I would agree that Chess960 doesn't have that much to offer (over classical chess) for weaker players who aren't steeped in opening knowledge to begin with.

My understanding is that the primary benefit of Chess960 was for strong players who want the game to be less dependent on opening preparation, and more dependent on tactical prowess. For example, I read this in regard to Aronian's win over Sokolov in Corus round 13:

"Aronian reached an even score with some impressive preparation against poor Sokolov's Slav. He and his team had prepared the exchange sacrifice on move twenty and indeed had prepared the entire game!"

I think the main point of Chess960 is to eliminate this type of preparation.

And even if there is a PGN standard, it may be that the PGN viewers used by <chessgames.com> don't recognize FRC. (For instance, the viewer must know how to castle in FRC.)

Leko vs Adams, 2001
Leko vs Adams, 2001

[Azaris] I read people saying chess960 positions are "weird". This perception shows how much of fundamental chess has long been hidden from us by traditional chess1. Unfamiliar and non-redundant are different than weird. Any position that arises from logical intelligent play is a proper part of general chess. Such positions are unrelated to the contrived or awkward positions known in chess mate-in-two puzzles and similar.

I believe chess1 is hiding more interesting chess from us than has yet been widely realized.

[YouRange] You wrote "I think the main point of Chess960 is to eliminate this type of [opening] preparation." Most people agree with that, but respectfully I think there are even better reasons to support chess960. Chess960 presents fascinating new dilemmas, and thus provokes fascinating new strategies.

In many setups it is unclear whether a player should hope to castle 'a' wing, 'h' wing, or even not at all. In static chess1 'h' wing castling is overwhelmingly popular.

There are many chess1 books devoted on how to attack the castled king fort. Such books have only a couple of common king fort styles to bother with: (A) Rf1 Kg1 Pf2 Pg2 Ph2 Nf3, or (B) Rf1 Kg1 Pf2 Bg2 Ph2 Nf3 Pg3. In contrast, chess960 leads to a stimulating variety of king fort layouts.

[YouRang] For chess960 two competing standards have emerged for the FEN portion of the PGN specification.

Reinhard Scharnagl's X-FEN emphasizes the maximization of backward compatibility with old-now-inadequate FEN, at the cost of relative complexity.

The competing Shredder-FEN spec says to heck with backward compatibility, in order to create an elegant simple spec. In old FEN the castling eligibilities are noted KQkq. This is how X-FEN usually encodes them too. Shredder-FEN notes them as HAha for the columns on which the eligible rooks stand.

Fritz9 chose Shredder-FEN for its chess960 PGN-FEN.

To me, it would have been more "in the spirit" of chess960 to make a rule like:

(1) Move the king 2 squares to the left (for a-side castling) or 2 squares to the right (for h-side castling). If it can't move 2 squares (because it is on the b or g file), then move it 1 square.

(2) Place the rook on the square next to the king, but on the opposite side.

Wikipedia has long had a comment urging that chess960 kings castle by moving 2 squares, rather than to c1 or g1. However, the 2 squares idea is bad IMHOpinion.

Chess960 and traditional "chess1" both suffer from draw rates that are much higher than we want in sporting competition. The 2 square castling idea would *Decrease* the already too-low rate of Opposite Wing Castling in chess. In chess1 only 8% of games have opposite wing castling, way too low.

When in chess960 the king starts on say g1, there is no way any sane player is going to castle his king to e1. Thus even more often players will be castling to the same wing as each other.

In my chess960 book I describe a better engineered castling rule. In brief, the idea is that the second player to castle should be given a broader choice of king destination squares if-but-only-if he castles to the wing opposite of the wing the first player already castled to. It is still fully fair to both players. It might also increase tension by motivating players to wait just a little longer before castling, which might risk an attack from the enemy on one's own still-centered king.

http://en.wikipedia.org/wiki/Talk:F...

As expected, I see that my proposal is unoriginal. Still, I found myself in agreement with the views of the Wikipedia writer. As he says, someone who learns chess960 apart from chess1 will find the castling rule to be arbitrary and unnatural. From a game-design perspective, it appears to be a flaw.

Chess960 is supposedly a generalization of standard chess. As such, its design should not be dependent on the rules particular to one of its 960 variations (i.e. standard chess). Rather, the reverse should be true.

You stated earlier that "Chess960 presents fascinating new dilemmas, and thus provokes fascinating new strategies." And yet, it seems to me that using the old castling rule will simply cause many games to become unnecessarily similar to chess1, with the king and rook in the familiar castled positions that players are (too) familiar with. Thus, it will continue to hide new strategies.

Your counter argument seems to be based on the idea that opposite wing castling is "too low".

But by what standard should it be considered "too low"? To me, this argument seems to be based on an arbitrary notion of how often it should occur. It's sort of like saying that the rate by which pawns are promoted to knights is too low. Well, it's low for reasons that are natural and germane to the game. However, just the fact that a pawn CAN promote to a knight adds a certain amount of beauty and intrigue to the game, even though it's rare. I think the same can be said of opposite wing castling, even (in some rare situations) castling the king to e1! :)

For chess960 to be a proper and convincing generalization of chess, the rules must be intuitive and simple. The move-king-2-squares proposal is both. The current chess960 rule is neither (except in the special case when the king is on the e file).

Draws are bad for chess or any game, and in chess they are waaayyy too frequent (tho we naturally become used to it). For instance, the phenomemon of the draw has hurt the general public's interest in the World Championship Chess matches more than anything else has.

I think the best way to lower the draw rate is to eliminate the rule that allows players to simply agree to a draw. The idea that the two combatants can quit while the game still appears to be undecided is the thing that's bad for chess (or any game).

I liked the rule used in the last MTel tournament, where draw-by-agreement was not allowed without the consent of the TD.

BTW, I am not saying that the move-king-2-squares (MK2S) rule is the only viable castling rule for chess960. I am saying that the chess960 castling rule should be simple and intuitive (and of course backward compatible with chess1).

In my view, the current chess960 castling rule violates the 'intuitive' part rather badly. The MK2S rule satisfies all the requirements, and I haven't seen another proposal that does so any better.

Before broad implementation of that draw rule, I would like there to be a PocketPC computer program that could assist the TD in making these judgments. I would cringe at seeing arbitors forced to make such controversial decisions unaided.

Professional grandmasters would likely object, saying tournaments already leave them exhausted. Part of the reason they are tired is they have to study the chess1 openings every night after playing a long game, to prepare for tomorrow. Chess960 would allow them to get a good night's sleep instead.

[] Instead or in addition, what about the idea of saying that draw offers expire one move-pair later than the traditionally have?! When you offer me a draw then press your clock, I get to make my reply move and see your reply, before I reach the point where I must either accept the draw or let the offer expire.

This way it would be foolishly risky to offer a draw in a position that still has life. Your opponent could make a speculative move, then wait to see whether your exploit its weakness or overlook its strength.

Regarding the rule to require permission from the TD for any draw-by-agreement, you said, <I would like there to be a PocketPC computer program that could assist the TD in making these judgments. I would cringe at seeing arbitors forced to make such controversial decisions unaided.> I could be wrong, but I think the idea is that the TD would simply deny the draw request if the game wasn't clearly drawn from a layman's perspective.

I'm concerned that involving a computer could sometimes be a form of TD interference. Suppose both players think it's a draw, but the computer sees that the player-to-move has a dazzling winning tactic! The TD (after consulting the computer) would deny the draw request, inadvertantly hinting to the player that he has a winning move! :-D

<Chess960 would allow them to get a good night's sleep instead.> Yes, I think this is one of the most appealing aspects of chess960 -- avoid all of the "pre-game" play. The winner will be the player who performs best over the board, not the one who did the best job of studying the opposing player's openings.