Chess will never be solved, here's why

Ok, I tested a few interfaces I have. Some assume castling is illegal and some don't. Mostly it seems to depend on whether you're starting with a full board or empty board. Castling is legal when starting with full board, and illegal when starting with empty.

But none allow it to be ambiguous and try to figure it out afterwards. After giving an engine a position, the FEN notation will tell you how the engine sees it. This is regardless of how you set up the position (manually, with FEN, or otherwise).

"Problem mode" and "game mode" are things you made up, so I have no way of responding to that.

Engines take positions as input, and assign a value to each legal move as output.

Interfaces can implement a "game mode" on top of this functionality by ending a game when time runs out and by not allowing the user to take back moves (etc), but this doesn't change the operation of the engine which is simply assigning values to legal moves.

That's not to say there aren't more versatile (or specialized) engines and interfaces, but if you had such an uncommon and specialized piece of software (one that rigorously tests whether castling is legal), you'd have mentioned it.

"Castling is legal when starting with full board, and illegal when starting with empty."
Well that looks like Progress.  And if that's always true - that would be better than looking at FEN codes.
There's been much complaining in the Tactics puzzles forums  from people asserting
"How are we supposed to know that castling is legal here?"
So much so - that the site had to reduce puzzles with castling in them.
Which is a really bad outcome.
But if you put in each problem in advance whether castling is legal or not - that would be both prohibitively cumbersome and prohibitively revealing.  

Regarding FEN codes - I found them useful only when transcribing puzzles in the tactics puzzles over to clubs or forums.  Found a way that worked OK.

Regarding how to make sure castling is always legal by default when setting up - 
two possibilities
1)  the site improves the options and features in the diagrams - making them about 300% more User Friendly.
That would greatly improve chess discussions throughout the site.
2)  a player could keep castling legal - laboriously. 
Start with a full board.
Remove a bunch of pieces and shift the others to where they belong without using the moves feature.   

This conversation has another use.
Even those 'familiar' don't seem to know how the defaults work.
Or might find out soon.

Rocky64 suggested in a private message that I clarify my comment to indicate that Ilama51 is correct in believing there are presently no chess programs that can make a practical test for castling legality in an arbitrary position. If you wanted to try doing it with Natch, which finds proof games, you would probably start by modifying the freeware to stop looking as soon as a proof game is found in which castling occurs, and to exclude any trial games in which either the king or both rooks move. Even then, there would be no guarantee of answering the legality question in a reasonable amount of time.

I don't know how Stockfish is programmed to determine castling rights in a given situation, and don't have any speculation to offer on that question.

"I don't know how Stockfish is programmed to determine castling rights in a given situation"
I like that.  'I don't know'.   

"Problem mode" and "game mode" are things you made up, so I have no way of responding to that."
I didn't 'make up' anything.
I suggested them as possibilities.
Why should we assume that Stockfish is always in the same mode all the time?
Try - 'Only an insider would know for sure exactly how chess.com runs its software' - and that's only if he/she has the time to look.
is an idea. 
But fact:
They've got a 'Dev team' here.  Lots of stuff to look at. 

Yeah, these discussions annoy me too. It's all very negative IMO. Apparently you feel the same way i.e. that you're not begin treated fairly.

Finding a good discussion on chess.com is so rare. Makes me sad.

We have yet to discuss here why castling in the diagrammed position would have to be illegal.
And this topic was/is mainly about solving chess by supercomputers -
not the issues of  precisely and thoroughly how chess.com programs it s Stockfish software. 
I remarked that Stockfish analysis button didn't regard castling as legal in the diagram.
That generated many 'inferences'.  But some might miss the point.
Which might have been made unintentionally by somebody.
That if the computer had to check for particular illegalities every time - it would 'lose' games too much on the clock.
It would be similiar with supercomputers taking on the task of 'solving' chess.

Well, I don't know how to salvage this discussion either.

Perhaps selfishly, I'd like to start by saying while I'm far from an expert on stockfish, I get the impression that I know a lot more about chess engines and their operation than you do. I get the impression that, frankly, you know basically nothing about chess engines.

As for why castling is illegal, such puzzles provide a lot of joy to many people... but I am not one of those people. I've never enjoyed that stuff.

'not being treated fairly' ?   How 'I' 'feel' ?
Consider that one chooses how one feels.  Many forget that.
As for the discussion 'annoying' - its more important what one does about that.  Is the suggestion.
And you could 'get the impression' of anything you'd like and in a personal way too.   'llama' account.  That seems familiar.
And people can worry a great deal about 'credentials'.
I guess you've never been to the moon - but does that mean I should suggest to you that you shouldn't talk about the moon?
I think I'd do better resonding to other posters -


We have yet to discuss here why castling in the diagrammed position would have to be illegal.
And this topic is/was mainly about solving chess by supercomputers -
not the issues of precisely and thoroughly how chess.com programs its Stockfish software. 
I remarked that Stockfish analysis button didn't regard castling as legal in the diagram.
That generated many 'inferences'.  But some might miss the point.
A poster might have unintentionally produced a point ...
That if the computer had to check for particular illegalities every time - it would 'lose' games too much on the clock.
It would be similiar with supercomputers taking on the task of 'solving' chess.  Having to check on whether the position 'could get there legally'  Or - much much worse. 

"but I am not one of those people. I've never enjoyed that stuff."

then why are you talking about that then?
You've already conceded that you're 'far from an expert on stockfish'
- progress.
you want to talk about 'salvage' of a discussion that you chose to have that you didn't have to be in and 'never enjoyed'? 

@Elroch
"Took me a while to work out the logic of why queenside castling is illegal after I found the solution to the problem. But Stockfish is not being clever: it's just that the FEN code tells it queenside castling is not legal"
Apparently you've now been more thoroughly shown to be mistaken.
Somebody found that when he set up positions using 'clear board' that castling is apparently deemed to be illegal by default.

Every day - anyone in the world might make a few or several mistakes.
Every day - a big percentage of people playing a chess game lose that game.
The same percentage win the game - the other percentage 'died'.
Every day - somebody might be exposed in a mistake they made ...
but you don't die !   You Live On !

Progress, heh.

And yet, when everyone seems to be telling you you're wrong, instead of admitting your utter lack of knowledge, you give yourself a pass via the false equivalence of pretending we're all ignorant, and justifying your ridiculous posts as an honest effort in moving the discussion forward. Simultaneously, of course, demonizing those who know more than you as being close minded and intellectually dishonest.

You're a ridiculous person.

I never said 'we're all ignorant'  that's not true.
And you just contradicted somebody else by announcing that the software assigned castling illegal after clearing the board - not by FEN code.
And you're sounding more and more like a person who 'went quiet' a few days ago.  
A member who always wants to personalize.

@llama51 joined chess.com Feb 23rd 2022.  

#1644
How long do Natch or Euclid take to generate a proof game e.g. for this position?

This is of interest to investigate how many of the 10000 randomly sampled positions without excess promotions
https://github.com/tromp/ChessPositionRanking/tree/noproms
are legal and sensible. It is too much work to do that manually.

I think in one of the posts - weren't the 5% or so of the 10,000 all found to be illegal?
Okay - black to move since white's in check.
But we're to assume white starts from bottom of the board ?

#1665
No: of 10,000 positions with allowed multiple excess underpromotions 5% were found to be legal, but none of these can occur in a reasonable game with reasonable moves.
This is another sample of 10,000 positions that Tromp prepared, but this time without excess promotions.
It would be of interest to know if any of these 10,000 can occur in a reasonable game with reasonable moves.

@tygxc
Its black's move - since he's in check.
white pawns are moving up the board?
The idea is to examine that position for legality?

#1667
Yes, the idea is to examine for legality by providing a proof game from the initial position to the given position if possible. Inspection of the accuracy of the proof game then determines if that give position can arrive in a reasonable game with reasonable moves.

Okay.
I'll add that if white's pawns are moving up the board -
then it appears the game could end with Kc4 and Bxd3 double check with mate.  

And if white's pawns are coming down the board - its already mate.
I'll assume for now you're referring to that position.  
If the supercomputers had to generate a 'proof game' for every checkmate position - 
that would be Terrible.
And many of those would probably be Helpmates.

Much better:
Generate checkmate and stalemate positions and hopeless draws and other positions by first excluding all possibilities that couldn't be legal anyway.  Like adjacent Kings.
Then exclude some more by instantly finding them illegal.  In nanoseconds.
In other words - don't worry about how it got there.  

Someday the supercomputers might find a trillion positions that are legal looking on the board but somehow 'could never have got there'.
How many types of such positions could there possibly be?