We will have to wait for 32 piece Table Bases! If Moore's Law hold's up I calculate 24-28 years.

>>
<Ishaan: I don't think Kasparov had a very great opinion of Topalov.

Magnus saw that after 37...Rxh5 38 Re6+ Kf7 39 Rxe5 Kf6 40 Rxd5 Rxf5 41 Rxf5+ Kxf5

click for larger view

Now the important point is that White can get to b4, ahead of the pawn, at a moment when Black cannot play Kb6, which would be a draw.

If Black wants to play the bishop to e6 and allow the capture Nxe6 he will not want the f7 pawn to be pinned, so as to preclude the recapture ..fxe6.

This suggests that an alternative to 16...Ke7 is 16..Kf8 17 Rg5 h6! 19 Rg1 and only now 19..Ke7 but then on 20 Bd3 Be6 21 Nc7 Rc8 22 Nxe6 fxe6 White has 23 Rg7+.

Definitely!

<carlsen like so many other young so called prodigies will fizzle out quite soon.>

yes, many prodigies fizzle out. However, historically speaking, I suppose all top players in the world - be it Kasparov, Fischer etc.- have been prodigies at some point. They were/are <the prodigies that didn't fizzle out...>

<We will have to wait for 32 piece Table Bases! If Moore's Law hold's up I calculate 24-28 years.>

Nope, we will never have 32 piece tablebases. Let's figure out how many positions would be in such a tablebase. We may place 32 pieces on 64 distinct squares. Among the pieces there are 8 indistiguishable pawns for black and white, and indistinguishable pairs of rooks and knights of each color. That yields (64!/ 32! 8! 8! 2! 2! 2! 2!) possible piece-placements. Some large fraction of these positions are illegal (both kings in check) or unreachable (a white pawn beyond the black pawns, with no captures played). So, put another factor of 1000 (a guess) in the denominator to eliminate positions that cannot happen in a game. Even so:

(64!/ 32! 8! 8! 2! 2! 2! 2! 1000) =~

(32)^127 e^(-32)/ 1.6*10^12 (Stirling's approximation) =~

e^380 =~ 10^165 reachable chess positions.

If you compare this number with the roughly 10^80 elementary particles in the known universe, you'll see that even if we had the whole universe at our command, and even if we could represent one chess position in memory by a single elementary particle, we could not construct a computer with sufficient memory to represent all possible chess positions. Not even close.

I don't know if chess will ever be "solved", but if it is, it must be solved by some theorem, not by constructing a complete tablebase.