| Italic Title: | no |
| Fischer random chess | |
| Image Link: | Chess960 example init position.png |
| Image Caption: | One of 960 possible starting setups. Black's setup always mirrors White's. |
| Years: | Since June 19, 1996 |
| Players: | 2 |
| Setup Time: | ≈1 min + 1 min to determine starting position |
| Random Chance: | Pieces are randomized |
| Aka: | Chess960 |
Fischer random chess, also known as Chess960 ('chess nine-sixty'), is a variation of the game of chess invented by the former world chess champion Bobby Fischer.[1] Fischer announced this variation on June 19, 1996, in Buenos Aires, Argentina.[2] [3] [4] Fischer random chess employs the same board and pieces as classical chess, but the starting position of the pieces on the players' is randomized, following certain rules. The random setup makes gaining an advantage through the memorization of openings impracticable; players instead must rely more on their skill and creativity .
Randomizing the main pieces had long been known as shuffle chess, but Fischer random chess introduces new rules for the initial random setup, "preserving the dynamic nature of the game by retaining for each player and the right to castle for both sides".[5] The result is 960 unique possible starting positions.
In 2008, FIDE added Chess960 to an appendix of the Laws of Chess. The first world championship officially sanctioned by FIDE, the FIDE World Fischer Random Chess Championship 2019, brought additional prominence to the variant. It was won by Wesley So.[6] In 2022, Hikaru Nakamura became the new champion.[7]
Before the game, a starting position is randomly determined and set up, subject to certain requirements. White's pieces (not pawns) are placed randomly on the first, following two rules:
Black's pieces are placed equal-and-opposite to White's pieces. (For example, if the white king is randomly determined to start on f1, then the black king is placed on f8.) Pawns are placed on the players' second ranks as in classical chess.
After setup, the game is played the same as classical chess in all respects, with the exception of castling from the different possible starting positions for king and rooks.
There are 4 × 4 × 6 × 10 × 1 = 4 × 4 × 15 × 4 × 1 = 960legal starting positions:
Usually, the players accept the conditions of the organizer to generate the starting position with software. If the software is not available or the players do not accept it, there are many ways to generate a random starting position with equal probability. This requires choosing a number randomly between 0 and 959 inclusive; this number is then used as an index to the Fischer random chess numbering scheme.
In 1998, Ingo Althöfer proposed a method that requires only a single standard die.[8] [9] (Re-roll if needed to get values in the range 1–4 or 1–5).
960 choices can be obtained in several ways by combining polyhedral dice without re-rolling; for example 4×12×20 or 6×8×20 or 8×10×12.
Shuffling marked objects (cards, pieces, pawns, dominoes, Scrabble letters) and use the permutations. For example, shuffle 14 marked cards a,b,c,d,e,f,g,h, N,N,Q,R,R,R and place them, in this random order, separated into three rows:
The first card of each row determines one bishop's place.
The variant has held a number of different names. It was originally known as "Fischerandom" or "Fischerandom chess", the name given by Fischer. "Chess960" is used by FIDE in its Laws of Chess.[10]
Hans-Walter Schmitt, chairman of the Frankfurt Chess Tigers e.V. and an advocate of the variant, started a brainstormingprocess for creating a new name, which had to meet the requirements of leading grandmasters; specifically, the new name and its parts:
The effort culminated in the name choice "Chess960" – derived from the number of different possible starting positions. Fischer never publicly expressed an opinion on the name "Chess960".
Reinhard Scharnagl, another proponent of the variant, advocated the term "FullChess". Today he uses FullChess, however, to refer to variants which consistently embed classical chess (e.g. Chess960 and similar variants). He recommends the name Chess960 for the variant in preference to Fischer random chess.
Starting 2019, whenever the Saint Louis Chess Club hosts Fischer Random chess tournaments, their tournaments are called Chess 9LX,[11] [12] where '9LX' is a combination of the Arabic numeral 9 and the Roman numeral LX (60), a possible reference to how the number '960' is often read as 'nine-sixty' instead of 'nine hundred sixty' when talking about 'chess960'.
As in classical chess, each player may castle once per game, moving both the king and a rook in a single move; however, the castling rules were reinterpreted in Fischer random chess to support the different possible initial positions of king and rook. After castling, the final positions of king and rook are the same as in classical chess, namely:
| + Examples of castling |
Castling prerequisites are the same as in classical chess, namely:
FIDE's recommended procedure for castling unambiguously is first to move the king outside the playing area next to its final square, then to move the rook to its final square, then to move the king to its final square. Another recommendation is to verbally announce the intent to castle before doing so.[15]
The study of openings in Fischer random chess is in its infancy,[21] [22] [23] [24] but fundamental opening principles still apply, including: protect the king, control the central squares (directly or indirectly), and develop rapidly, starting with the less valuable pieces.
Unprotected pawns may also need to be dealt with quickly. Many starting positions have unprotected pawns, and some starting positions have up to two that can be attacked on the first move. For example, in some Fischer random chess starting positions (see diagram), White can attack an unprotected black pawn on the first move, whereas in classical chess it takes two moves for White to attack, and there are no unprotected pawns.
It has been argued that two games should be played from each starting position, with players alternating colors, since the advantage offered to White in some initial positions may be greater than in classical chess.[25]
However, Sesse[18] [19] (which used Stockfish 9) evaluated the starting positions to be between 0.00 and 0.57, with an average of 0.18 advantage for White. BBNNRKRQ (SP 80) was the least balanced position with 0.57 advantage for White, while 27 different starting positions evaluate as equal, or 0.00 advantage for White or Black, ex. BBRNNQKR (SP 432). The standard chess starting position RNBQKBNR (SP 518) was evaluated at 0.22 advantage for White. Hence, on average, a Fischer Random starting position is actually 18.2% more balanced than the standard starting position. Furthermore, 923 starting positions lie within two standard deviations of the 0.18 mean advantage for White, i.e. between 0 and 0.371, with a standard deviation of 0.0955. This demonstrates that the advantage for White, at least for a chess engine, is on average lower for Fischer random compared to the standard position, and at worst only marginally higher. However, since humans can't analyze to the depth of a chess engine, it is possible that the practical advantage for White or Black could skew in either direction from one position to the next, depending on the player skill level.
See also: Displacement chess.
The concept of random chess, also known as shuffle chess, was first proposed by the Dutch chess enthusiast Philip Julius van Zuylen van Nijevelt (1743-1826). In his renowned treatise on chess, La Supériorité aux Échecs, published in 1792, Van Zuylen van Nijevelt articulated his disdain for the repetitive patterns often found in standard chess openings. He proposed the idea of randomizing the starting positions of the main pieces to create a vast array of unique situations, eliminating the possibility of pre-game memorization or extensive opening theory. Van Zuylen van Nijevelt's book, with its full title La Supériorité aux Échecs mise à la portée de tout le monde, et particulièrement des dames qui aiment cet amusement ("Superiority in Chess brought into the reach of all, and particularly of ladies who love that amusement"), gained significant popularity and was reprinted several times. Its influence extended beyond the Dutch-speaking world, as it was subsequently translated into multiple languages, spreading the idea of randomizing the initial positions of chess pieces to a wider audience. This early conception of random chess by Van Zuylen van Nijevelt laid the foundation for what later became known as Fischer random chess, a variant that gained prominence in the 20th century.
Van Zuylen van Nijevelt's innovative approach to chess not only offered a solution to the repetitive nature of traditional openings but also paved the way for the exploration of chess variants that deviate from the conventional starting position. His quote within La Supériorité aux Échecs emphasizes the core principle behind random chess, stating, "This produces a huge number of different situations, so that no one can study them beforehand," reflecting his desire to introduce an element of unpredictability and originality into the game of chess. The legacy of Van Zuylen van Nijevelt's contributions to the evolution of chess remains significant, with his early insights serving as a cornerstone for the development of various randomized chess variants, including Fischer random chess and its contemporary iterations.[26] [27] [28]
The pioneering work of Van Zuylen van Nijevelt found continued development through the efforts of his nephew, the Jonkheer Elias van der Hoeven (1778-1854), a Dutch diplomat. Van der Hoeven took the concept of shuffle chess further, potentially sharing his insights with Aaron Alexandre, evident from Alexandre's incorporation of the theory into his Encyclopédie des échecs in 1837.
The earliest documented games of shuffle chess were played between Van der Hoeven and Alexandre in Mannheim in 1842, with Alexandre emerging as the victor with a score of 3-0. One of these games is preserved in Sissa, the Netherlands' first long-standing chess journal, demonstrating an initial position with two advanced pawns on each side. A later game played by Van der Hoeven was against Baron von der Lasa (1818-1899), adhering more closely to the contemporary rules of random chess, except for the monochromatic bishop pairs.
In 1851, Van der Hoeven visited Willem Verbeek, the editor-in-chief of Sissa. Verbeek and Hancock, Verbeek's chess companion in Amsterdam during the 1850s, delved into shuffle chess, with their initial findings documented in the pages of the Sissa journal.
Van der Hoeven's modifications to the original concept of random chess were published in Alexandre's Encyclopédie in 1837 and later republished in Sissa by an individual known as T. Scheidius. This variant began to be referred to as "schaakspel, naar de wijze van jhr. Van der Hoeven" or "schaakspel à la Van der Hoeven."
Following Van der Hoeven's visit, the Sissa Chess Society received an invitation from the Philidor Chess Society in Amsterdam, in collaboration with Van der Hoeven, to organize a shuffle chess tournament. The aim was to promote the dissemination and popularity of the chess variant attributed to Van der Hoeven. Originally intended as a tournament among the eight prize winners of the 1851 Philidor-organized event, logistical challenges led to invitations being extended to other chess societies. Ultimately, a tournament with seven players from Amsterdam, along with the 74-year-old Van der Hoeven, was organized. Notable participants included Maarten van 't Kruijs (Philidor), J. Seligmann (Philidor), H. Kloos (La Bourdonnais), M.M. Coopman, and F.G. Hijmans / S. Heijmans, supplemented by Mohr and J. van Praag. Van ’t Kruijs emerged as the winner of the tournament, reinforcing the growing sentiment that the removal of opening theory allows true chess talent to shine.
Fischer's modification "imposes certain restrictions, arguably an improvement on the anarchy of the fully randomized game in which one player is almost certain to start at an advantage".[29] Fischer started to develop his new version of chess after the 1992 return match with Boris Spassky. The result was the formulation of the rules of Fischer random chess in September 1993, introduced formally to the public on June 19, 1996, in Buenos Aires, Argentina.[30]
Fischer's goal was to eliminate what he considered the complete dominance of openings preparation in classical chess, replacing it with creativity and talent. In a situation where the starting position was random it would be impossible to fix every move of the game. Since the "opening book" for 960 possible opening systems would be too difficult to devote to memory, the players must create every move originally. From the first move, both players must devise original strategies and cannot use well-established patterns.[31] [32] Fischer believed that eliminating memorized book moves would level the playing field.
During summer 1993, Bobby Fischer visited László Polgár and his family in Hungary. All of the Polgar sisters (Judit Polgár, Susan Polgar, and Sofia Polgar) played many games of Fischer random chess with Fischer. At one point Sofia beat Fischer three games in a row. Fischer was not pleased when the father, László, showed Fischer an old chess book that described what appeared to be a forerunner of Fischer random chess. The book was written by Izidor Gross and published in 1910. Fischer then changed the rules of his variation in order to make it different.[33] [34] There are games of shuffle chess recorded as early as 1852 but Fischer is generally credited with fixing the colors of bishops alongside king placement between the rooks and defining the castling process.[35] In a later radio interview, Fischer explained his reasoning for proposing a revision of shuffle chess, rather than a game with new pieces (and a larger board), as the “new chess“:
See also: World Chess960 Championship.
Note: None of the Mainz championships were recognized by FIDE. Furthermore, they were all played with rapid time controls.
| Year | Championship | Open | Women's Championship | Computer Championship | |
|---|---|---|---|---|---|
| 2001 | Péter Lékó (4½–3½ vs. Michael Adams) | ||||
| 2002 | |||||
| 2003 | Peter Svidler (4½–3½ vs. Péter Lékó) | ||||
| 2004 | Peter Svidler (4½–3½ vs. Levon Aronian) | ||||
| 2005 | Peter Svidler (5–3 vs. Zoltán Almási) | Spike | |||
| 2006 | Levon Aronian (5–3 vs. Peter Svidler) | Alexandra Kosteniuk (5½–2½ vs. Elisabeth Pähtz) | Shredder | ||
| 2007 | Levon Aronian (2–2, 1½–½ vs. Viswanathan Anand) | Rybka | |||
| 2008 | Alexandra Kosteniuk (2½–1½ vs. Kateryna Lahno) | Rybka | |||
| 2009 | Hikaru Nakamura (3½–½ vs. Levon Aronian) | Rybka |
In 2005, chess program The Baron[55] played two Fischer random chess games against Chess960 World Champion Peter Svidler, who won 1½–½. The chess program Shredder, developed by Stefan Meyer-Kahlen of Germany, played two games against Zoltán Almási from Hungary, where Shredder won 2–0.
TCEC has held TCEC FRC since 2019 where Stockfish has won every edition except the 2021 edition which was won by Komodo.[56]
From February 9 to 13, 2018, a Fischer random chess match between former classical World Chess Champion Magnus Carlsen and the unofficial Fischer random chess world champion Hikaru Nakamura was held in Høvikodden, Norway. The match consisted of 8 rapid and 8 blitz games, with the rapid games counting double. Each position was used in two games, with colors reversed. Carlsen prevailed with a score of 14–10.[57] [58]
2018 – From September 11 to 14, 2018, the Saint Louis Chess Club held a Fischer random chess event,[59] but they did not yet call their event 'Chess 9LX'. (They started next year.) The playing format consisted of individual matches, each pair of players facing the same five different starting positions, with 6 rapid games (counting 2 points each) and 14 blitz games (counting 1 point each). Players and scores:[60]
2019 – The playing format once again consisted of individual matches. Players and scores:[61]
2020 – The playing format changed to a round robin. The event was won by both (There was no tiebreaker) former world (standard) chess champion Magnus Carlsen and Hikaru Nakamura. The reigning FIDE world Fischer random chess champion Wesley So placed fifth out of the ten players. So lost only once, namely to Alireza Firouzja.[62]
2021 – The playing format was once again a round robin. The event was won by Leinier Domínguez Pérez. The reigning FIDE world Fischer random chess champion Wesley So placed second out of the ten players, tied with Sam Shankland and Maxime Vachier-Lagrave. So lost only twice, namely to Leinier Domínguez Pérez and Sam Shankland.[63]
2022 – The playing format was once again a round robin. The event was won by Fabiano Caruana who defeated Alireza Firouzja in armageddon. Firouzja had previously placed tenth out of ten (last place) in the 2020 tournament. The reigning FIDE world Fischer random chess champion Wesley So placed fifth out of the ten players. So lost only thrice, namely to Shakhriyar Mamedyarov, Alireza Firouzja (again) and Caruana.[64]
See main article: FIDE World Fischer Random Chess Championship 2019.
See main article: FIDE World Fischer Random Chess Championship 2022.
On April 20, 2019, the first world championship in Fischer random chess officially recognized by FIDE was announced. It ended on November 2, 2019. In the finals, Wesley So defeated the former and four-time world chess champion Magnus Carlsen 13½–2½ (4 wins, 0 losses, 2 draws) to become the inaugural world Fischer random chess champion.
In the announcement, FIDE president Arkady Dvorkovich commented:[65]
It is an unprecedented move that the International Chess Federation recognizes a new variety of chess, so this was a decision that required to be carefully thought out. But we believe that Fischer Random is a positive innovation: It injects new energies and enthusiasm into our game, but at the same time it doesn't mean a rupture with our classical chess and its tradition. It is probably for this reason that Fischer Random chess has won the favor of the chess community, including the top players and the world champion himself. FIDE couldn't be oblivious to that: It was time to embrace and incorporate this modality of chess.
On August 19, 2022, the second world championship was announced for later in 2022, in Iceland. This is exactly half a century after the World Chess Championship 1972 held in Iceland between Fischer and Boris Spassky. On October 30, Hikaru Nakamura played the finals against Ian Nepomniachtchi, who had earlier knocked out Magnus Carlsen. Nakamura won in the armageddon after drawing the match 2–2.[66]
See main article: Freestyle Chess G.O.A.T. Challenge. The eight-player Freestyle Chess G.O.A.T. Challenge was the first major Fischer Random Chess tournament that used classical time controls. It took place in Germany from February 9–16, 2024. Fischer Random world champion Nakamura did not play in the event. Magnus Carlsen won the tournament by defeating Fabiano Caruana in the finals.
See main article: Fischer random chess numbering scheme. Recorded games must convey the Fischer random chess starting position. Games recorded using the Portable Game Notation (PGN) can record the initial position using Forsyth–Edwards Notation (FEN), as the value of the "FEN" tag. Castling is notated the same as in classical chess (except PGN requires letter O, not number 0). Note that not all chess programs can handle castling correctly in Fischer random chess games. To correctly record a Fischer random chess game in PGN, an additional "Variant" tag (not "Variation" tag, which has a different meaning) must be used to identify the rules; the rule named "Fischerandom" is accepted by many chess programs as identifying Fischer random chess, though "Chess960" should be accepted as well. This means that in a PGN-recorded game, one of the PGN tags (after the initial seven tags) would look like this: [Variant "Fischerandom"].
FEN is capable of expressing all possible starting positions of Fischer random chess; however, unmodified FEN cannot express all possible positions of a Chess960 game. In a game, a rook may move into the back row on the same side of the king as the other rook, or pawn(s) may be underpromoted into rook(s) and moved into the back row. If a rook is unmoved and can still castle, yet there is more than one rook on that side, FEN notation as traditionally interpreted is ambiguous. This is because FEN records that castling is possible on that side, but not which rook is still allowed to castle.
A modification of FEN, X-FEN, has been devised by Reinhard Scharnagl to remove this ambiguity. In X-FEN, the castling markings "KQkq" have their expected meanings: "Q" and "q" mean a-side castling is still legal (for White and Black respectively), and "K" and "k" mean h-side castling is still legal (for White and Black respectively). However, if there is more than one rook on the baseline on the same side of the king, and the rook that can castle is not the outermost rook on that side, then the file letter (uppercase for White) of the rook that can castle is used instead of "K", "k", "Q", or "q"; in X-FEN notation, castling potentials belong to the outermost rooks by default. The maximum length of the castling value is still four characters. X-FEN is upwardly compatible with FEN, that is, a program supporting X-FEN will automatically use the normal FEN codes for a traditional chess starting position without requiring any special programming. As a benefit, all 18 pseudo FRC positions (positions with traditional placements of rooks and king) still remain uniquely encoded.
The solution implemented by chess engines like Shredder and Fritz is to use the letters of the columns on which the rooks began the game. This scheme is sometimes called Shredder-FEN. For the traditional setup, Shredder-FEN would use HAha instead of KQkq.
There are several variants based on randomization of the initial setup. "Randomized Chess, in one or other of its many reincarnations, continues to attract support even, or perhaps especially, that of top players."[67]
| Variant | Under condition (and all of the above) | Positions with symmetry a1 to h1 = a8 to h8 | Positions with symmetry a=h, b=g, c=f, d=e (king=queen) | |
|---|---|---|---|---|
| Shuffle chess | Random | 5040 | 48 | |
| Chess2880 | Bishops opposite color | 2880 | 48 | |
| Fischer random chess (FRC, Chess960) | King between rooks | 960 | 16 | |
| Chess18 | King and rooks fixed on traditional starting squares | 18 | 2 |
Remarks
Any variant with N starting positions can exist with mirroring (or rotating) white and black otherwise it means another (double) variant with N2 starting positions. In any variant the castling is not possible in any case or is possible only when king and rook are on their traditional starting squares, or as follows:
After castling with the nearest rook to the column:
The double chess2880 without castling is known as Transcendental chess (or TC).
Chess18 is the subset of Chess960 in which the kings and rooks are fixed, so that castling is always orthodox, preserving more of the feel of ordinary chess. This allows some opening knowledge to still have practical relevance – one could practically learn a few short lines for each of the eighteen starting positions – though it does not reach the extent of orthodox chess, in which one has to memorize many long computer lines. Moreover, this excludes all starting positions where White has a large advantage, and makes early blunders less common by making it impossible for a bishop to attack an undefended pawn after White's first move.[68] [69] [70] [71] However, due to having fewer starting positions and the requirement that the king and rooks must start on traditional starting squares, Chess18 has much less positional diversity than is offered by Chess960. Additionally, while memorizing openings is more difficult in Chess18 than in traditional chess, opening preparation has the potential to be a much more significant aspect of the game than in Chess960, especially for grandmasters and top players. Chess18 positions are more likely than Chess960 positions to resemble structures found in traditional chess due to the limited starting positions in Chess18 in comparison to the diversity of starting positions found in Chess960.
As discussed above in 'Castling rules', Chess870 and Chess90 are the partitioning subsets of Chess960 in which a player, respectively, never needs or may need to give up castling rights on one side to castle on the other side.[17] [20]
Chess480
In "Castling in Chess960: An appeal for simplicity", John Kipling Lewis proposes alternative castling rules which Lewis has named "Orthodoxed Castling".[72]
The preconditions for castling are the same as in Chess960, but when castling,
... the king is transferred from its original square two squares towards (or over) the rook, then that rook is transferred to the square the king has just crossed (if it is not already there). If the king and rook are adjacent in a corner and the king cannot move two spaces over the rook, then the king and rook exchange squares.
| + Examples of Chess480 castling |
Unlike Fischer random chess, the final position after castling in Chess480 will usually not be the same as the final position of a castling move in traditional chess. Lewis argues that this alternative better conforms to how the castling move was historically developed.
Lewis has named this chess variation "Chess480"; it follows the rules of Chess960 with the exception of the castling rules. Although a Chess480 game can start with any of 960 starting positions, the castling rules are symmetrical (whereas the Chess960 castling rules are not), so that mirror-image positions have identical strategies; thus there are only 480 effectively different positions. The number of starting positions could be reduced to 480 without losing any possibilities, for example by requiring the white king to start on a light (or dark) square.
There are other claims to the nomenclature "Chess480"; Reinhard Scharnagl defines it as the white queen is always to the left of the white king.
David O'Shaughnessy argues in "Castling in Chess480: An appeal for sanity"[73] that the Chess480 rules are often not useful from a gameplay perspective. In about 66% of starting positions, players have the options of castling deeper into the wing the king started on, or castling into the center of the board (when the king starts on the b-, c-, f-, or g-files). From Wikipedia article Castling: "Castling is an important goal in the early part of a game, because it serves two valuable purposes: it moves the king into a safer position away from the center of the board, and it moves the rook to a more active position in the center of the board." An example of poor castling options is a position where the kings start on g1 and g8 respectively. There will be no possibility of "opposite-side castling" where each player's pawns are free to be used in, as the kings' scope for movement is very restricted (it can only move to the h- or e-file). These "problem positions" play well with Chess960 castling rules.
Bibliography