Alternatives to Classical Chess

@dboing said in #4:

1 (non-random, random over 960 does not allow learning over many games the subtleties in the long run about the backrank geometries))

I think it can allow for learning, over many games, the subtleties of various setups. But it doesn't allow for move memorization.

Perhaps the notions of intuition building are not really digested in the world of random 960. Or it is not an acceptable thing, to be able to learn from experience about the back-rank associations with downstream position characteristics.

Intuition building is a part of Chess960. As you play many starting positions that have similar (but maybe not exactly the same) starting positions, you begin to see the different strengths and weaknesses of the positions. Move memorization might not be a big part of the game, but knowledge and understanding of different types of positions as well as chess principles are a big part of the game. Arguably bigger than in chess because there are new ideas and players can't rely on rote memorization.

@CheerUpChess-Youtube said in #10:

I won't argue again with you, as we already did, just telling you that this is wrong and Fischer has said it himself in the interview. There has always been a lot of theory and book moves, fromout the beginning - not as much as there is today, but there has been theory. Always. Chess has NEVER been a game of pure creativity. There have been chess principles, chess theory, book moves, known openings, a lot of studying and preparation also back then.

You're just flat out wrong. Chess may have existed for a thousand years before the beginning of theory.

There's some evidence that chess may have been created around the 6th century: https://www.britannica.com/topic/chess/History

"The history of chess goes back almost 1500 years. The game originated in northern India in the 6th century AD and spread to Persia."

https://simple.wikipedia.org/wiki/History_of_chess

"The earliest printed work on chess theory whose date can be established with some exactitude is Repeticion de Amores y Arte de Ajedrez by the Spaniard Luis Ramirez de Lucena, published c. 1497"

"The first author to attempt a comprehensive survey of the openings then known was Aaron Alexandre in his 1837 work Encyclopédie des Échecs. According to Hooper and Whyld, "[Carl] Jaenisch produced the first openings analysis on modern lines in his Analyse nouvelle des ouvertures (1842-43)."

https://en.wikipedia.org/wiki/Chess_theory

So some rudimentary theory may have started around ~1500 AD, but theory really started taking hold around the 1800's. Chess is far older than that.

@Prophiscient said in #11:

I think it can allow for learning, over many games, the subtleties of various setups.

Well it would be nice to start matching human time scale knowledge to that statement. It would also need a question of similarity definition . But that can be convened or iterated. Which i find the op allowing by its initial gamut of directions (iteration).

I glad to see though a common goal being mentioned.

I am also not into what i call move-chess. (not that moves don,t exist, don't get me wrong, hahaha, but as shorthand for what has become a chore of chess at high level, and percolating down to us (maybe not you) amateurs only trying to understand chess more from a position information focus. Rather than advantageous tunnel sequences (if cooperatively followed, consciously or not, the not maybe fashion induced for example, or engine style induced).

And I enjoy going to 960 (only from position, where i can start a daily or weekly scale accumulating experience of the lost kind).
sometimes i let the third player random choose.. but then i would rematch.. many times....

memory, is a bigger thing. one need to be careful about just saying "memorization".. I think it is about remember sequences of moves more than remembering sequence of positions. one can abstract the positions into similarity blobs, the moves are very concentrated information bits that are critical to how the positions evolve..

off game versus in-game.. what do you come with in limited storage.. do you have a long term memory type (intuition or some seamless mixture of intuition and theory) of position evolution. or move only sequences.. One could think that with limited in-game storage to prepare for, learning only the move sequences or a few of them might be more advantageous that using that space for sequences of positions. And it seems that tournament chess, without all the extra bells and whistles above the core game, has evolve toward reward only one strategy of memorization (=learning, but not the rational (wide sense above) kind, the imitation kind).

@dboing said in #13:

Well it would be nice to start matching human time scale knowledge to that statement. It would also need a question of similarity definition . But that can be convened or iterated. Which i find the op allowing by its initial gamut of directions (iteration).

It is human time scale. The type of knowledge we get from 960 comes in the form of principles (king safety, control the center, develop the pieces, have good pawn structure, etc). This is common knowledge between the old chess and Chess960. The difference between them is that the randomness of Chess960 prevents rote memorization. And this is how chess was made to be played and was played for a thousand years.

The ancients didn't memorize their moves. Sure, they understood principles. But the modern move memorization is a relatively modern phenomenon that has gotten orders of magnitude worse due to the modern chess engines.

Chess960 brings us closer to the way the ancients used to approach the game. This is why I think it's better. Creativity should come first in a game of creativity.

@CheerUpChess-Youtube said in #3:

I would like 960 chess, but with improvements: castling should look more understandable and less weird, bishops or knights shouldn't start in the corners of the board, the pawns should maybe be randomized a bit too (for example using the 3rd/6th rank aswell), there should be no weaknesses or unprotected pawns in the starting position.

You're listing these things as if they're problems with Chess960, but you're not explaining WHY they're problems. What is wrong with bishops or knights in the corners? What is wrong with the castling? Pawns shouldn't be in different places because choosing your pawn structure is a fundamental part of chess. Unprotected pawns can be protected.

It seems like you just don't like the new tactical and strategic dynamics of diverse starting positions. That's fine, but that isn't a problem with Chess960. It's just a product of your mindset.

@Professor74 said in #1:

If you had to promote any alternatives to Classical Chess, having to choose some of the following options, which would you prefer?

1. Chess960
2. Chess18 (uses only the starting positions from Chess960 in which the king and the rooks are placed on the same home squares as in Classical Chess)
3. No Castling Chess
4. Reform Chess (boards-size 8x6, 6x8, 9x6, 5x8)
5. Double-King Chess
6. Double-King Chess960

I would promote Chess960 and Double-King Chess960. The former to revitalize the old chess and the latter for exhibition tournaments.

@njswift said in #2:

But honestly I think it would be a good idea to just eliminate castling from 960 before it becomes more popular, because it's even more complicated and sometimes even absurd.

Honestly, I thought castling in Chess960 was problematic at first, too. I even thought about Chess204 which is the subset of 960 positions where the king starts on the e-file.

However, I realized that the castling in 960 is actually great. The castling rules aren't as crazy as it seems. The king and rook just go to the same end positions as the old chess with the same basic restrictions (pieces can't be in between the king and rook or their start and end squares, no pieces checking squares between the king's start and end squares, etc.)

If you're able to remember the end squares of castling in the old chess, understanding the 960 castling rules is actually very simple. I think it's also important to keep castling to make 960 as chess-like as possible.

Also, I think the castling in 960 introduces much more interesting possibilities.

The main issue I have with the castling is that the end squares feel contrived just to copy the old chess. But Chess960 is a progression of the old chess, so I think it's fine.

No Castling Chess is the most interesting option for me. I've played it several times.

@Prophiscient said in #17:

Honestly, I thought castling in Chess960 was problematic at first, too. I even thought about Chess204 which is the subset of 960 positions where the king starts on the e-file.

How many such characterized subsets of 960 are out there? not the engine valuations ones only, but also those with human geometric concepts, as this Chess204 (actually same question, I had for op chess18).

No Castling chess looks like the only viable alternative at higher levels.

Reforming chess would also work, but it's difficult to find an agreement. Capablanca already tried back in the days, with no luck.

Chess960 isn't really suited for classical time control, because of the randomness it introduces in tournament formats. Although no position is crazily unbalanced, some starting setups give White an even bigger advantage than classical chess, while others are equal. This means that, if starting positions change every game, there could be someone who gets luckier than other players (e.g. always getting the best positions as White and the equal ones as Black), and this would give them an edge. In order to prevent that, you should have every player play each position both with White and Black... but this means that every player, after the first game, could prepare their second game using an engine in-between rounds, which goes against chess960 spirit.

The only way to solve the problem is to have shorter rounds (e.g. rapid), so that players can immediately play their second game with the same position after their first one, without having time to prepare for it. But I don't see chess960 as a viable alternative to traditional chess in slower time controls.