David Paulowich wrote on Sat, Sep 8, 2001 12:00 AM UTC:Excellent ★★★★★
This is David Paulowich writing in support of the standard rules for pawn promotion, which seem to be unpopular with some players and chess variant designers. If, for example, pawns could only be promoted to previously captured pieces, then many beautiful games would no longer be legal. My databases contain over 400 games with 4 Queens on the board, 2 White and 2 Black, including: Capablanca - Alekhine, 1927 (Thirteenth World Chess Championship Match, game 11) and Borsony - Koch, 1956 (Second World Correspondence Chess Championship). In 1936 Reinle checkmated Lange in this 'extra promotion' game: 1. e4 e5 2. f4 f5 3. exf5 e4 4. Qh5+ g6 5. fxg6 h6 6. g7+ Ke7 7. Qe5+ Kf7 8. gxh8=N#
Such games, with one player having nine pieces other than pawns, used to be rare (only ten were played between 1856 and 1963). In modern times the opening: 1. b4 e5 2. Bb2 Bxb4 3. f4 exf4 4. Bxg7 Qh4+ 5. g3 fxg3 6. Bg2 gxh2+ 7. Kf1 hxg1=Q+ 8. Kxg1 (from Kucharkowski - Walter, 1982) has been repeated in over 200 games. Incidentally, White is winning, by about 150 to 50.
I am pleased to see the PBM Preset for Unicorn Great Chess is still available on this website. I had been planning to post a rules page with a Zillions game file several years ago, but did not get around to it before my hard drive crashed. Anyway, here are some brief notes on this variant.
Unicorn Great Chess adds four 'Lions' to the board - on the second and ninth ranks. The Lion is identical to Ralph Betza's 'Half-Duck' (HFD, a piece that can move one square diagonally or jump two or three squares orthogonally), from the Remarkable Rookies army. It was designed to be at least as strong as the Rook on the 8x8 board, but it will be weaker on larger boards. I chose the name Lion as a reminder of the 3-square leaper in Grande Acedrex (1283). That game also has a piece called the Unicorn (or Rhinoceros). The remaining ten White pieces are arranged CRNBUKBNRQ on the first rank. Placing the Chancellors and Queens in the corners (with the Pawns in front of them also defended by the Knights) is an opening setup I have used on 64, 80, and 100 squares.
My somewhat rusty memory tells me that a King and Lion can force checkmate against a lone King. 'A Pawn promotes to a Chancellor, Queen, or Unicorn of the same colour upon reaching the player's tenth rank.' Perhaps I should amend that rule to also allow promotion to a Lion, if desired. Note that the Lion has the longest 'single leap' of any piece on this board, which makes it capable of delivering a Fool's Mate:
1. Lion f2-f5, Lion f9-f6
2. Pawn e3-e4, Lion f6-f4 (mate)
My suggested game values (allowing for Pawns being a little weaker on the 10x10 board) are as follows: Pawn 100, Knight 300, Bishop 350, Lion 450, Rook 550, Chancellor 900, Queen 1000, Unicorn 1000. [EDIT] I decided to bump the Lion up to 500, see the Rules section on my Unicorn Great Chess page.
I firmly believe that Q = R+B+P and Q+P = R+R on 64 or 100 squares. My (limited) playing experience on 64 squares leads me to conclude that a Chancellor is worth a Rook and Knight and half a Pawn. So I used that formula to calculate a 900 point value in this game. WARNING: the Lion and Unicorn values are mostly based on guesswork. It certainly would be convenient if the Chancellor was worth exactly two Lions (or a Rook and a Bishop). Other, more experienced, chess variant players may come up with different values. --- David Paulowich
David Paulowich wrote on Sun, Aug 1, 2004 11:51 PM UTC:Excellent ★★★★★
Every chess variant can raise complicated rules questions. Here are some for Shatranj.
First Position. White: King e5, Knight e1 Black: King e3, Rook a1
MOVES 1. Nc2 check, Kd3 2. Nxa1, Kc3 3. Ke4, Kb2 4. Kd3, Kxa1 is a simple 'two bare Kings draw' in FIDE chess. I wonder if centuries ago there were Shatranj tournament rules concerning bare King draws that require more than one move.
Second Position. White: King c1, Knight e1 Black: King a1, Pawn a2, Rook e2
MOVES 1. Nc2 check, Rxc2 check 2. Kxc2 stalemate(?) Applying the Bare King rule exactly as stated, White lost the game before he could capture the Rook and win by stalemate. This seems unfair. But if the rules did allow White to play his second move, should a stalemate by a bare King count as a win or only a draw?
Try having the Kings and Rooks move exactly three squares when castling. Thus 'White O-O-O' ends up with Kc1 and Rd1. Also, I wish Capablanca had given the simple and elegant promotion rule: to Queen, Marshal(l), or Janus of the same color. There is absolutely no need for underpromotion in these games - except to avoid the bother of borrowing a piece from another set.
We need to look at the geometry of the chesspieces. Rook, Knight, and Bishop each attack different squares. Let us start with one exceptional case: if promotion to any one of Rook, Knight, Bishop results in a stalemate, than so does promotion to any of the compound pieces Amazon, Marshall, Queen, or Janus. Assuming that the situation is not that bad: [1] if promoting to a Rook stalemates - choose a Janus, [2] if promoting to a Knight stalemates - choose a Queen, [3] if promoting to a Bishop stalemates - choose a Marshall.
<p> As for the Amazon - it is just too strong for my taste. I prefer to have just the three pieces (Marshall, Queen, Janus) to choose from. Actually I use the Unicorn (Bishop + Nightrider) instead of the Janus (Bishop + Knight) in my own chess variants.
Michael Howe raised a valid point for all variants that do not have a set of 'perfect promotion choices.' In FIDE chess, there is the common situation in King-and-Pawn endings where promotion to a Queen stalemates and promotion to a Rook does not - but a master will probably have resigned before this happens on the board. Also, sometimes you need to promote to a Knight and give check. For example:
45. Ra6 f1=N+ 46. Ke2 Rxd2+ 47. Kxf1 Rxc2 1/2-1/2 Promoting to a Knight (with check) on move 45 was the only way to avoid losing, according to my computer.
David Paulowich wrote on Sun, Aug 22, 2004 02:32 AM UTC:Excellent ★★★★★
Eric Greenwood uses this piece in his 80-square variant 'Eric's Great Chess', calling it the Giraffe. In his 84-square variant 'TamerSpiel', a Lion promotes to a Warlord (=Amazon).
Definitely one of the weaker pieces. Dave McCooey writes (in his Endgame statistics with fantasy pieces)
'Two Camels, even on different colors, cannot checkmate a lone King.
A Camel and a Zebra cannot checkmate a lone King.
A Camel and a Wazir can checkmate a lone King, which is surprising. It can take as long as 77 moves (154 halfmoves) to force mate.'
Perhaps inventors would find the Camel-Dabbabah combination to be a more useful piece. Still limited to one color, but jumping two squares orthogonally gives the piece 'close-range' power.
NeuroDoc is correct - perpetual check is 'chess slang', commonly used for a forced sequence of moves. Note for computer programmers: any time a position repeats five times in a game, it must be true that the same position with the same player to move has repeated at least three times. Just mentioned this in case a plain 'fivefold repetition rule' is easier to code.
<p>In my database is the game [R. Pert - M. Franklin, 1996] in which both players have two rooks on the board. Black sets up a possible stalemate position on move 33 by advancing his passed Pawn to h3. All White needs to do is sacrifice both Rooks. After 21 consecutive Rook checks, they agreed to a draw. Of course, Black can always end the checks (and stalemate White) by capturing the last Rook.
In some endgames the weaker side can sacrifice pieces to achieve a stalemate draw. I just posted a comment to 'Perpetual check explained with animated gif' in 'The rules of chess', citing a game that ended with 21 consecutive Rook checks. The 50-move rule is probably required to actually force an end to that game.
<p>I like the value classes! The CardinalRider (Unicorn) and ChancellorRider (Varan) are listed in Abecedarian Big Chess (ABChess) on this site. Back in 1991, G. P. Jelliss called these pieces the Banshee and the Raven.
Try P=1, N=3, B=3, R=5, N+B=7, N+R=8.50, Q=9
<p>Also NN (nightrider) = 4.75, while NN+B (unicorn) = 8.75
<p>I invented Unicorn Great Chess to show off the power of a unicorn on a 10x10 board, but the Queen still looks a little bit stronger on the usual 64 squares. If we were to accept Sam's high value for the N+B, then the Unicorn (worth at least 1.50 pawns more) would get pushed higher than the Queen. I know his value is popular, but I am not convinced that it is correct. As for the colorbound camel+bishop piece, one pawn more than a nightrider seems generous. Not every compound piece deserves the same 'one pawn value' bonus as the queen and unicorn.
The Encyclopedia Britannica (1911) says: 'The chancellor of an order of knighthood discharges notarial duties and keeps the seal.' Sounds good to me!
<p>Looks like the dates for Carrera's Chess and The Duke of Rutland's Chess are switched above.
Too many windows open on my computer! My previous comment was supposed to be sent to the Piececlopedia entry for Chancellor.
<p>I am not aware of any computer programs implementing my King's Leap revision above. So I have written a Zillions file (kingleap.zrf) for King's Leap Chess and it is now being playtested. This uses my third and final version of the rules, allowing the Kings to leap either orthogonally or diagonally.
David Paulowich wrote on Sat, Aug 28, 2004 12:59 AM UTC:Excellent ★★★★★
Replying to Charles Gilman: Legler probably allows 'castling' using the Kings and the pieces in the corners of the board. So leaving the Rooks on the h-file means that they will still be available for castling after the Chancellors have left their home squares. When I independently came up with Paulowich's Chancellor Chess in 1997, I started off with that kind of castling rule. My variant just adds Chancellors (on the a-file), but also switches the Queens with the remaining Rooks. The revised version replaces castling with a King's Leap rule. Currently playtesting my Zillions file (kingleap.zrf) for this King's Leap Chess.
<p>Michael Howe cites Schmittberger as giving the Marshal and the Queen equal value in 'Grand Chess.' Wow! I consider the Chancellor (Marshal) to be at least a pawn lower than the Queen in Unicorn Chess and Unicorn Great Chess. Note: the Piececlopedia entry for Camel has recently turned into a mini-debate on piece values.
David Paulowich wrote on Sat, Aug 28, 2004 10:44 PM UTC:Excellent ★★★★★
1. P b2-b3
1... p d7-d5
2. R a1-p1; I-p1 // pocket nightrider
2... b c8-d7
3. I p1-b2 // is a PMChess Fool's Mate. Neat!
This variant has the potential to go beyond excellent. Bringing Class 4 up to four pieces yields 21 different pieces for the game, including the King. And 21, being the product of the magic numbers 3 and 7, makes a traditional choice for a complete set. I suggest adding the SuperAlibaba to Class 4, as its WFAD moves make a nice change from long range pieces.
David Paulowich wrote on Tue, Aug 31, 2004 10:14 PM UTC:Excellent ★★★★★
Repeating my 2001 comment. One web page for Cardinal Super Chess states: 'Because of the Cardinals' unique movement, a combination of a knight and a bishop, it gathers the initiative into one sweeping action.' This naturally leads to the mistaken conclusion that it is the usual B+N piece. But the second web page given for this commercial variant shows the move to be a non-leaping Camel. I tested the piece on the applet provided and saw the program move a Bishop to block my Cardinal check.
In his 12x12 variant GANYMEDE CHESS, Mark E Hedden calls the camel+bishop piece a 'Flying Dragon'. He considers it to be 'worth a bit less then a rook' and the griffon to be 'worth almost as much as a queen'. That surprised me, until I remembered that rooks and queens promote to stronger pieces in his game, which increases their value.
David Paulowich wrote on Wed, Sep 1, 2004 04:51 PM UTC:Excellent ★★★★★
Ralph Betza, in his article on More Halflings, writes: 'A Halfling Rose can move to all the squares a Rose can move to! The only difference is that it cannot go past the halfway mark and return to its starting square, and therefore it does not attack most squares as often. The normal Rose attacks every square it can reach at least twice -- once clockwise and once counterclockwise -- and also attacks half of the squares it can reach four times, where the circles intersect. The Halfling Rose still attacks twice ...' [when the half-circles intersect]
<p>Both pieces were intended for use on 12x12 and larger boards. I personally prefer a half-circle moving Rose to one capable of moving in a complete circle and ending its turn on the same square as it started from.
The paths d1-b2-a4-b6-d7 and d1-f2-g4-f6-d7 indicate that a Half-Rose on d1 can, so to speak, 'double-check' a King on d7. I will follow Duke in defining the Half-Rose to be a circular nightrider making from one to four leaps. On a 12x12 board this piece seems to be worth as much as a Rook, perhaps a little more than a Nightrider. [EDIT 2007] Rose Chess XII has these three pieces, with the Half-Rose definitely looking stronger than the other two.
Seven years ago I looked at the combination of Knight and (3,3) leaper, decided that it was too strange, and went on to invent the Quarter-Rose. Replacing the Queens in the standard game with this two-step leaper can lead to: 1. d1-c3 c7-c5, 2. c3-b5 mate! Black's foolish move left both c7 and d6 empty, which means that the mate cannot be blocked. Clearly this can be a dangerous piece on an 8x8 board. Looks like King, Knight, and Quarter-Rose can force mate against a lone King.
To be precise, your 'QuickPawns' move exactly like the Pawns on the 12x12 board of Perfect Chess, a non-competing entry in the Large Variant 99 Contest. Jean-Louis Cazaux states on his rules page:
<p>'It can advance one or two square from ANY position on the board. However, its capturing move is unchanged: one square diagonally forward. As a consequence, the en-passant capture is possible every time the opposite Pawn has advanced two squares. When the Pawn reaches the last row it can promote to one of the three major pieces: Queen, Lion or Gryphon. '
The Lion used by Antoine Fourrière, in Bilateral Chess and Jacks and Witches 84, is a weaker chess piece. But the Murray Lion itself is currently quite popular. Glenn Overby uses it in his Abecedarian Big Chess, Beastmaster Chess, and Decimal Chess - now called Meirìqí (Beautiful Sun Chess). Some other chess variants using this piece are:
Year *** Game *** (Piece Name) *** Game Inventor
1999 Centennial Chess (Murray Lion) John William Brown
1999 Millennial Chess (Murray Lion) John William Brown
Ralph Betza started off his first article with '... a Halfling Knight is exactly the same as a normal Knight.' So any statement about halfling piece values he made presumably refers only to halfling pieces which move to about half as many squares as the original piece.
<p>You almost need a lawyer to deal with the subject of halfling pieces. Just look at Eric Greenwood's Cavalier, a piece designed to have multiple paths to each square it attacks. I guess that the Halfling Cavalier can always make its nonleaping Knight moves. And these would be the only moves allowed from a central square on an 8x8 board.
David Paulowich wrote on Sun, Sep 5, 2004 02:19 AM UTC:Good ★★★★
'Three pieces can checkmate the Shaolin Master by themselves: the Praying Mantis, Ng Mui, and the promoted Bruce Lee.' Of course, there are many combinations of Shaolin Master and a single piece which can force mate against a lone Shaolin Master. The (unpromoted) Bruce Lee will do the job, as it is identical to the centuries-old Griffon (see the Griffon entry in the Piececlopedia). The companion piece, Wing Chun, does not occur in any chess variant that I know of. Eric Greenwood's 'Duke' is the only piece that even comes close.
This is David Paulowich writing in support of the standard rules for pawn promotion, which seem to be unpopular with some players and chess variant designers. If, for example, pawns could only be promoted to previously captured pieces, then many beautiful games would no longer be legal. My databases contain over 400 games with 4 Queens on the board, 2 White and 2 Black, including: Capablanca - Alekhine, 1927 (Thirteenth World Chess Championship Match, game 11) and Borsony - Koch, 1956 (Second World Correspondence Chess Championship). In 1936 Reinle checkmated Lange in this 'extra promotion' game: 1. e4 e5 2. f4 f5 3. exf5 e4 4. Qh5+ g6 5. fxg6 h6 6. g7+ Ke7 7. Qe5+ Kf7 8. gxh8=N#
Such games, with one player having nine pieces other than pawns, used to be rare (only ten were played between 1856 and 1963). In modern times the opening: 1. b4 e5 2. Bb2 Bxb4 3. f4 exf4 4. Bxg7 Qh4+ 5. g3 fxg3 6. Bg2 gxh2+ 7. Kf1 hxg1=Q+ 8. Kxg1 (from Kucharkowski - Walter, 1982) has been repeated in over 200 games. Incidentally, White is winning, by about 150 to 50.
Unicorn Great Chess: Inventor's Comments
I am pleased to see the PBM Preset for Unicorn Great Chess is still available on this website. I had been planning to post a rules page with a Zillions game file several years ago, but did not get around to it before my hard drive crashed. Anyway, here are some brief notes on this variant.
Unicorn Great Chess adds four 'Lions' to the board - on the second and ninth ranks. The Lion is identical to Ralph Betza's 'Half-Duck' (HFD, a piece that can move one square diagonally or jump two or three squares orthogonally), from the Remarkable Rookies army. It was designed to be at least as strong as the Rook on the 8x8 board, but it will be weaker on larger boards. I chose the name Lion as a reminder of the 3-square leaper in Grande Acedrex (1283). That game also has a piece called the Unicorn (or Rhinoceros). The remaining ten White pieces are arranged CRNBUKBNRQ on the first rank. Placing the Chancellors and Queens in the corners (with the Pawns in front of them also defended by the Knights) is an opening setup I have used on 64, 80, and 100 squares.
My somewhat rusty memory tells me that a King and Lion can force checkmate against a lone King. 'A Pawn promotes to a Chancellor, Queen, or Unicorn of the same colour upon reaching the player's tenth rank.' Perhaps I should amend that rule to also allow promotion to a Lion, if desired. Note that the Lion has the longest 'single leap' of any piece on this board, which makes it capable of delivering a Fool's Mate: 1. Lion f2-f5, Lion f9-f6 2. Pawn e3-e4, Lion f6-f4 (mate)
My suggested game values (allowing for Pawns being a little weaker on the 10x10 board) are as follows: Pawn 100, Knight 300, Bishop 350, Lion 450, Rook 550, Chancellor 900, Queen 1000, Unicorn 1000. [EDIT] I decided to bump the Lion up to 500, see the Rules section on my Unicorn Great Chess page.
I firmly believe that Q = R+B+P and Q+P = R+R on 64 or 100 squares. My (limited) playing experience on 64 squares leads me to conclude that a Chancellor is worth a Rook and Knight and half a Pawn. So I used that formula to calculate a 900 point value in this game. WARNING: the Lion and Unicorn values are mostly based on guesswork. It certainly would be convenient if the Chancellor was worth exactly two Lions (or a Rook and a Bishop). Other, more experienced, chess variant players may come up with different values. --- David Paulowich
First Position. White: King e5, Knight e1 Black: King e3, Rook a1 MOVES 1. Nc2 check, Kd3 2. Nxa1, Kc3 3. Ke4, Kb2 4. Kd3, Kxa1 is a simple 'two bare Kings draw' in FIDE chess. I wonder if centuries ago there were Shatranj tournament rules concerning bare King draws that require more than one move.
Second Position. White: King c1, Knight e1 Black: King a1, Pawn a2, Rook e2
MOVES 1. Nc2 check, Rxc2 check 2. Kxc2 stalemate(?) Applying the Bare King rule exactly as stated, White lost the game before he could capture the Rook and win by stalemate. This seems unfair. But if the rules did allow White to play his second move, should a stalemate by a bare King count as a win or only a draw?
Michael Howe raised a valid point for all variants that do not have a set of 'perfect promotion choices.' In FIDE chess, there is the common situation in King-and-Pawn endings where promotion to a Queen stalemates and promotion to a Rook does not - but a master will probably have resigned before this happens on the board. Also, sometimes you need to promote to a Knight and give check. For example:
H. Reinle - M. Lange (Murnau 1936)
1. e4 e5 2. f4 f5 3. exf5 e4 4. Qh5+ g6 5. fxg6 h6 6. g7+ Ke7 7. Qe5+ Kf7 8. gxh8=N# 1-0 is a checkmate underpromotion.
Another example: J. Levin - A. E. Santasiere (Pittsburgh 1946)
1. e4 e5 2. Nf3 Nc6 3. Bc4 Nf6 4. Ng5 d5 5. exd5 Na5 6. Bb5+ c6 7. dxc6 bxc6 8. Be2 h6 9. Nf3 e4 10. Ne5 Qc7 11. f4 Bc5 12. c3 Nb7 13. Qa4 O-O 14. Qxc6 Qxc6 15. Nxc6 Re8 16. b4 Bb6 17. Na3 Nd6 18. Nc2 Nd5 19. c4 Nxf4 20. c5 Nxe2 21. Kxe2 Bg4+ 22. Kf1 Bc7 23. cxd6 Bxd6 24. Na5 Bd7 25. Bb2 Ba4 26. Nb3 f5 27. g3 g5 28. Kg2 f4 29. gxf4 gxf4 30. Rhg1 Kf8 31. Kf2 Be7 32. Bg7+ Kf7 33. Bxh6 Bh4+ 34. Ke2 f3+ 35. Ke3 Rad8 36. Rg7+ Kf6 37. Rag1 Bxb3 38. axb3 Ke5 39. Bf4+ Kf6 40. Bg5+ Bxg5+ 41. R1xg5 f2 42. R5g6+ Kf5 43. Rg5+ Kf6 44. R5g6+ Ke5
45. Ra6 f1=N+ 46. Ke2 Rxd2+ 47. Kxf1 Rxc2 1/2-1/2 Promoting to a Knight (with check) on move 45 was the only way to avoid losing, according to my computer.
Definitely one of the weaker pieces. Dave McCooey writes (in his Endgame statistics with fantasy pieces) 'Two Camels, even on different colors, cannot checkmate a lone King. A Camel and a Zebra cannot checkmate a lone King. A Camel and a Wazir can checkmate a lone King, which is surprising. It can take as long as 77 moves (154 halfmoves) to force mate.'
Perhaps inventors would find the Camel-Dabbabah combination to be a more useful piece. Still limited to one color, but jumping two squares orthogonally gives the piece 'close-range' power.
This variant has the potential to go beyond excellent. Bringing Class 4 up to four pieces yields 21 different pieces for the game, including the King. And 21, being the product of the magic numbers 3 and 7, makes a traditional choice for a complete set. I suggest adding the SuperAlibaba to Class 4, as its WFAD moves make a nice change from long range pieces.
Repeating my 2001 comment. One web page for Cardinal Super Chess states: 'Because of the Cardinals' unique movement, a combination of a knight and a bishop, it gathers the initiative into one sweeping action.' This naturally leads to the mistaken conclusion that it is the usual B+N piece. But the second web page given for this commercial variant shows the move to be a non-leaping Camel. I tested the piece on the applet provided and saw the program move a Bishop to block my Cardinal check.
The paths d1-b2-a4-b6-d7 and d1-f2-g4-f6-d7 indicate that a Half-Rose on d1 can, so to speak, 'double-check' a King on d7. I will follow Duke in defining the Half-Rose to be a circular nightrider making from one to four leaps. On a 12x12 board this piece seems to be worth as much as a Rook, perhaps a little more than a Nightrider. [EDIT 2007] Rose Chess XII has these three pieces, with the Half-Rose definitely looking stronger than the other two.
Seven years ago I looked at the combination of Knight and (3,3) leaper, decided that it was too strange, and went on to invent the Quarter-Rose. Replacing the Queens in the standard game with this two-step leaper can lead to: 1. d1-c3 c7-c5, 2. c3-b5 mate! Black's foolish move left both c7 and d6 empty, which means that the mate cannot be blocked. Clearly this can be a dangerous piece on an 8x8 board. Looks like King, Knight, and Quarter-Rose can force mate against a lone King.
The Lion used by Antoine Fourrière, in Bilateral Chess and Jacks and Witches 84, is a weaker chess piece. But the Murray Lion itself is currently quite popular. Glenn Overby uses it in his Abecedarian Big Chess, Beastmaster Chess, and Decimal Chess - now called Meirìqí (Beautiful Sun Chess). Some other chess variants using this piece are:
Year *** Game *** (Piece Name) *** Game Inventor
1999 Centennial Chess (Murray Lion) John William Brown
1999 Millennial Chess (Murray Lion) John William Brown
1999 Jester Chess (Murray Lion) Thomas Havel
2003 Treeleaders Chess (Murray Lion) Erez Schatz
25 comments displayed
Earlier⇧Reverse Order⇩ Later⇧ Latest⇩Permalink to the exact comments currently displayed.