Losing Chess Information

Topic Editor: Guy Haworth

Contributions are welcome and the ICGA will, with permission, list contributor's names and email addresses on its Contributor's page.


Losing Chess (LC) is also known as Giveaway Chess, Killer Chess, Loser's Chess, Qui Perde Gagne, Suicide Chess, Take-all Chess and Vinciperdi. The alternative names make a better distinction between Losing Chess and losing chess, a common experience in Western Chess.

Losing Chess is a popular variant of Western Chess, and there is more to it than at first appears. Endgames are compulsory and more varied, complex and subtle. See www.pion.ch/Losing/index.html for advice on openings, middlegame, endgame, tactics and some games.

Since compulsory capture limits options significantly, this is a sharply tactical game in which computers excel. The top 13 or so places in the FICS league are taken by computers, headed by Ben Nye's ascp.

Rules of the Game

The varieties of LC are numbered here. The variants revolve around stalemate, the King, castling, and P-conversion. LC1 is the 'International' ruleset; LC2 is the FICS ruleset.

  • LC1: The objective is not to have a move, i.e. to have no pieces or be stalemated. Capturing is compulsory but you have the choice of capture if there is one. The King is an ordinary piece to which a Pawn can promote but Western Chess concepts of castling, check and mate do not exist.

  • LC2: FICS version: as LC1 but stalemate means that the side with least pieces wins.

  • LC3: As LC2 except that Pawns promote only to Queens, and stalemate is a draw.

  • LC4: As LC2 but, on stalemate, the player 'passes' and the opponent plays on until the stalemate is relieved.


" known to be only value-preserving move; ' unique optimal in some metric; ° only legal move; ?? loses


  • LC1 variant, "a well fought draw by correspondence." (Pritchard, 2000; p. 37) E.T.O. Slater v H. Klüver (1954?): 1. e3 b5 2. Bxb5° c6 3. Bxc6° Nxc6 4. b4 Nxb4° 5. a3 Nxc2° 6. Qxc2° g6 7. Qxc8 Qxc8 8. Nf3 Qxc1° 9. Ng5 Qxb1 10. Rxb1 Rb8 11. Rxb8 Kd8 12. Nxh7 Rxh7° 13. Rxd8° Rxh2° 14. Rxf8 Rxh1° 15. Rxg8 Rxe1° 16. Rxg6 Rxe3" (not 16. fxg6?? 17. e4!) 17. dxe3 fxg6° 18. e4 g5 19. e5 e6 20. f4 gxf4° 21. g3 fxg3° 22. a4 g2 23. a5 g1=Q 24. a6 d6 25. exd6° Qb1 26. d7 Qg6 27. d8=K" (alternatives lose by stalemate) 27. ... Qb1 28. Kc7 Qb8 29. Kxb8° (KP-PP =) e5° 30. Kxa7° e4° 31. Kb7 e3° 32. a7 e2° 33. a8=R e1=K" ½-½.

  • F.Sandström-F.Liardet (1st Losing Chess World Ch., Utrecht, 2001, www.pion.ch/Losing/wcgame1.html for notes. This game won the tournament Endgame Prize): 1. g3 g6 2. b4 b6 3. Nh3 a5 4. bxa5° Rxa5 5. d3 Rxa2° 6. Rxa2° Bh6 7. Bxh6° Nxh6° 8. Ra6 Bxa6 9. Qc1 Bxd3° 10. Qxh6 Bxe2 11. Qxh7 Rxh7 12. Bxe2 Rxh3° 13. Bf3 Rxh2 14 Rxh2° c5 15. Bc6 Nxc6 16. Rh8 c4 17. Rxe8° Qxe8° 18. Kd2 Qh8 19. Kd3 cxd3° 20. cxd3° Qc3 21. Nxc3° d5 22. Nxd5° e5 23. Nxb6° e4 24. dxe4° f5 25. exf5° gxf5° 26. g4 fxg4° 27. Na4 Nb4 28. f3 gxf3° 29. Nc3 Na2 30. Nxa2° f2° 31. Nc3' f1=N" 32. Resigns {32. Na2 Ne3' 33. Nc1' Nc2' 34. Na2 Nb4' 35. Nxb4} 0-1.

  • d3-loss: (www.pion.ch/Losing/d3.html for annotation) 1. d3?? g5 2. Bxg5° Bg7 3. Bxe7° Bxb2 4. Bxd8° Bxa1 5. Bxc7° Bc3 6. Nxc3 d5 7. Nxd5 Nf6 8. Nxf6 Rg8 9. Nxe8 Rxg2 10. Bxg2 f6 and now:
    1. 11. Nxf6 Nc6 12. Nxh7 b5 13. Bxc6° Bb7 14. Bxb7° Rd8 15. Bxd8° a5 16. Bxa5° b4 17. Bxb4° 0-1.
    2. 11. Bxb8 Rxb8° 12. Nxf6° b6 13. Nxh7° Ra8 14. Bxa8° Bb7 15. Bxb7° a6 16. Bxa6° b5 17. bxb5° 0-1.
    3. 11. Bxb7 Bxb7° 12. Nxf6 Bxh1° 13. Nxh7 Be4 14. dxe4 (14. Bxb8 Rxb8° 15. dxe4° Rb8) Nd7 16. Qxd7° Rb8.

  • d4-loss: (www.pion.ch/Losing/d4.html for annotation) 1. d4?? e5 2. dxe5° Qg5 3. Bxg5 (3. Qxd7 Kxd7° 4. Bxg5° is similar) Kd8 4. Bxd8° 5. a6 Bxc7° Ra7 6. Bxb8° b6 7. Bxa7° a5 8. Bxb6° h5 9. Bxa5° Bb4 10. Bxb4° Ne7 11. Bxe7° Rd8 12. Bxd8° g5 13. Bxg5 h4 14. Bxh4° d6 15. exd6° f6 16. Bxf6° Bh3 17. gxh3° 0-1.

  • e4-loss: (www.pion.ch/Losing/e4.html for annotation) 1. e4?? b5 2. Bxb5° Nf6 3. Bxd7° Nxe4 4. Bxe8 Nxf2 5. Kxf2 Qxd2 6. Bxf7 Qxc1 7. Qxc1° Rg8 8. Bxg8° g6 9. Bxh7° e5 10. Bxg6° e4 11. Bxe4° Nc6 12. Bxc6° a5 13. Bxa8° c6 14. Bxc6° a4 15. Bxa4° Bd7 16. Bxd7° Ba3 17. Nxa3 0-1.

  • Top C-C Game: ascp (Elo 3098) v RoboAmn (2960), FICS, 2002/10/04 (300"+10"/m), notes by angrim: 1. e3 c5 2. Bd3 h5 3. Qxh5° Rxh5° 4. b4 cxb4 5. Ba6 Nxa6 6. d3 Rxh2° 7. Rxh2° b3 {taking ascp out of book} 8. cxb3 Nb8 9. Bd2 a5 10. Bxa5° Rxa5° 11. Na3 Rxa3° 12. Rh6 gxh6 13. Rd1 Rxb3 14. axb3° b5 15. e4 h5?? {seen later to be the losing move: nevertheless, both computers now play optimally} 16. Nh3 h4 17. Kd2 Ba6 18. Rh1 f5 19. exf5° d6 {evaluations had been slightly in favour of ascp, which only here saw the forced win} 20. f6 Nxf6 21. f3 Ne4 22. dxe4 d5 23. exd5° Qxd5° 24. f4 Qxb3 25. Ke3 Qxe3° 26. Ng5 Qxf4° 27. Rxh4° Qxh4 28. Nf7 Kxf7° 29. g4 Qxg4° 1-0.

Didactic Positions

Curiosities and Studies

  • (Beasley, 1997a) after Gyorgy Evseev. wNc2g1/bPd3, win by 1. Ne2 and now 1. ... dxc2 2. Nc3 (not 2. Nd4 c1=B 3. Nc6 Bh6) c1=B 3. Na2 or 1 ... dxe2 2. Nd4 (not 2. Ne2 e1=B 3. Ng2 Bh4) e1=B 3. Ne6.

  • (Beasley, 2001a/b). The only B-QN draw ... Bd1 qd6 nh8 w. 1. Ba4" Qf4" ½-½ ... repeating position 1w.

  • Deepest 4-man endgames: 87 moves to finish,
        a) BNP-K, Bb2 Nc1 Pf3 kg1 w and BNP-P, Bd2 Ng1 Pf3 pg3 w
    Deepest 5-man endgame (Beasley, 2002c): 78 moves to Conversion, KRP-KP, Ke8 Rf2 Pc2 ke4 pc6 w.

Past Achievements

Opening Theory

Pritchard (2000) says that 1. Na3/Nh3/b3/e3(favourite) seem safe.

LC2 opening theory has been developed by the search-method PN2 (Breuker, 1998) as implemented by syzygy in Sjaak: c.f. lenthep.dyndns.org:8080/suicide/proofs/

1. Nc3/Nf3/b4/d3/d4/e4/f4/h4??; 1. b3 e6 2.a3/c3??
1. e3 Na6/Nf6/a5/a6/e5/f5/f6/g6/h5/h6??; 1. g3 Nc6??

Endgame tables

1992: First LC endgame computed ... N-NN (Evseev and Poisson, 1993; Beasley, 2001a/b).
1997: All pawnless 3-man endgames computed (Beasley, 1999b, 2001a/b).
1998: All 3-man endgames computed by Laurent Bartholdi (Beasley, 2001a/b).
1999: R-KKK computed by Laurent Bartholdi (Beasley, 2001a/b).
1999: All 4-man DTF(inish) endgames computed by Angrim Ben Nye (Beasley, 2001a/b).
2002: All 5-man DTC(onversion) endgames computed by Angrim Ben Nye (Beasley, 2002c).

Let the number of ordered n-sequences of k characters ° S(n, k).
Then S(1, k) = k and S(n, k) ° Si = 1..k (n-1, i).
Then, for mi < 9:
    the number of mi-mj LC endgames with mi > mj is S(ni, k) x S(nj, k) where ni = mi and nj = mj
    the number of mi-mi LC endgames is S(ni, k) x [S(ni,k)+1]/2 where ni = mi
For LC, k = 6. For Western Chess with a single mandatory King per side, k = 5 and ni = mi-1.


Future Challenges

  1. Retrieve, for the web, Stan Goldovski's pages on Losing Chess.

  2. Published opening proofs with the full use of 5-man EGTs for variant LC2.

  3. A publicly-available opening book.

  4. 3-5-man databases for variant LC1.

  5. 6-man databases for variant LC2.

  6. Getting the editor to understand (Proof Number) pn-search ... and pn2-search.

Relevant Associations

Game Associations

Services and Sources of Information

Game-playing Servers

Key People and Programs

  • Laurent Bartholdi: computed endgame databases.

  • John Beasley johnbeasley@mail.com main source for much here.

  • Johan Bosman: author of Pindakaas.

  • Gyorgy Evseev: endgame calculations.

  • Stan Goldovski: sadly, deceased. Copy of his webpages on LC being sought.

  • Fabrice Liardet: www.pion.ch lots of advice on LC, nabla@pion.ch.

  • Ben Nye (angrim): author of top program ascp and complete 5-man (LC2) endgame databases.

  • Gian-Carlo Pascutto: Sjeng,gcp@sjeng.org

  • Tim Remmel: won the 1st Losing Chess World Championship.

  • syzygy: author of Sjaak.

  • Lenny Taelman: lenthep.dyndns.org:8080



Journals and Magazines

...not exclusively focused on this game.

Books and Papers

  • Beasley, J.D. (1996). Elementary Duels in the Losing Game. BESN, Special No. 4, pp. 2-3.

  • Beasley, J.D. (1997a). NN-P Study F1674. The Problemist, Vol. 16, No. 1, p21.

  • Beasley, J.D. (1997b). Promotion Studies in the Losing Game. BESN, Special No. 8, pp. 2-3.

  • Beasley, J.D. (1998). Computer Discoveries in the Losing Game. BESN, Special No. 13, pp 2-3.

  • Beasley, J.D. (1999a). Paradoxical Play in the Losing Game. BESN, Special No. 18, pp. 2-3.

  • Beasley, J.D. (1999b). Three-man pawnless endings in Losing Chess. Privately published research pamphlet.

  • Beasley, J.D. (2000a). A first survey of Losing Chess endgame material published up to the end of 1999. Privately published research pamphlet.

  • Beasley, J.D. (2000b). Two Classic Games of Losing Chess. Variant Chess, Vol. 5, No. 35, pp. 34-35.

  • Beasley, J.D. (2001a). Further Discoveries in the Losing Game. BESN, Special No. 24, pp. 2-3.

  • Beasley, J.D. (2001b). Computer Discoveries in Losing Chess. ICGA Journal, Vol. 24, No. 2, pp. 102-103.

  • Beasley J.D. (2002a). The First Unofficial Losing Chess World Championship. Variant Chess. Vol. 5, No. 39. pp. 106-107.

  • Beasley J.D. (2002b). The First Unofficial Losing Chess World Championship (cont'd). Variant Chess. Vol. 5, No. 40. pp. 124-125.

  • Beasley, J.D. (2002c). Losing Chess: Burning the Candle at Both Ends. Variant Chess, to appear.

  • Binnewirtz, R.J. (2000). Schlagabtausch im Räuberschach. Mädler, Dresden. ISBN 3-9256-9124-3.

  • Breuker, D.M. (1998). Memory Versus Search in Games. Ph.D. thesis, University of Maastricht. ISBN 9-0901-2006-8.

  • Evseev, G. and Poisson, C. (1993). Finales de cavaliers en "qui perd gagne". Rex Multiplex, No 41 (April), pp. 2048-2049. Paris.

  • Goldovski, S. (1999). Problem F1844. The Problemist, Vol. 17 No. 2 (March), p. 51. ISSN 0032-9398.

  • Liardet, F. (2002). www.pion.ch/chess.html La page d’échecs de Fabrice Liardet.

  • Pritchard, D.B. (1994). The Encyclopedia of Chess Variants. Everyman Chess, esp. pp. 176-9. ISBN: 0-9524-1420-1.

  • Pritchard, D.B. (2000). Popular Chess Variants. B.T. Batsford, London. ISBN 0-7134-8578-7.

Web Sources