Thanks for the link. That is very funny! I think that 17...Nxe4!! is too counter-intuitive for even most Grandmasters.

18.fxe4 Qxe4 forks the bishop and knight. However, they can normally be saved, especially in such proximity to each other.

Leko and Naroditzsky pretty easily were able to find the win for White and I wondered why Magnus could not see it, but part of Stockfish winning sequence includes this fortress position after 65...Rg5+

click for larger view

1) +12.63 (37 ply) 54.Rxd5 cxd5 55.Kxd5 Ra2 56.c5 Kf7 57.Kc4 Ke7 58.Kb5 Rc2 59.Bb4 Kd8 60.a6 Kc7 61.c6 Rxc6 62.a7 Rb6+ 63.Ka5 Rxb4 64.a8=Q Rg4 65.Qe8 Rg5+ 66.Kb4 Kd6 67.Kc4 Rg4+ 68.Kd3 Rg2 69.Ke3 Rg1 70.Kf3 Rf1+ 71.Kg2 Rf6 72.Qg8 Ke5 73.Qxg7 Kf5 74.Qd7+ Ke5 75.Qe7+ Re6 76.Qg7+ Rf6 77.Kg3 Ke6 78.Kg4 Ke5 79.Qe7+ Re6 80.Qc5+ Kf6 81.Qf8+ Ke5 82.Qf4+ Kd5 83.Kf5 Rb6 84.Qc7

Also Karjakin not playing 20....Nxf2+ in Carlsen vs Karjakin, 2016. Not a winning move, probably would have led to a draw, but it would have brought SK a giant step closer to winning the title.

Also the fact that a queen exchange is in the air seems to make such a trick even less likely to work. At least, I think it would to me.

"Not easy" is an understatement; I should not have challenged the engines, but I did.

Now I have to live with the consequences.

Because even at top level "thinking" is now regarded as "boring". And then we are not speaking about chess, but about life itself.

Classical chess is mostly not thinking but an endless self-doubt sesseion. Normally, you know your move within the first minute. What follows in the next minutes, is self-doubt without any reason other than having more time. Whence the "think long, think wrong" phenomenon.

I recall a number of years ago that tests showed good players normally have their move within a few seconds of seeing an opponents move and 7-8 times out 10 that is the move played.

This sounds a bit ambiguous. A clear-cut example would be, "okay Ne5 doesn't work yet, how can I prepare it better?" But what about, "Darn, Ne5 never works, I'd better look for another plan... Oh, how about b4, which I never thought of before?"

It ran upon the lines of looking at Ne5 (the first reaction choice) and discovering 3 moves in it was not on because of a pawn to b4 move.

So then attention was switched to stopping b4 to get Ne5 to work, stopping b4 was not idea that was not first considered until it was spotted during analysis of Ne5.

You will experience this yourself when you do the POTD. One move will suggest itself within seconds.

Often it is correct but upon analysis you spot a flaw, so you keep your first reaction move up your sleeve and perhaps either remove or deflect a piece covering a sensitive square or play out the combo using a different move order.

To me, that is the difference having more time makes. As a low-level master player, my two main weaknesses are: a) Misevaluation of positions (especially the resulting positions after some moves, as I see them in my head), and b) Getting stuck on a strategic idea that is sub-optimal and not realizing I have to totally change plans.

This said, I do believe you that at the highest level of play, GMs are very likely to play the first move they think of, or an alternative, better move, that supports <the same plan>... because they have been thinking of the good plan all along (while correctly evaluating positions).

By coming to late on purpose, adjusting all your pieces while the game has begun already.

Then you play Queen's Gambit Accepted. Fumble with the captured pawn asap.

That's chess! According to Kasparov c.s. He found some great pupils in Carlsen and Pragg.

<So, as you said, it depends on the position, but it also depends on player's strength.>

The tests were not done under game conditions. players ranging from 1600 - GM were given selected positions and had to state their first reaction, who was better what move would they play etc...

The good players rarely changed their first reaction choice. Some lower rated players actually spotted what was deemed the best move but analysed themselves out of it.

Of course if you were in the middle of a Kingside sac attack or a long endgame plan then the chances are you will have seen what was going to be played before your opponent plays it.

It was run in the BCM (or it could have been CHESS, it was a while ago and I used to get both.)

Round about the same time a TV program ran a chess memory test. players were shown middle game positions for 20-30 seconds and then asked to recreate it on another board. Stronger players did a lot better than weaker players.