How many moves a GM can think?

sorry was down pub havin' Sunday lunch, didn't get to see your post, tiresomebug ,but that was the way the story was reported when I read it nearly 40 years ago; and the story is extremely aposite to the OP's question - or don't they do lateral thinking in Singapore either ?

There are two points where a world class player might go into a long think.  One being if they've reached a critical position or otherwise need to find the thread (theme, idea) of the game.  The other is when the position is very tactical in nature and has many forcing moves.  Although recaptures are not the only time a move is the only move, recaptures are a good example.  You're somewhat right that after these thinks the moves do come faster, at least comparatively.

However other than this, like Tonaydal said, you don't quite hit the nail on the head.

First of all, calculating 20 moves isn't the same as looking into a crystal ball -- it's not possible to see every variation, you just try to find the best moves for you opponent to work against your ideas.  If your opponent has a particularly dangerous way to work against you, and you're a world class player, (and there are some forcing lines e.g. re-captures) then 20 moves is not impractical or impossible, and yes it happens, even every day.

To the amateur each move is made in isolation, one at a time, therefore each move gets equal amount of calculation.  To a world class player all the moves are connected and flow in a logical order.  This is one reason they remember games from 10 years ago, not because of photographic memories, most of them do not have this, it's just like you can remember the plot of a 2 hour movie a week, month, or even years after going to the theaters, because the information has order and flows logically from beginning to end.

They don't bang out those 20 moves because the opponent might find a slightly different way to challenge the idea that suites their playing style -- as long as this different way doesn't change the theme though, the following calculation will not take nearly as long as the initial deep think.

can someone tell me please how you is s'posed to calculate one move ahead when u is dealin' with ur girlfriend ? - me bin workin' on that one for 40 years an' has got precisely nowhere - somehow me is always 'wrong'

Absolutely, but I think we're all curious about, nevertheless, how much they can calculate. I don't really know, all I know is really well in all areas (not just depth of course)

Just how forcing was this? Was it just lots of queen checks? I mean all tactical puzzles are forcing, but that doesn't mean there aren't lots of alternatives to consider for each side every ply. Were they just pawn endings? I wouldn't be so quick to dismiss yourself from calculating that long, I'm class B-A and I think that's really impressive. Unless it was just a queen dancing around or a pawn ending, in a complex position my absolute limit is probably 8 accurate moves, and to get that I have to be really lucky.

It seems the claim of 10,12,14 moves was exaggerated on my part.  When I'm determined to find the answer, other than taking sometimes 20-30 minutes to find the solution, I'll also calculate moves past the book solution just to be sure.  I had remembered writing solutions 10+ moves long, but all these moves were not necessary and not part of the solution in the back, example below.

In respect to actual book solution moves, my max is around 8 which are usually king hunts / straight out mate in 8s (of course I'm happy to get a mate in 3 right, 8 is an extreme case)

To this I recorded the solution: 1. Rh8+ Bxh8 2. Qh7+ Kf8 3. Bc5+ Re7 4. Qxh8+ Qg8 5. Qh4 Bxf5 6. Qxe7+ Kg7 7. Bd4+ Kh7 8. Qh4+ Kg6 9. Qf6+ Kh7 10. Qxf5+ and wrote "and with Bc3 white will even pick up more pawns"

The book solution only went to move 5 saying you get full credit if you saw that in any case black is lost.

Of course then there are cases like these where for whatever reason, I simply couldn't find the solution.

After looking at moves like Bh3, Bg4, and ideas like sacing on f2 I finally gave up -- the puzzle was rated easy, the solution was one half move long, simply Bd7, I was very frustrated

Look I don't pretend to be a chess whiz or anything, but as far as I can tell strong positional abilities have to come before you can calculate any sort of real variations (with any consistency that is).

It's absurd when you see 1500 CC players claiming that they routinely see 5-6 moves ahead. What exactly does that entail? Anyone can see ten possible moves ahead, but it takes a player with very, very well developed positional abilities to see multiple variations of ten LIKELY moves ahead (we're talking strong GM here).

If you don't have at least expert level positional and tactical abilities then I wouldn't touch your four, or even three move variations with a nine and a half foot pole.

I agree, it's all about accurate evaluation at the end of any length of calculation.  In a tournament game I bust my arse trying to see 2-3 "good" moves ahead, and am very happy when I can.  Most of the time though, my 2 to 3 moves are just poor to okay.

Even in CC chess, sub master's positional understanding is so low as to create positional blunders at least once a game.  This is why I also wouldn't want to touch it with a 10 foot pole, I don't care how long they spent on their analysis board, just because it doesn't drop a pawn, doesn't mean the master wont be thinking to himself after a positional blunder "just a matter of technique from here, too easy :) "

Murray Chandler writes, in How to Beat Your Dad at Chess, in reference to Daimiano's Mate:

"An eight-move combination, sacrificing a knight and two rooks to force mate.  Fantastic!  Yet when shown the position, the famous Grandmaster John Nunn found the combination in just two seconds.  Interestingly the PC program Fritz 5, running on a Pentium, took several minutes to find the forced win.  It is remarkable that a human brain is able to solve a purely tactical position faster than a powerful computer..."

Yes, positional recognition and thinking are far more valuable than sheer mental power when finding winning moves.  You don't need to look ahead a great deal, you just need to be able to see how things will fit together.

 Didn't Garry Kasparov said that as well ?

A different way of looking at the same thing: It's hard for us mere mortals to understand how good the REALLY good are. Or, as Bobby Jones once said of Jack Nicklaus, "He plays a game with which I am not familiar."

The point being, unless you play at nearly the same level as your opponent, you really aren't even playing the same game.

I can't find the source of this quote...

"Masters don't play the same game we do, it just happens to have the same rules."

Is a good expression of anyone whose played a master / grandmaster has probably felt

I think this is a little harsh. I don't think it's right to equate making positional mistakes to having a positional understanding that is 'so low'. I think there are plenty of players with (FIDE stylee) ratings of 1800+ or 2000+ who have a pretty good positional understanding, and will play many good plans during a game.

I think it's more that there are certain gaps missing, certain dogmas still existing, certain bits of knowledge not ingrained... so that there are certain instances where they will falter. This doesn't mean that a lot of their positional play isn't actually pretty decent.

In my opinion, you cannot figure too many moves in advance, because the player doesn't always respond to what's expected.  When you figure moves too far down the line, you are playing lottery odds of predictability.  An effective player, during a complex-middle game, predicts a few moves at most.  People, don't look at things so deeply.  I bet the masters move so predictably to their advicary(sorry for spelling) that they can make calculated moves, because the difference in ability and education is so similar.  They know what masters think.

It doesn't really matter what your opponent actually does, though. So long as you make the strongest continuation against their BEST moves, anything inferior to that only increases the strength of your position. In that case you'd be calculating again from a new position, but still only considering the best possible replies to his moves.

15-50 moves. 15 at least. Most GMs can calculate up to 35 moves ahead.

Indeed, most Grandmasters calculate about 35 moves ahead, in all variations, that's why most games end around move 15, because they already know how the game will play out, there is no reason to play it.

I've seen a mate in 238 (at least, I think it was 238) puzzle that I'm sure a GM could calculate most of in his or her head. It all depends on the situation...

EDIT: Ah, it was mate in 271. Here: http://chessbase.com/puzzle/games/puz7.htm

I don't think GM can think 35 moves a head 

It is like Super human power

Hmmm....I could be totally wrong here, but an eight-move combination seems a little implausible to "fully" solve (by which I mean every possible reply by the opponent is considered, even "weird-looking" moves) in 2 seconds unless the position was "strictly" forcing (by which I mean there were few legal moves for the opponent at each step, not just that there were few intuitive moves). Of course, I'm nowhere near being a strong player so I'm kind of shooting in the dark here. Plus I guess if most of the replies were intuitive, then a 2 second analysis should be enough to spout off the main line; which would impress most people enough that they wouldn't think about whether the GM saw some funny knight move in the far corner of the board that could have required a totally different variation.

We do plenty of lateral thinking here. Which is what leads me to the inevitable and totally logical conclusion that inability to admit mistakes must be a flaw in everyone from Spain rather than a simple failing of your personal character.

http://www.chessbase.com/newsprint.asp?newsid=2326

Link includes the interview of kasparov give after retirement dated 14.04.05

According to interviewer Kasparov made 35 moves preparation during match against Kramnik in 1999, a match were all computers gave +3 win to Kramnik but Some how Kasparov thought 35 moves ahead and drew the match. Kasparov didn't deny the statement. So we can assume that he might have actually thought 35 moves A head!!!!!!  Amazing!!!