Chess is finite game. The number of moves of any game is finite. So you can, in principle, check out for yourself if a given move can force a win or not. If the position is such that there is no way to force a win, then there is no intrinsic advantage to the move the computer suggests. (Remember the computer evaluates on the basis of an evaluation function that is defined by the programmer, a man. Its parametrization comes from our chess lore and the practical experience programming chess engines the programmer may have.)

My answer to your first question is "Not Necessarily". A lot depends on the chess engine used, the depth that it was allowed to search, and the "cleverness" of the engine's heuristics in selecting which branches of the search tree it searches. And, regardless of the circumstances, any engine is quite capable of making mistakes.

Chess engines are different and some are simply better than others in evaluating the position resulting from the candidate moves they investigate. All things being equal, the deeper an engine is allowed to search the more confident we should be that it will find the "best" move since it will have looked at more potential positions. But all things are never equal. No chess engine, regardless of how powerful a computer it runs in, can look at all the moves in its search tree at the depth required to establish confidence in its evaluation in a reasonable amount of time. So engines restrict (prune) the branches of the search tree that they look at in depth according to some heuristics (a.k.a. "guesses"). The more aggressive an engine prunes its search tree, the deeper it can search in the same amount of time (and therefore increase our confidence in its final evaluation of the position). But at the same time it increases the chances the branch of their search tree that contains the "best" move will not be looked at, so the quality of its search depends on the quality of its "guesses". So search engine developers must perform a number of tradeoffs and determine what approaches work best in the majority of positions.

Oh, yes, the majority of positions. Chess engines currently tend to excel in tactically complex open and semi-open positions, and they tend to do less well in strategically complex closed positions and endgames. To make it even harder to generalize, if two engines are used to analyze a position, the first one might find the "best" move and the second one may not. And, if they are given a different position to analyze, the second engine may find the "best" move and the first one may not. So whenever possible, I try to analyze a position using at least 3 different engines, and average their evaluations to reduce one engine's possible evaluation bias.

To try to answer your second and third questions, IMO the "best" move is the one that will cause the player on the move to reach the position's most likely conclusion in the smallest number of moves. I phrased my answer that way because as you pointed out in some positions a win can be demonstrated, in some positions a win is likely but not capable of being demonstrated, and some positions are just dead drawn and it's not reasonable to downgrade a human or a computer for their inability to find a winning move in that position, assuming equally good play from their opponent.

But I'm not really satisfied with that definition myself. If two moves give the same result, but one of the moves causes that result to be delayed one or more moves, is it proper to consider one move "better" than the other? After all, the end result was the same. And, if you enjoy making your opponent suffer (yes, there are some chess-playing sadists out there), you might even be able to argue that the move that prolongs the inevitable result the longest is the "best" move.

As you also pointed out the criteria for determining the "best" move can also include esthetics And as <Once> pointed out, the criteria can also include practicality. So Morozevich's 32...0-0-0 was arguably better practically than 32...d2 even though the latter may have lead to a win in a smaller number of moves. Not implying at all that Morozevich might have a sadistic streak in him; no, not at all. He was just being practical.

And what about a player who in a lost position plays a move that leads to complications that are the most likely to lead his opponent astray even though it's not the "best" move in the sense that the "best" move would delay the otherwise inevitable loss the longest (the flip side of my definition of "best" move)? Kind of like the horizon effect affecting chess engines. And what if as a result of that less than "best" move one's opponent goes astray and a certain loss is transformed into a draw or even a win? Who would then argue that the move actually played wasn't "better" than the objectively "best" move?

There is more than one game of chess. First and (probably) foremost there is the struggle. Two people over the board trying to beat each other. And here the best move is the one that wins. That might not be the absolute best move in theoretical terms. But chess is a struggle, a fight, a sport.

If you are winning, keep it simple. If you are losing, look for ways to complicate. And don't care too much about the "best" move.

Stockfish 17 says:

28. .. Ne4

(28. .. Ne4 (1. ... Ne4 2.Bxe7 Nxf2 3.Bb4 0-0-0 4.Re1 d2 5.Qxf2 d1Q 6.Rxd1 Rxd1+ 7.Kg2 Qxf2+ 8.Kxf2 Rxh1 9.Nf8 Rh2+ 10.Kf3 Rxb2 11.e6 fxe6 12.Bc3 Rc2 13.Bh8 Nd2+ 14.Kf4 b5) +5.33/45 650)

I'll give an example. I googled a question a while back about the record price for gold and it's current price and got back absolute rubbish.

This is the answer I got today to the same question:

The all-time high for gold price was around £1,708.57 per ounce, recorded on March 8, 2024. Today, May 10, 2025, the gold price is fluctuating around £2,508.55 per ounce.

Still serving up the same old garbage.
Moral: The jury's still out on so called AI.

Anyway many thanks Chessgames for a wonderful website and may it continue to entertain and educate us for many years to come.

Doing so, I really wanted to add something: that all we here are not young guys anymore. Many just left us, like our friend Brenin (may I say, RIP?), just to remember a strong chess analyzer (and player) out here.

Computers changed all in our life. In my Engineering case, one day I said to a manager on my area, during a job interview, in 1983, that nearly soon, drawing would not be made by designers at drawing boards, instead with mouse and computers. This forecast became laugh and despise about me on him and he concluded that I was mad: "that never will happen!"

Twenty years passed, I've to close doors of my office, because eventhough I have goodname, several clientes, services and so on, we have suddenly no steel for our steel structures, since our government (it was 2003 with Lula, who else?) have set out steel industries to great exportations and inside here, we had not steel to continue our designs. More than 25 companies closed doors, many factories throw their employers out... only later, on 2008 we would know that the main reason for this, was the supercharged prices of concrete in public construtions which becomes the called "big monthly" scandall, discovered by the famous car-wash investigation.

Yes, they premeditated everything. By that prices, steel could be used instead of concrete, so they throw overseas this kind of steel, for the big engineering companies, partners of this assalt to Brazilian finances, could rise their prices without any problem.

Trying to survive, I catch a job as designer in one yet surviving office, who works to the steel producers and who I find there? Yes again, the same manager, to who I asked: "well, I thought that you still work with drawing board". He remembered me and didn't laugh anymore, the designing room was plenty of computers and no board.

I always win easily the common computers with the first Chess games programs, until I run Chessmaster. I had to work my skill, it was not so easy to win some of his simulated players. When after years I could win at least once every person created by the program and get on the top of the piramyd, I begin to challenge playing against simulated masters (Reti, Capablanca, Smyslov, Flohr...). It was very very uneasy to say the less. I never did win a game against the own program Chessmaster and never used it on my analyzes or in my turn of play. But, the CD became damaged and the program doesn't run anymore. Some games I showed like comments, on KAB page of games.

I doubt that our sons (among many people) or family members will ever dedicate their minds to play our loved Chess, will undestand why we lose time here and why we fill so glad to do so. Everyday that pass, many seem to be less interessed in fight against the machine on a Chessboard. My son preferres eletronic games.

Nowadays with Stockfish and 8 core CPUs
and 20 Gigs or more of memory allocated
it's impossible to win - even for grandmasters.
The CPU can consider billions of board positions in 5 minutes. Now - all the grandmasters have their computer resource to check new openings out and they do - so chess has been changed forever for anyone who does some homework on openings.

Computers have killed off the art of interesting and speculative analysis and fantasy variations.
These days an author dare not suggest an alternative without running it through a machine because they know it's the first thing an interested reader would do.

But we can still jump in and give reasons as to why the human played a move not suggested by S.F.
They did not see it. Yes it can be as obvious as that.
Saw it but misanalysed it, had no time to look at it fully, saw a safer win, needed a win so took a risk,set a trap, went for complications, never had the nerve to play it because their opponent had a much higher rating than theirs...In fact we fill in all the gaps that S.F. creates.

<Mel>

"After all - what does it really matter what I might say? It could be nonsense "

Of course it matters. What people say keeps the forums ticking over. I think you will find a lot of people will be interested in a genuine question as to why a move was played. There are some good players here, you will get some good answers

And as for 'nonsense' Go for it. Based on human v human play and what I think is plausible I have often had a fantasy line squashed. No big deal. I'm happy that I can still spot the occasional shot and that someone took the time to look at it.

For instance. here;

click for larger view

I can see me going for 15.Bd7+ Kxd7 (forced) 16. 0-0-0 with e3 ideas. Black position looks hard to play, White easier, (something a machine cannot understand).

But times have changed.
A whimsical move that looks OK might prove invalid say 4 moves later when the computer is asked. I notice that the youtuber Agadmator is always referring to the engine, to back up what he is saying. Still - I like the comments here too.

If people start worrying what a computer will say about their suggestion then we might as well give up. Sometimes a plausible move not considered by a computer can lead to interesting play and whimsy proposals add to the fun.

On the whole human v human games are decided by blunders so iffy suggestions by people who are actually thinking on their own two feet are part of the game.
I and I'm sure others have never think any less of someone coming up with an idea that has a hole in it.
Even if a computer says it is wrong we can see where they are coming from and in some cases defend their choice by agreeing it is a plausible move (and be glad they posted it first because they were just about too. 😊)

Doubled Pawns:
Personally I never bothered about having a doubled pawn. It means I have a half open file for a Rook to do things. Meanwhile my opponent will switch into 'get the ending mode' and walk into a two move trap.

One thing we can never accuse a modern computer of is being dogmatic. They have opened up a few eyes with their total disregard of pawn structures and King safety.
Especially the latter, if it does not end in mate then grabbing material whilst walking their King across the board or throwing up their King pawn cover does not matter.