World Fischer Random Championship

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Chessqueen
Posts: 1043
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: World Fischer Random Championship

Post by Chessqueen » Wed Oct 30, 2019 6:25 pm

Chessqueen wrote:
Wed Oct 30, 2019 5:50 pm
Nordlandia wrote:
Wed Oct 30, 2019 5:08 pm
lkaufman wrote:
Tue Oct 29, 2019 8:48 pm
Ovyron wrote:
Tue Oct 29, 2019 12:46 pm
So, if after rolling one of the 960 positions you get the opening one from chess, do people just sit and play a normal game?

If not, perhaps the best name for the game would be Fischer 959 :mrgreen:
Yes, the rules say that you just play normal chess in that case, although I think that some frc events have added the rule to disallow normal chess, so they are playing chess 959, not 960.
What about the mirror position. RNBKQBNR. Will FRC events accept that position?

Sorry but the position you posted is #534 in Chess960, which is NOT exactly the standard or normal position # 518 in Normal Chess :shock:
https://www.chess.com/blog/friscodelros ... sition-534

Our Normal chess is position #518 in Chess960, also notice that playing our Normal Standard Chess Position in Chess960 the Castling is done differently watch this video carefully, you can NOT castle by simply picking up the King and dropping it on the g square, you have to drop the King on top of the Rook sitting on the h square :roll:
https://www.youtube.com/watch?v=9dbAHd5Yd5k

lkaufman
Posts: 4290
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: World Fischer Random Championship

Post by lkaufman » Wed Oct 30, 2019 6:35 pm

Chessqueen wrote:
Wed Oct 30, 2019 5:50 pm
Nordlandia wrote:
Wed Oct 30, 2019 5:08 pm
lkaufman wrote:
Tue Oct 29, 2019 8:48 pm
Ovyron wrote:
Tue Oct 29, 2019 12:46 pm
So, if after rolling one of the 960 positions you get the opening one from chess, do people just sit and play a normal game?

If not, perhaps the best name for the game would be Fischer 959 :mrgreen:
Yes, the rules say that you just play normal chess in that case, although I think that some frc events have added the rule to disallow normal chess, so they are playing chess 959, not 960.
What about the mirror position. RNBKQBNR. Will FRC events accept that position?

Sorry but the position you posted is #534 in Chess960, which is NOT exactly the standard or normal position # 518 in Normal Chess :shock:
https://www.chess.com/blog/friscodelros ... sition-534

Our Normal chess is position #518 in Chess960
https://www.youtube.com/watch?v=9dbAHd5Yd5k
Because of the special castling rules, the "mirror" position 534 is fundamentally different from normal chess, it's not just a cosmetic difference, so there is no reason to discard that position.
Komodo rules!

User avatar
Ovyron
Posts: 4373
Joined: Tue Jul 03, 2007 2:30 am

Re: World Fischer Random Championship

Post by Ovyron » Wed Oct 30, 2019 6:57 pm

And that's why I don't like Chess 960 per se. Castling rules would be better if they just had king moving two spaces towards an unmoved rook that has from him the same distance as in normal chess, then we'd not have those weird instances where castling doesn't even move the king and it's just the rook that goes through him :shock:

It's unnatural that kings and rooks end in the same squares as in normal chess after castling in completely different positions, and the position that is the mirror of the opening one highlights it (the king will move 1 square when castling queenside and 3 squares when castling kingside? Isn't that unnecessarily confusing and shouldn't it just behave like a mirrored normal chess?)

But I guess it's too late to change it...

lkaufman
Posts: 4290
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: World Fischer Random Championship

Post by lkaufman » Wed Oct 30, 2019 7:30 pm

Ovyron wrote:
Wed Oct 30, 2019 6:57 pm
And that's why I don't like Chess 960 per se. Castling rules would be better if they just had king moving two spaces towards an unmoved rook that has from him the same distance as in normal chess, then we'd not have those weird instances where castling doesn't even move the king and it's just the rook that goes through him :shock:

It's unnatural that kings and rooks end in the same squares as in normal chess after castling in completely different positions, and the position that is the mirror of the opening one highlights it (the king will move 1 square when castling queenside and 3 squares when castling kingside? Isn't that unnecessarily confusing and shouldn't it just behave like a mirrored normal chess?)

But I guess it's too late to change it...
If the rules made mirrored positions behave similarly, then it would be effectively chess 480, not 960. I prefer the larger number myself, but there are arguments both ways.
Komodo rules!

User avatar
Ovyron
Posts: 4373
Joined: Tue Jul 03, 2007 2:30 am

Re: World Fischer Random Championship

Post by Ovyron » Wed Oct 30, 2019 7:40 pm

But the rules of chess castling state:

"Castling consists of moving the king two squares towards a rook on the player's first rank, then moving the rook to the square over which the king crossed."

What is happening on Chess 960 isn't even castling. It is considered a King's move even on games when it doesn't move?

If the only difference between 480 positions and the other 480 is their weird castling rules (and mirror) then it has been revealed to me that most people wrongly think it's twice as varied as what it mostly is.

lkaufman
Posts: 4290
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: World Fischer Random Championship

Post by lkaufman » Wed Oct 30, 2019 9:40 pm

Ovyron wrote:
Wed Oct 30, 2019 7:40 pm
But the rules of chess castling state:

"Castling consists of moving the king two squares towards a rook on the player's first rank, then moving the rook to the square over which the king crossed."

What is happening on Chess 960 isn't even castling. It is considered a King's move even on games when it doesn't move?

If the only difference between 480 positions and the other 480 is their weird castling rules (and mirror) then it has been revealed to me that most people wrongly think it's twice as varied as what it mostly is.
It comes down to whether you consider the process of castling or the result to be more important. I side with Fischer on this; castling has a purpose, and castling according to the normal rule won't accomplish this purpose in many of the 960 positions. Actually I think that the castling rules of normal chess are illogical, in that queenside castling should result in Kb1, Rc1 (as many novices think it does) as this is the mirror of kingside castling and is a much more likely stable formation than Kc1, Rd1. This change would make queenside castling more common, and would make chess less drawish I think, but I know it's not going to happen.
Komodo rules!

duncan
Posts: 11512
Joined: Mon Jul 07, 2008 8:50 pm

Re: World Fischer Random Championship

Post by duncan » Wed Oct 30, 2019 9:41 pm

lkaufman wrote:
Wed Oct 30, 2019 3:29 pm
duncan wrote:
Wed Oct 30, 2019 12:54 pm
lkaufman wrote:
Tue Oct 29, 2019 8:55 pm
It seems to me that Fischer Random is working out better than I expected; I thought it would solve only the preparation problem in chess, but it seems to be doing wonders for the draw rate as well, as has already been shown in computer chess in the CCRL testing.
Do you have any thoughts on why Fischer Random should reduce the draw rate. Is it because without the 100 year background of opening moves research, moves are objectively worse leading to many more errors in the opening?
Yes, that's right. The more moves players have to make on their own, without the benefit of knowing what others have played or what computers recommend, the more chances for mistakes, and since chess (both standard and frc) is almost certainly drawn without mistakes the more chance for error, the more chance for a decisive result.
Do you have any idea what the percentage of draws komodo gets in self play comparing Fischer Random, regular chess without opening book?

todd
Posts: 56
Joined: Thu Apr 19, 2018 7:09 pm

Re: World Fischer Random Championship

Post by todd » Wed Oct 30, 2019 10:03 pm

duncan wrote:
Wed Oct 30, 2019 12:54 pm
Do you have any thoughts on why Fischer Random should reduce the draw rate. Is it because without the 100 year background of opening moves research, moves are objectively worse leading to many more errors in the opening?
As others pointed out, this is true.

However, also, having to think earlier in the game effectively reduces the amount of thinking time per move. In general, draw rate goes up as the quality of play goes up. So more thinking time per move usually means more draws (and vice versa).

lkaufman
Posts: 4290
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: World Fischer Random Championship

Post by lkaufman » Wed Oct 30, 2019 10:17 pm

duncan wrote:
Wed Oct 30, 2019 9:41 pm
lkaufman wrote:
Wed Oct 30, 2019 3:29 pm
duncan wrote:
Wed Oct 30, 2019 12:54 pm
lkaufman wrote:
Tue Oct 29, 2019 8:55 pm
It seems to me that Fischer Random is working out better than I expected; I thought it would solve only the preparation problem in chess, but it seems to be doing wonders for the draw rate as well, as has already been shown in computer chess in the CCRL testing.
Do you have any thoughts on why Fischer Random should reduce the draw rate. Is it because without the 100 year background of opening moves research, moves are objectively worse leading to many more errors in the opening?
Yes, that's right. The more moves players have to make on their own, without the benefit of knowing what others have played or what computers recommend, the more chances for mistakes, and since chess (both standard and frc) is almost certainly drawn without mistakes the more chance for error, the more chance for a decisive result.
Do you have any idea what the percentage of draws komodo gets in self play comparing Fischer Random, regular chess without opening book?
We never test komodo in normal chess with no opening book, because the variety would be too limited. I suppose you mean to test with four threads or so in order to rely on MP to produce some variety, but it wouldn't be enough to be meaningful. We have tested with Kai's three move book, which is about as close to no book as you can get for a meaningful test. It does reduce the draw percentage, but not by the large amounts we see with frc. One reason for less draws in frc (in engine play) may be that same side castling is not so overwhelmingly likely as in normal chess.
Komodo rules!

duncan
Posts: 11512
Joined: Mon Jul 07, 2008 8:50 pm

Re: World Fischer Random Championship

Post by duncan » Wed Oct 30, 2019 10:38 pm

lkaufman wrote:
Wed Oct 30, 2019 10:17 pm
duncan wrote:
Wed Oct 30, 2019 9:41 pm
lkaufman wrote:
Wed Oct 30, 2019 3:29 pm
duncan wrote:
Wed Oct 30, 2019 12:54 pm
lkaufman wrote:
Tue Oct 29, 2019 8:55 pm
It seems to me that Fischer Random is working out better than I expected; I thought it would solve only the preparation problem in chess, but it seems to be doing wonders for the draw rate as well, as has already been shown in computer chess in the CCRL testing.
Do you have any thoughts on why Fischer Random should reduce the draw rate. Is it because without the 100 year background of opening moves research, moves are objectively worse leading to many more errors in the opening?
Yes, that's right. The more moves players have to make on their own, without the benefit of knowing what others have played or what computers recommend, the more chances for mistakes, and since chess (both standard and frc) is almost certainly drawn without mistakes the more chance for error, the more chance for a decisive result.
Do you have any idea what the percentage of draws komodo gets in self play comparing Fischer Random, regular chess without opening book?
We never test komodo in normal chess with no opening book, because the variety would be too limited. I suppose you mean to test with four threads or so in order to rely on MP to produce some variety, but it wouldn't be enough to be meaningful. We have tested with Kai's three move book, which is about as close to no book as you can get for a meaningful test. It does reduce the draw percentage, but not by the large amounts we see with frc. One reason for less draws in frc (in engine play) may be that same side castling is not so overwhelmingly likely as in normal chess.
Is it possible to ban same side castling in normal chess, or would that be unfair to black.?

Post Reply