For example, this combination is one of my favorites of all time because it leads not to immediate mate but to an endgame with a better pawn structure, suggesting the opponent's slip was somewhat smaller.

It's certainly decisive, no doubt about that.

"Unncessary" is an inherently subjective term, but I would not use it unless the move put the win in doubt or made it harder just for the sake of flashiness. This move does just the opposite, puts the win beyond all doubt, and brings a complicated game to an immediate conclusion. Nobody would avoid this move, nor should they, unless they didn't see it or were uncertain about its effectiveness.

"Not really all that brilliant" is a joke.

But if we are talking about the greatest move ever played then it needs to be "better" than all other moves across every chess games in history.

And that is where I have a problem. Of course, it depends on what you need from a move to be the best ever. If you are looking for a move that needs a lot of calculation, then Qg3 doesn't quite work. The lines are relatively short. Our typical Sunday POTD is usually a lot deeper.

Or you might want a move that is the culmination of a deep plan, say one of Nimzo's zugzwang games. And again Qg3 slightly disappoints. Marshall grabs a loose Bh3 and finds a clever way to hold on to it. Opportunism rather than the culmination of a grand strategy.

Or you might want a move that overturns a material deficit or wins a bucketful of bits or makes a daring queen sac. Fischer's game against Byrne, perhaps? Qg3 doesn't do any of this. Before the move black is a knight up. After the move he is still precisely a knight up.

If Qg3 was an album it would be Bat out of a Hell. A great album, makes a huge impression, bombastic, one of my favourites of all time ...

Game Collection: Game of the Day Pun Index (A-C)

The new pun on the third time came because the game was used for a different reason, on the 100th anniversary of the day it was played.

The only move I recall with a similar effect was Fischer's 21...Qd7 in R Byrne vs Fischer, 1963. In fact, I think that Fischer's move is even more remarkable because it was so unspectacular, yet Byrne's resignation afterwards was shocking to all those who were witnessing the game. I also wish I knew how long Byrne thought after Fischer played 21...Qd7 before he resigned.

But 21...Qd7 could hardly be considered the best move ever since it was part of a calculated sequence of moves. If any move in this game deserves to be called the best it's 18...Nxg2. It's this move which was, as Byrne said, "the shocker", because it was so unexpected.

So perhaps that's one requirement for a move to be considered the best ever. It must be unexpected both to the opponent and to anyone watching the game. Another requirement might be that the move is part of a deep combination that changes the course of the game, turning either a loss or a draw into a win.

First would be flashiness. How stunning, beautiful and unexpected a move looks. Well 23... g3 is a perfect ten in that area.

Second one would be depth, number and difficulty of variations involved to justify the move. In this case they are just two very short variations and one a bit longer but everything is obvious and forced. Fischer's 18... xg2 in Byrne-Fischer is much stronger in that area (or say 17. h4 in Polugaevsky vs E Torre, 1981 ).

Third would be efficiency of the move : it should be the best move of the position and by the biggest margin possible. For instance 17. d5 in Anderssen vs Kieseritzky, 1851 would fail (it makes a mess of an easy win). Well in this game 23... g3 is not necessary at all and some other moves are not much worse, although there are no move obviously stronger and by forcing a liquidation it's not without practical value. By contrast, Alekhine vs Verlinsky, 1918 gives an example of an insane queen move which is the only solution to a complicated problem.

The criteria presented are:
1. aesthetic appeal
2. complexity
3. efficiency

but i would suggest a FOURTH criteria..
4. effect
i.e., what consequence does the move have in the game?

For all it's visual appeal (and it IS undeniably spectacular) it only wins a piece. After 23. ..Qg3 24. Qxg3 Ne2+ 25. Kh1 Nxg3+ 26. Kg1 Ne2+ 27. Kh1 Rc3 black has won a knight, and though the position is resignable, it is not forced. White could have chosen to play on a piece down. Whereas in some other games the final blow removes ALL defense with immanent mate that is unavoidable. In such case THAT move would rate higher for it's effect / consequences. It may be true that capturing the queen by a pawn ends in mate... hxg3 Ne2++ or
fxg3 Ne2+ Kh1 Rxf1++ and
Qe4 (to prevent Qxh2++) Nf3+ Kh1 Rxh2++