-space of all the drawish games from move 1 with best moves for each move and each player, and then see if the game falls within that set.

 So going back to my first reply to you, home computers do not have enough processing power to solve chess - there's so many positions they can evaluate in 3 days if they work 24/7 - so any of those correspondence chess games is unlikely to be examples of perfect games.

#997
Carlsen, Caruana, Nepo did not use home computers, but cloud engines.
A cloud engine achieves 10^9 nodes/s, that is 1000 times faster than a desktop.
The last world championship matches make believe that the Petrov draws for black.
https://www.chessgames.com/perl/chessgame?gid=1937789
https://www.chessgames.com/perl/chessgame?gid=1937912

https://www.chessgames.com/perl/chessgame?gid=2127155

https://www.chessgames.com/perl/chessgame?gid=2135056

https://www.chessgames.com/perl/chessgame?gid=2136005 
Carlsen varied variations, as none got him anywhere.
The only win was a blunder by Nepo in a drawn position.

Post #998 would suggest that e4 e5 Nf3 allows a forced 'draw' by black replying Nf6 for his second move.  
'Make believe' ?  Yes - that does sound like 'make believe'.

If Carlsen plays the Petroff against Stockfish 14 he will almost certainly lose.

according to me chess will never be solved by any supercomputer because there are 10^120 games possiable (approx) it might take some million years 

that's just my thought

Couldn't find 'FIDE' ratings for chess engines and supercomputers.
Maybe there is no such thing.
Carlsen is in the high 2800's it seems.  Might break 2900 someday?

But some engines/computers have an Elo rating.
I saw more than one over 3400.
Elo is the same as FIDE ?  Somehow I doubt that ...  but its a long time since I researched the rating systems.
But if the strongest computer chess is 500 points stronger than Carlsen - then maybe they could even spot him material !
Maybe that can be looked up too !  

World class players struggle against top engines with odds.

Apparently.  
Here's something from seven years ago.
In the last two games of the four - 
Nakamura was given a pawn and white - but apparently  still finished behind with 1.5 to 0.5 ...  same result in the first two - with no pawn.
https://www.chess.com/news/view/stockfish-outlasts-nakamura-3634

Stockfish is a lot stronger than it was then! The big step was adding neural network enhancements (after AlphaZero's success), and there has also been a sizeable distributed computing project, with about 5 million hours of time donated to test improvements.

he

See post #19 here https://www.chess.com/forum/view/general/chess-will-never-be-solved-heres-why?page=1 

Please try not to propagate the myth.

You don't know if there is no improvement for white (or black) until chess is solved . 

You've always had the cart before the horse, which is why you don't get anywhere by pretending Stockfish can discern when there is "no improvement" to be made.  Clearly, there is no way to even posit that Stockfish can discern best play, since each successive release of Stockfish wins more often against previous versions. 

What does it tell you that meaningful improvements are still coming out for Stockfish every few weeks or months?  What does it tell you that Alpha Zero turned the chess world on its ear after only 4 hours of playing itself?  Because those things seem to clearly indicate that humans and engines are nowhere near reaching the zenith of best play...so, trusting Stockfish to take you to Tablebase Town accurately is folly.

You seem to be criticising tyxgc in general rather than specific terms (as in "you've always had the cart before the horse, which is why you don't get anywhere by pretending") and yet Elroch posted yesterday to tell us that ty is completely correct, apparently in all points, according to Elroch. I believe you also criticised me for arguing in far less general terms than you against ty. What gives? Incidentally, this phrase "until chess is solved" is rather glib, since there's nothing to indicate that it will ever be "solved".