will chess ever die out once its fully solved?

#90
Fully solving chess requires about 10^19 legal, sensible, and relevant positions.
Cloud engines reach 10^9 nodes / second.

The solution for all sports is known - play perfectly and never lose. Sport marches on.

Citation needed.

It depends on what your definition of solved is.

It took 5 years in 2007 (according to you).

Today's technology is much faster.

Well,maybe no

#93
Solved = weakly solved, like checkers
https://en.wikipedia.org/wiki/Solved_game

We know from when checkers was solved, that its proof involves about the square root of the number of possible positions.
Chess has about 10^38 legal and sensible positions, thus by analogy it can be conjectured that about 10^19 are relevant. Point is that each capture and each pawn move renders vast numbers of legal and sensible positions irrelevant.

Chess has about 10^53 possible positions (including illegal positions).

Where does 10^38 come from?

#96
Google, IBM, the Lomonosov University all have the capability to solve chess, but somebody has to allocate a few million $ for 5 years of cloud engines and human assistants.
It is like sending humans to Mars: it can be done, but the money must be there first.

Well, I doubt any single entity could solve chess in five - ten years. I also doubt "million" is correct number for amount of £, $ and € we would need to spend. But every single university and tech company working on only chess for five - ten years would surely make some progress.

Is it possible to splitt solution between many different super computer and networks? If not, then I would agree about how it is impossible.

#99
No, 10^53 is an obsolete figure.
The number of possible positions is lower.
The number of legal positions is even lower.
The number of sensible positions is even lower.
The number of relevant positions is even lower. 

See
https://github.com/tromp/ChessPositionRanking 

Why do you continue to say 5 years when your source said it would take 5 years with 14 year old technology?

If you believe your source, then you believe it would be much less than 5 years with today's technology.

Anyone who is not an idiot knows that.

However that's not what I asked you. I asked you where 10^38 comes from.

I'm not trying to be argumentative. I'm willing to entertain that a weak solution requires approximately the square root of relevant positions, so if you can give me a reasonable explanation for where 10^38 comes from I won't disagree so vocally... I'll still disagree (there are plenty of nit picky things) but at least you won't sound like a crazy person to me, and I'll be willing to keep it to myself.

#104
See 
https://github.com/tromp/ChessPositionRanking 
From these data follows the 10^38.
The estimate of 5 years stands, maybe with less computers now.

#105
It is all opinion until it has been done, then it will be fact.
John Tromp has counted all possible positions and then by sampling has established 5% of these are legal. He also has counted the numbers of positions with 0, 1, 2... 16 excess promotions. Some positions with 1 or 2 excess promotions are senisble, e.g. 4 queens, but many of the positions without excess promotions are still not sensible, hence approximately number of sensible positions = number of positions without excess promotions.

Ah ok, so ~10^38 is for legal positions with no promoted pawns.

That sounds reasonable.

Of course not. "Solve" means that one side wins if both players make best moves. Players already play "suboptimal" moves because they are trying to catch opponents out of prep, so the "losing" side will simply avoid those lines. There are kajillions of possible responses and there's no way that one side could remember solved lines for all of them. And even if there were solved lines for "best" play, there can't be solutions to ALL lines. Suboptimal for the "losing" side can be as little as .0001 centipawns.

The idea that things would change is dumb.

#108
Pawns can be promoted to pieces already captured, but not to excess pieces that have to be borrowed from another box of chess men. There are some more details and so on, but to investigate the feasibility we are only interested in the order of magnitude.

An important speedup is to start analysis with positions of 26 men. The same research of John Tromp has shown that most chess positions are with 28 men, i.e. 2 captures/recaptures. Thus the human assistants with ChessBase should set up 26 men starting positions to launch the cloud engine calculation.

#109
No, not at all. To solve means to demonstrate that black always has a move that draws against alle sensible white moves.

@tygxc - it doesn't matter if solving involves winning or drawing. there's just no way a human can memorize all drawing lines of every possible variation. known drawn lines will be avoided easily.