http://www.xs4all.nl/~timkr/chess2/... [entry 123 near the bottom]

[Quote]

... How Fischer ever got the chess world to accept tying his name to this old idea, is a mystery. Shuffle chess has been known for over 200 years under various names, such as Pre-chess, Baseline Chess, Varied Baseline Chess, Meta Chess, Array Chess and others. The pawns are on the second rank, the pieces are shuffled on the first. Sometimes the arrays must be symmetrical; sometimes the players choose them; sometimes they are determined by lot.

"Fischer Random" is an unpractical form, because a computer is needed to randomize the arrays, and it has some wacky castling rules - the first forcing you to have a computer at hand if you want to play it, the second making it impossible to play it against a computer when no human Fischerrandom player is around.

Advocates of Fischers stolen invention also tend to forget that it sprang from his delusions. He invented it so Kasparov and Karpov couldn't fix their games anymore. He added the randomizing computer because otherwise they could still fix their games. Now if somebody thinks an invasion of purple potatoes from outer space is about to happen, and it can only be stopped if we all swapped our left and right shoes, that's OK with me, but it would be silly to have wrongfooted dancing contests as a result.

Possibly the first time the shuffling idea was mentioned, was in 1792, in the first original Dutch chess book by Philip Julius, Count Van Zuylen van Nijevelt, a Dutch army general, senator of the French Empire, and one-time Napoleonic governor of Amsterdam. He did not like the openings with their boring repetition of patterns, which enabled weaker players to memorize moves with which they could beat stronger ones. He suggested determining the places of the pieces by lot, "because the positions can then be changed infinitely, and it will certainly not be possible anymore to study them beforehand."

His idea does not seem to have caught on, but it was often reinvented. The oldest surviving shuffle games are from 1842; here is a slightly younger one.

Van der Hoeven - Von der Lasa, Baden-Baden 1851
It is not known whether this symmetrical array was determined by chance or by the players themselves. Note the two pairs of Bishops of opposite colors.

[The randomized starting position may be seen here, at entry 123 near the bottom of the page]: http://www.xs4all.nl/~timkr/chess2/...

1.Nf3 b5 2.d4 d6 3.Ba5 f6 4.Ne3 e5 5.Qe1 Bxf3 6.exf3 exd4 7.Rxd4 Ne7 8.b3 Nc6 9.Nd5 Rb7 10.Re4 Bg6 11.Ne7+ Nxe7 12.Rxe7 Nd7 13.Qe6 Qg8 14.Qxg8 Rxg8 15.Bd4 Ne5 16.Bxe5 fxe5 17.Kb2 Kd8 18.Rxc7 Rxc7 19.Rc1 Kd7 20.Bxc7 Kxc7 and White resigned....

...Note: I found the counts shuffle idea in an impressive new book by the Dutchman Henk Mesman - an almost 1000-page treatise of the artistic chess study, from the Arabic beginnings to 1900. I'll come back to it.

[end quote]

(: ♗B :)

You raise some interesting issues, <BishopBerkeley>!

While I agree with Tim Krabbé that FischerRandom is not really new, I do believe it is distinct. In trimming away two-thirds of the possible opening setups with his Rook and Bishop rules (2,880 possible starting positions for Shuffle Chess as opposed to 960 for FischerRandom), and in providing Castling rules (which are absent from Shuffle Chess), Bobby Fischer has created a distinct variation of an old idea with its own peculiar flavor.

Furthermore, I don't know that Bobby Fischer ever tried to present it as anything other than this. Perhaps he did, perhaps he didn't - I just don't know.

I've been playing Shuffle Chess against Shredder (Computer) and Fritz (Computer) lately, and I find it quite invigorating! I'm not sure that FischerRandom would be an obvious improvement. I'm also not sure that the game suffers terribly from the absence of Castling.

Indeed, on the subject of Castling, I note that Dr. Emanuel Lasker , longtime World Chess Champion, had the following to say in discussing "The Fianchetti" openings (the King's Fianchetto: 1.e4 g6; and the Queen's Fianchetto: 1.e4 b6):

"[These are] Ancient Openings, presumably invented at a time before our rule of Castling was introduced, and when another form of Castling, a jump of the King over two squares was lawful. These were sane rules, and it would have been better if they had not been changed. After that unfortunate change, the Fianchetti lost their original purpose of providing a safe square for the King and developing the Rooks. To-day they present a weakness in that they leave the centre of the board in the control of the enemy. True, they do not present targets in the centre either, but a fighter is used to being a target as well as a shot...."

from "Lasker's Manual of Chess" by Dr. Emanuel Lasker, p.88 of the Dover edition, ISBN: 0486206408:
http://www.amazon.com/exec/obidos/t...

I wonder if Shuffle Chess supplemented with the OLD Castling rule would be a good combination?

By the way, Edward Collins provides us with a method of selecting a Fischerandom starting position with a single die (plural: dice): no computer needed!

http://www.edcollins.com/chess/fisc...

(: ♗ Bishop Berkeley ♗ :)

Nihil obstat.

It is actually very easy to get a true random setup consistent with the Fischer-random rules, just using a dice. Do like this:

By the number of a square, I mean the first vacant square on the first row counted from the right. First you decide the postion of the black-squared bishop. There are 4 options. Roll the dice until you get a number between 1 and 4, and place the bishop on the corresponding (black) square. Do likewise with its white-squared fellow (only considering white squares). The bishops has now been placed, and as for placing the remaining pieces, the square color is of no importance. Six squares left, roll for the placing of the queen, then 5 squares for the first knight and four for the last. On the remaining three squares you place the king in the middle, and the rooks on the last two squares.

<It is actually very easy to get a true random setup consistent with the Fischer-random rules, just using a dice.>

<oh guys the thing that "you need a computer in order to shuffle pieces" is so ridicolous!!!! there's a new invention, it's called "die".>

Perhaps the ideal solution would be to have a "rollable Rook": a Rook with a hexagonal base. You just roll the Rook as if it were a cylinder, and one of the hexagonal faces will disclose its tiny, barely visible number, from 1 to 6.

You could call it a Fischerollable Rook!

(Like this, only hexagonal rather than octagonal):

http://www.cosy.sbg.ac.at/~held/pro...

1: "Those dirty stinkin' jews are out to get me."

2: "Bush and the US are mass murderers. They had it coming."

3: "They stole my valuable property and beat me up like thugs."

4: "You know, all those matches between Karpov and that filthy Kasparov were fixed. Every single move was prearranged."

Curiouser and curiouser...

The rules of placement are identical for FRC and ShuffleChess except for the castling. It therefore suffices to consider only the king and rooks.

There are three methods of placing the king and rooks on eight squares: king between the rooks, both rooks to the left of the king and both rooks to the right of the king.

In positions where the king is between the rooks, there are 6 ways of placing the king. The number of rook placements for each king placement is as follows:

King on 2nd: 1 * 6 = 6 ways

King on 3rd: 2 * 5 = 10 ways

King on 4th: 3 * 4 = 12 ways

King on 5th: 4 * 3 = 12 ways

King on 6th: 5 * 2 = 10 ways

King on 7th: 6 * 1 = 6 ways

The total of placements is thus 56 ways.

Now consider the case where both rooks are on one side of the king. There are 6 ways to place the king and the corresponding number of rook placements are:

King on 1st: (7 2) = 21 ways

King on 2nd: (6 2) = 15 ways

King on 3rd: (5 2) = 10 ways

King on 4th: (4 2) = 6 ways

King on 5th: (3 2) = 3 ways

King on 6th: (2 2) = 1 way

The total of placements is thus also 56. The case for the other side is identical.

I had (mistakenly, I think) believed that Shuffle Chess allowed for Bishops on one side to be on the same color squares (thereby ensuring that the Bishops on the opposite side were forced to move in a parallel but separate universe, B X B never being possible). Indeed, I do think this was true in earlier versions of Shuffle Chess (like the one that Tim Krabbé shows in entry 123 near the bottom of this webpage: http://www.xs4all.nl/~timkr/chess2/... ). But in "Shuffle Chess 2880", which appears to be the modern version, I think it is not allowed. (I've been recasting the positions over and over in the Shredder (Computer) interface, and none of them show Bishops on the same side on the same color squares.)

Given the minimal differences between the two games, then, I suppose an elegant divisor is not so very strange!

(: ♗B :)

http://www.100bestwebsites.org/quak...

If I had the opportunity to play a game against Paul Morphy with Morphy setting the condition that it had to be an odds game, this would be the setup I would request (with Morphy playing White).

This gives us a less elegant divisor vis-a-vis FRC:

5040 / 960 = 5.25

Relative to ShuffleChess 2880, we have:

5040 / 2880 = 1.75

We could easily triple the 5040 possibilities without too extreme an alteration of rules (all positions symmetric for White and Black about the horizontal centerline):

Random position 1 of 3:

White Pieces randomized on Rank #1
White Pawns on Rank #2
[This is the standard pattern]

Random position 2 of 3:

White Pieces randomized on Rank #1
White Pawns on Rank #3

Random position 3 of 3:

White Pieces randomized on Rank #2
White Pawns on Rank #3

Double-pawn openings would be allowed in all cases.

Perhaps we would discover some absurdities somewhere in these patterns, some flaw that would give White (or Black) a massive, guaranteed advantage from some of the random deployments. But even in FRC 960, ShuffleChess 2880, or ShuffleChess 5040, I feel confident that not all starting positions present the same balance between White and Black.

Of course, one could enlarge the number of randomization schemes on and on! When the White Pawns are on Rank #3, one could randomly distribute the pieces throughout Ranks #1 and #2!

All fun to think about!

Thanks again, <azaris>!

Of course, I suspect that this project would be rather far down the Chessgames.com To Do List.

Ah, the To Do List, that Tireless Terrible Tyrant!