Chess will never be solved, here's why

Sort:
Avatar of Elroch
AerisChua wrote:
Checkers was solved tho. Oh and did you know that white’s first move options are 18

No, I didn't know that. I've always understood there were 20.

Avatar of tygxc

@13257

'variation of this that is not skewed to one side of the board, which is entirely reasonable'
++ Not really, white has more reasonable moves than black.

"assuming an average true branching factor of 35" ++ That is way too much.
10^38 = 3^80, hence the average branching factor without transposition must be 3.
There may be positions with 31 legal moves, but most of these can be pruned right away,
and many will just be transpositions, i.e. other move orders leading to the same positions.
Engines use transposition tables and do not recalculate.

Avatar of MEGACHE3SE
Elroch wrote:
MEGACHE3SE wrote:

dont forget that tygxc only assigns 1 node per position in his calculations.

I am not sure what you mean.

tygxc's solving chess in 5 years claim is literally 10^17/however many nodes per second. he only assigns one node figure out each of the 10^17 positions in the game tree.

https://www.chess.com/forum/view/general/chess-will-never-be-solved-heres-why?page=1#comment-66772279

Avatar of MEGACHE3SE

https://www.chess.com/forum/view/general/chess-will-never-be-solved-heres-why?page=5#comment-66899947 heres the full tygxc calculation, he says top computers will make errors in blitz games in the comments before mine, but yet he also claims that computers will have perfect accuracy looking at millions of positions a second.

Avatar of Elroch
tygxc wrote:

@13257

'variation of this that is not skewed to one side of the board, which is entirely reasonable'
++ Not really, white has more reasonable moves than black.

"assuming an average true branching factor of 35" ++ That is way too much.
10^38 = 3^80, hence the average branching factor without transposition must be 3.

You are forgetting two crucial factors but, to keep the discussion simple, I will point out that if you replace 35 with 3, that means there are over 6,000,000 distinct moves that could refute the analysis at every move. The consequence is the same.

There may be positions with 31 legal moves, but most of these can be pruned right away,
and many will just be transpositions, i.e. other move orders leading to the same positions.
Engines use transposition tables and do not recalculate.

You seem very slow to get the point that you can prune as many lower ranked moves as you like for the proponent (the only risk is that you will miss a better move), but pruning a single move for the opponent means your entire evaluation can be refuted if a single zero ply evaluation is wrong.
And this applies recursively to branches where the opponent plays a sequence of alternative moves, all analysed except the last.

All you have is a bet that it won't matter, backed up with overconfidence (and apparent lack of understanding).

Avatar of Thee_Ghostess_Lola

Dr. Phil is incompetent.

ur just jealous cuz hes extraordinarily wealthy, a dad success, and sectsee smart.

Avatar of Thee_Ghostess_Lola

[...]about AI you would know that it cant think on your own

riiiiiiight...about 5 yrs ago. which u wooda been abt 65 at that time lol !

Avatar of Justanotherfolkh

O AI will control my friend. I might be old but you would wish you were dead when AI start to control you

Avatar of Justanotherfolkh

Which might happen

Avatar of Justanotherfolkh

This world will be run by human brain not bots like you...

Avatar of Thee_Ghostess_Lola

There may be positions with 31 legal moves, but most of these can be pruned right away,and many will just be transpositions, i.e. other move orders leading to the same positions.Engines use transposition tables and do not recalculate.

so true. and recall 2. Ba6 ?...never EVER will white overcome being dn a bishop. elrock ?...repeat after me - "Inherent". and plz dont self-argue ur own common sense into being uncommon. ur better than that...i should hope.

Never EVER will one side be down over ~3 points w/out compensation in a superdupersilicon game. and dont expect a doozie combo that overcomes this height of decrepancy. itll be seen way before that.

we will solve chess. trust me. and itll happen in rachets. cold porridge > just right > goldilox burnt her tongue. not predicting the platinum solution will outcome a win. expecting a piece count alone of many (12-14 ?...if so). i can bob my foot & hold my breath. dont you dodo due. you'll pass out (burp).

Avatar of Thee_Ghostess_Lola

This world will be run by human brain not bots like you...

u may wanna take ozempic (more so for ur head).

Avatar of Justanotherfolkh

U might go to psychologist...

Avatar of Justanotherfolkh

You need it bro...

Avatar of DiogenesDue
AerisChua wrote:
Checkers was solved tho. Oh and did you know that white’s first move options are 18

Lol. Count again.

Avatar of DiogenesDue
Thee_Ghostess_Lola wrote:

Dr. Phil is incompetent.

ur just jealous cuz hes extraordinarily wealthy, a dad success, and sectsee smart.

Yeah...no. He's just an incompetent and ignorant man riding a cresting tide of ignorant people. I do feel more sorry for his fans, though, I'll give you that.

Avatar of MARattigan
Elroch wrote:
MEGACHE3SE wrote:

dont forget that tygxc only assigns 1 node per position in his calculations.

I am not sure what you mean.

...

@tygxc defines "position" as FEN-{move number,ply count} and then assumes his positions correspond 1-1 with nodes in the competition rules game tree and also in the ICCF game tree.

Avatar of DiogenesDue
Thee_Ghostess_Lola wrote:

There may be positions with 31 legal moves, but most of these can be pruned right away,and many will just be transpositions, i.e. other move orders leading to the same positions.Engines use transposition tables and do not recalculate.

so true. and recall 2. Ba6 ?...never EVER will white overcome being dn a bishop. elrock ?...repeat after me - "Inherent". and plz dont self-argue ur own common sense into being uncommon. ur better than that...i should hope.

Never EVER will one side be down over ~3 points w/out compensation in a superdupersilicon game. and dont expect a doozie combo that overcomes this height of decrepancy. itll be seen way before that.

we will solve chess. trust me. and itll happen in rachets. cold porridge > just right > goldilox burnt her tongue. not predicting the platinum solution will outcome a win. expecting a piece count alone of many (12-14 ?...if so). i can bob my foot & hold my breath. dont you dodo due. you'll pass out (burp).

It's easy for you to hold a position when your grasp of it is so soft and fluffy. Something you should ponder in general, not just for this discussion.

Avatar of MARattigan
Thee_Ghostess_Lola wrote:

There may be positions with 31 legal moves, but most of these can be pruned right away,and many will just be transpositions, i.e. other move orders leading to the same positions.Engines use transposition tables and do not recalculate.

so true. and recall 2. Ba6 ?...never EVER will white overcome being dn a bishop. ...

Demonstrably false. I let SF16.1 play the position as White and it already has overcome being dn a bishop. 

In fact the position is a White mate in 45. Here is a @tygxc style proof.

And engines can validly use only transpositions of nodes, not @tygxc's positions under competition rules, to avoid recalculating. There aren't too many of those.

Avatar of BigChessplayer665
llama_l wrote:
tygxc wrote:

"He's rated lower than me OTB"
++ Maybe he lacks the nerves to play over the board with a clock ticking.
Maybe he is good at positional play, but cannot calculate tactics.

For what it's worth, my strength is endgames / positional play. I'm weakest at attacking / combinations. When I was a beginner I thought tactics were only tricks that worked on bad opponents, and I thought GMs only won in the endgame so that's what I studied... my rating improved very slowly many years later I realized attacking was not "tricks" and it was better for beginners to focus on tactics but at least I had fun doing it my way.

10 years ago I'd regularly disagree with the quick analysis of engines (chess.com or lichess or on my PC). They would suggest "bad" moves, but if I gave them more time or forced them to look at different lines, they'd eventually agree with me... that doesn't happen anymore. Sure sometimes I prefer the 2nd, 3rd, 4th choice of the engine because it's more practical, but I can't prove them wrong (so to speak) anymore... now... could a top GM laugh at an engine's anti-positional move and be proven right? I don't know, I'm not even close to that level... but neither are ICCF players. In fact they're even lower rated than me... and that's my point.

Hi I'm prob going to send this to someone else lol