Latest Stockfish with old time managment

[syzygy]

That would have totally crippled the variant by effectively removing castling.

Maybe for humans.

For engines the variant could have been way more popular since everyone could participate, instead we have a "you must add extra code for entry."

But Fischer created the variant to relieve human chess players of having to spend most of their time memorising thousands and thousands of lines just to be able to stay competitive.

For a computer-friendly game, first get rid of en passant, then of castling.

Another criticism is insisting on mirrored starting position, a lot more could be playable if you let white and black to roll positions individually, like these jewels.

Without this rule, many openings would be too unbalanced.

[Ovyron]

But Fischer created the variant to relieve human chess players of having to spend most of their time memorising thousands and thousands of lines just to be able to stay competitive.

How much success did it have? I think that pretty much any chess variant serves those purposes, you can jump right in into any variant and be competitive just because their variations haven't been explored in depth as in chess.

Lichess is currently the place where most humans play chess variants, and you can get a game on demand for any variant very quickly, that's the kind of place where Chess960 should thrive, instead you have variants like Crazyhouse and Atomic being more popular.

What I'm saying is that if from the get go Chess960 was backward compatible with chess, it's possible it could have been more popular. The weird castling rules that it has make it seem alien and unlike chess (people can't even agree to a standard, Shredder has some implementation, Arena another. How many GUIs have you castling by clicking on the King, then on the Rook? Not all, it's a nightmare and counter-intuitive, you have to learn how the GUI you're using expects you to input castling, and thus far, even more alien-looking "captured piece swaps color and can be dropped by capturer" has fared better.)

For atomic chess you can jump right in without knowing the rules, lose a few games and get what's it about, then you can beat some beginner already, because without explanation, its quirks are self-evident (say, you learn very fast that putting the opponent in check doesn't stop anything if someone captured a piece adjactent to your king.) In Chess960, in contrast, as soon as the opponent castles you'd ask "how did they do that?", and would need to read the rules to figure out castling rules, because they're a mess (can a king castle if it's already on the square he'd arrive at after castling? Is it still castling if only the rook moves through the king? The answers are arbitrary.)

Another criticism is insisting on mirrored starting position, a lot more could be playable if you let white and black to roll positions individually, like these jewels.

Without this rule, many openings would be too unbalanced.

The "jewels" I linked to are the most unbalanced starting positions out there, and even, engines have no idea who is winning and give 0.00 scores to positions where one side has an edge, or have the advantage of the sides backwards, they don't know what's going on.

And honestly, they seem very fun to play, what is unbalanced regarding white win percentage anyway? Atomic chess has some 66% win advantage for white in the opening position and that doesn't stop the fun.

Having humans in those positions would just add to the opening as they have to figure out who has the advantage, who must be attacking and defending, and why. They seem very rich starting positions.

[syzygy]

But Fischer created the variant to relieve human chess players of having to spend most of their time memorising thousands and thousands of lines just to be able to stay competitive.

How much success did it have?

It was hugely successful. It was so successful, that others appropriated the variant and gave it another name. FIDE has now adopted Fischer's variant in the "Fide Laws of Chess", calling it "Chess960".

What I'm saying is that if from the get go Chess960 was backward compatible with chess, it's possible it could have been more popular.

So we have diametrically opposed views on the question whether chess without castling could have gained any popularity as a serious variation of the regular game of chess.

How many GUIs have you castling by clicking on the King, then on the Rook?

So you're saying castling should never have been included in the rules of chess because those people in the 16th century should have realised the problems it would create with GUIs.

[Dirt]

There are 18(?) FRC opening positions that leave the rooks and the king in the standard starting positions. Would that make it easier to set up engines to let them play chess18?

Just as a stopgap, of course.

[Ovyron]

Let me exemplify what I mean because it hasn't been made clear at all.

You should be able to take a look at a position and be able to analyze it and make objective statements about it, but you're not. Here's such a position:

[d]r5kr/pQ3pp1/3P3p/2b5/b5N1/1ppB4/q4P1P/3RR1K1 b ha -

Hmm, it seems that white is winning, After 1...Rd8 Bc4 a move like 2...Bd7 loses to 3.Bxf7 Kf8 and it's a mate in 7. Black has actually to play 2...Bxf2+, but it seems this defensive attempt will fail as well...

So, what does black play?

1...O-O

[d]r4rk1/pQ3pp1/3P3p/2b5/b5N1/1ppB4/q4P1P/3RR1K1 w - -

Now all of black's problems are solved, and the tables have been turned around, white's d pawn is dropping and it would be lucky if it manages to draw this...

See the problem?

This doesn't look like chess at all, the best move could as well have been 2...Bxf2# making the king explode like in Atomic Chess, or 1...N@a3+ dropping a knight in Crazyhouse, or whatever.

Allowing the king to castle without moving is unnatural and arbitrary, Fischer could as well have let the black king castle with the g rook in this position:

[D]r5rk/pQ3pp1/3P3p/2b5/b5N1/1ppB4/q4P1P/3RR1K1 b D -

Why not? Who says the king must start between the two rooks? Arbitrariness.

Heck, when composing these positions my GUI asked me if the white king was able to castle with the d rook or with the e rook, something is wrong with your variant if you need to specify more than if O-O and O-O-O are available for the sides.

Chess960 isn't the most played chess variant and its lack of backward compatibility with chess (that could have been an edge) played a factor.

[syzygy]

See the problem?

Not really, no.

[Ovyron]

Not really, no.

The problem is that 99% of engines can't play O-O in that position, and 80% of GUIs can't even allow an engine that can to play it in such a position, and for those that do, a "switch" has to be turned on so they're aware of the posibility.

That was the reason this discussion was started, Vasik Rajlich supported Chess960 in Rybka but removed the code in later versions, that the code has to exist in the first place is a problem, when the only thing the variation should require is the shuffling of the starting positions.

There's a problem when, instead of more and more programmers supporting the variant in their engines, some start to remove it from their engines, even though in its inception the variant could have allowed free entry for all chess playing entities of the future, not just humans.

[Uri Blass]

But Fischer created the variant to relieve human chess players of having to spend most of their time memorising thousands and thousands of lines just to be able to stay competitive.

Humans do not need to spend most of their time memorizing thousands of lines to stay competitive.

I think that it is generally a mistake to do it.
top programs can easily beat all humans when they get out of book in 3 moves.

knowing how to play without memorizing is more important and humans probably memorize nothing after lines like 1.e4 e6 2.d4 d6 when they will probably lose them with white against top chess programs.

[Ras]

Humans do not need to spend most of their time memorizing thousands of lines to stay competitive.

They do, at least top players. It is not uncommon to see the first 20 moves in a human match be played by memory. It's also common to prepare deep lines at home and force the opponent to deal with it under time pressure. If nothing else, the prepared one will have a time advantage at the end of the opening.

[syzygy]

But Fischer created the variant to relieve human chess players of having to spend most of their time memorising thousands and thousands of lines just to be able to stay competitive.

Humans do not need to spend most of their time memorizing thousands of lines to stay competitive.

Tell that to Fischer

Fischer's goal was to eliminate what he considered the complete dominance of openings preparation in chess today, replacing it with creativity and talent.

I don't know when, but I think we are approaching that [the end of chess] very rapidly. I think we need a change in the rules of chess. For example, I think it would be a good idea to shuffle the first row of the pieces by computer ... and this way you will get rid of all the theory. One reason that computers are strong in chess is that they have access to enormous theory [...] I think if you can turn off the computer's book, which I've done when I've played the computer, they are still rather weak, at least at the opening part of the game, so I think this would be a good improvement, and also just for humans. It is much better, I think, because chess is becoming more and more simply memorization, because the power of memorization is so tremendous in chess now. Theory is so advanced, it used to be theory to maybe 10 or 15 moves, 18 moves; now, theory is going to 30 moves, 40 moves. I think I saw one game in Informator, the Yugoslav chess publication, where they give an N [theoretical novelty] to a new move, and I recall this new move was around move 50. [...] I think it is true, we are coming to the end of the history of chess with the present rules, but I don't say we have to do away with the present rules. I mean, people can still play, but I think it's time for those who want to start playing on new rules that I think are better.

OK, the "most of their time" part may have been exaggerated, but still. Fischer created the games for humans, not for computers.