Let's be frank - the idea that even this class of endings can be adequately dealt with using rules of thumb based on tiny samples is ridiculous.
Chess will never be solved, here's why
Sort:
May 27, 2024
0
#7925
MEGACHE3SE wrote:
" at 10^7 positions/second/engine "
why do you continue to make this error when it has been repeatedly pointed out?
you falsely construe nodes as a full positional evaluation.
The only "full positional evalation" is the one that solving chess needs to achieve for the initial position.
No need to do 10^7 of those: one will do. 
May 27, 2024
0
#7926
MEGACHE3SE wrote:
hypothetical question: how does P vs NP affect chess solving?
P v NP can only be relevant to classes of problems with a size parameter that has no upper bound. For example, solving tictactoe on boards of size N x N. It is about how hard it is to find a solution compared with to verify a solution.
May 27, 2024
0
#7927
Elroch wrote:
MEGACHE3SE wrote:
hypothetical question: how does P vs NP affect chess solving?
P v NP can only be relevant to classes of problems with a size parameter that has no upper bound. For example, solving tictactoe on boards of size N x N. It is about how hard it is to find a solution compared with to verify a solution.
well chess can be expanded to larger boards (maybe just expanding the height?), but with the problem of say "is there a set of white moves which guarantees a draw", the solution set can be verified much easier than the solution could be found, and I was considering if that degree of difference was enough.
@10625
"We can't conclude that perfection has been reached."
++ We can, because all games end in draws.
That can only be if Chess is a draw and if all errors come in pairs that undo each other.
Any unpaired error would cause a decisive game.
"This does not prove that missed winning lines didn't exist in ie. 100 games played"
++ It does, especially because the drawing lines are redundant. On 1 e4: 1...e5 as well as 1...c5 as 1...e6, on 1 d4: 1...Nf6 as well as 1...d5, on 1 Nf3: 1...d5 as well as 1...Nf6.
Further: on 1 e4 e5 2 Nf3: 2...Nc6 as well as 2...Nf6. On 1 d4 Nf6 2 c4: 2...e6 as well as 2...g6 etc.
On 1 d4 d5 2 c4: 2...e6 as well as 2...c6 and 2...dxc4.
As to 1 e4 e6 I am willing to accept the unlikely possibility that all 3 ICCF (grand)masters are wrong and that 1...e6 could be a mistake and 3 Nc3 a missed win.
If you want, you can discard the 2 French games.
@10630
"which of the quadrillion opposite coloured bishop endings with more than 7 men are known? You do understand that in chess, the exact position of all of the pieces matters?"
++ That is why it takes 2 ICCF (grand)masters to agree on the draw after ample consideration with engines of the exact position when both have no hope of winning. It would be pointless to force them to continue until a 50-moves draw, i.e. 500 days or almost 2 years.
Not all opposite colored bishop endings are draws, but this one is
https://www.iccf.com/game?id=1360123
Not all rook endings are draws, but this one is
https://www.iccf.com/game?id=1360174
Optimissed wrote:
Why do you imagine I don't mean what I say? I certainly don't see much value in playing lip service to people who can't understand that they are making fools of themselves by going round and round in circles arguing completely unproductively with tygxc. Elroch isn't bright enough to add anything useful. It's as simple as that.
I do agree that you can't always rely on rules of thumb. But for instance, take the Kp vs KNN ending. There's a diagram used where you can enter the pawn position and you immediately know if it's a win, because it's a rule of thumb that depends on how far advanced the pawn is and which file it's on.
tygxc and Elroch are too much like each other and they both have weaknesses. The rest of you allow yourselves to be led by them and so can you imagine what it's like for someone who knows their conversation is unproductive and that it can't be anything else? MAR is a troll and deliberately messes around with people by talking rubbish, Dio is a troll who is almost incapable of making arguments concerning subject matter, you hang on Elroch's coat tails. You say you're a second year maths student at university? Act like it. Stop pandering to them and use your brain if you have one. Don't be scared of them. Of course they'll try to make it seem you don't know what you're talking about. That's what they always do and you have to learn sometime or other.
Maybe. Well, I suppose you don't and you can be another failure. I've been too busy to be involved in the past few days, I return and it's the same nothing arguments. Just two fools arguing that black's white or an indistinct shade of grey.
You should take longer vacations from this thread. There's nothing you can contribute anyway.
May 27, 2024
0
#7931
llama_l wrote:
MEGACHE3SE wrote:
Elroch wrote:
MEGACHE3SE wrote:
hypothetical question: how does P vs NP affect chess solving?
P v NP can only be relevant to classes of problems with a size parameter that has no upper bound. For example, solving tictactoe on boards of size N x N. It is about how hard it is to find a solution compared with to verify a solution.
well chess can be expanded to larger boards
Yes, and that was his point. Chess as we play it is only on an 8x8 board, so asking whether it (and it alone) is P or NP isn't an appropriate question.
just because im only interested in the 8x8 solution doesnt mean that the P v NP of a generalized form of chess wouldnt give insight on the 8x8 solution, besides, upon a little more reflection, the verification is on too high an order to apply to p v np
May 28, 2024
0
#7932
Of the last 60 posts - a considerable number were Elroch and MarAttigan discussing the three move repetition rule.
They're both very good posters but I tend to agree with Elroch about 3 move repetition.
When trying to form tablebases starting from two Kings obviously one doesn't have information from the beginning of the game and since - relevant to the 50 move rule nor to 3-fold repetition.
If we have two pawns in an en passant formation then computers could in theory solve both cases as to whether en passant is available or not.
Such would not add substantially to the task because a very low percentage of positions contain such formations.
If we have at least one King in castling formation with either or both of its rooks - we don't know whether the King or its rook or rooks have previously moved and returned to their original squares - or not - but again both cases could be solved for without adding substantially to the overall task.
Again because of low percentage of occurrences although considerably higher than en passant formations.
And would include positions with in-between pieces since they could be moved out of the way and then castling might be legal or not and might be relevant to solution in both cases.
-------------------------------------------------------
Point - the tablebases are a route towards thoroughly solving chess - even though it could take trillions of trillions of years.
But if the computers had to 'conceive' of all the permutations of games resulting in any particular position ...
then that could take many septillions of septillions of septillions of years.
Permutations of possible move orders greatly outnumber the maximum possible number of chess positions with 32 or less pieces even if illegal and legally unreachable positions are included.
-----------------
The forum subject appears to concern solving chess given the board position and whose move it is.
Not the game history previously.
If that is factored in too - then the task becomes even more unsolveable in the foreseeable future.
What has happened here apparently is that tygxc realized immediately at the inception of the forum that chess is not solvable given today's technology -
and therefore that there would be nothing to discuss on that unless a somewhat different discussion occurred.
Namely - 'variations' on solving.
The more pronounced the variations the more such activity resembles players playing.
@10650
"chess is not solvable given today's technology"
++ Chess is not strongly solvable to a 32-men table base with 10^44 positions given today's technology, but very well weakly like Checkers or ultra-weakly like Hex.
May 28, 2024
0
#7934
Elroch wrote:
MEGACHE3SE wrote:
hypothetical question: how does P vs NP affect chess solving?
P v NP can only be relevant to classes of problems with a size parameter that has no upper bound. For example, solving tictactoe on boards of size N x N. It is about how hard it is to find a solution compared with to verify a solution.
You cracked the conundrum. I was trying to work out what was special about KPvKNP.
May 28, 2024
0
#7935
@Optimissed has a point tho, all of you except him seem to be going in circles. he kind of is the only one bringing up new topics while @Elroch and all the others just seem to contradict him no matter how valid his argument is.
ardutgamersus wrote:
@Optimissed has a point tho, all of you except him seem to be going in circles. he kind of is the only one bringing up new topics while @Elroch and all the others just seem to contradict him no matter how valid his argument is.
There one other that's not going in circles is megan s but he's just not talking to the other people because he doesn't want to deal with trolls besides txgc(who's not a troll)
May 28, 2024
0
#7937
tygxc wrote:
@10625
"We can't conclude that perfection has been reached."
++ We can, because all games end in draws.
That can only be if Chess is a draw and if all errors come in pairs that undo each other.
Any unpaired error would cause a decisive game.
False. It could also happen if Chess is a win and all games contains an odd number of half point blunders.
These games are played by the same engine as the ICCF games. The starting position is a win and all games contain an odd number of half point blunders. (Incidentally I've also posted you examples of sets of games that start from a drawn position end each contain an even number > 0 of half point blunders - why do you call that perfection?)
May 28, 2024
0
#7938
tygxc wrote:
@10630
"which of the quadrillion opposite coloured bishop endings with more than 7 men are known? You do understand that in chess, the exact position of all of the pieces matters?"
++ That is why it takes 2 ICCF (grand)masters to agree on the draw after ample consideration with engines of the exact position when both have no hope of winning. It would be pointless to force them to continue until a 50-moves draw, i.e. 500 days or almost 2 years.
Not all opposite colored bishop endings are draws, but this one is
https://www.iccf.com/game?id=1360123
Not all rook endings are draws, but this one is
https://www.iccf.com/game?id=1360174
Proof?
A recording of your big red telephone conversation perhaps?
@10655
"the same engine as the ICCF games"
++ And the same time 5 days/move?
Blitz games do not count.
@10616
"what are agreed draws meant to prove?" ++ When both ICCF (grand)masters and their engines have lost all hope to win, then they agree on a draw. It is pointless to play on in a known opposite colored bishop endgame with more than 7 men..
Could you tell me exactly which of the quadrillion opposite coloured bishop endings with more than 7 men are known? You do understand that in chess, the exact position of all of the pieces matters?
I created a related topic a couple of years back. Is this one "known"?
https://www.chess.com/forum/view/endgames/opposite-coloured-bishops-always-draw
This one is obviously known to @tygxc at least, because he posted a link to it on your thread, though it's not entirely clear if it falls under his definition (viz. ++ There are centuries of endgame knowledge). Maybe it does, because Tarrasch at least seems to have eventually lost all hope to win and decided it was pointless to play on.
But on the other hand I think ++ There are centuries of endgame knowledge is probably intended to be taken to mean Both Stockfishes show 0.