# This file contains positions where INSUFFICIENT MATERIAL is the theme. # # Dabbaba solves all of these positions. # #Not all the positions have a single move as the right( wrong) solution, #because this test is also used to reveal when a program claims mate, draw or #don't claim anything. # #Be aware that some positions may be solved by a program without the right #knowledge, if it calculates deep enough and fx detects that a pawn is #promoted. #Such positions should be considered 'not solved'. #Most positions should be solved 'immediately'. Only two positions is allowed #a longer search (they are in INSULONG.EPD). This is described below. # KB1k4/8/8/8/8/8/8/8 w claim draw! KN1k4/8/8/8/8/8/8/8 b claim draw! # kbN5/8/K7/8/8/8/8/8 w Nc8-b6+! both have insuff. material, but claim mate! KNb5/8/1k6/8/8/8/8/8 b Bc8-b7+! both have insuff. material, but claim mate! # #These positions are almost as those above... kb1N4/8/K7/8/8/8/8/8 w both have insufficient material; (claim?) draw! KN1b4/8/1k6/8/8/8/8/8 b both have insufficient material; (claim?) draw! # #If a program does not know that two knights cannot mate, it will capture the #knight on b6, as this move wins both knights which are more worth than a #rook. 8/rKn5/1n6/8/8/k7/8/8 w Kb7xa7! # #Here comes the same position twice; with both black and white to move. #Black should capture the white pawn on h2; white should save it. #There are some rare positions, where you can win with two knights if the #opponent has a pawn, so a stalemate can be avoided. #But I doubt this is the case here, so concentrate on the white pawn on h2 #instead of the temtating knight on h1. 8/7N/7p/7p/7K/8/6kP/7N b Kg2xh2! best chance for a draw... 8/7N/7p/7p/7K/8/6kP/7N w h2-h3! gives an easy win # #Many programs will not capture on a8, which is a sure draw. #They play -,Kc6 and feel an advantage of more centralised pieces! N6K/1k6/8/8/4b3/8/8/8 b Kb7xa8! # #Here the positions show, if your program will sacrifice a piece to get a #draw. #It is likely, that the program will not show a draw score, as black will not #capture the piece, but prefer a more centralised position as above. 1B5K/1n6/1p6/2k5/8/8/8/8 w Bb8-a7! 7K/Bn6/1pk5/8/8/8/8/8 w Ba7xb6! # #The same as above, but leaving black with a lone bishop instead of a knight. #This is a little more tricky, as Nc8-b6+ in a short search seems to win the #black bishop, but a little longer search shows that black with Kc5-c6-b7 wins #the white knight. #The position ought to be solved 'immediately' (3-ply search; no extensions). b1N4K/8/8/1p6/2k5/8/8/8 w Nc8-d6+! b6K/8/3N4/1pk5/8/8/8/8 w Nd6xb5! # #Here will Nd8-e6+ win the bishop without the knight gets caught, but a longer #search (not allowed in this test) will show, that black promotes his pawn. 1b1N3K/8/8/8/3k1p2/8/8/8 w Nd8-e6+! # #Here again some positions with the lone knight bishop as the theme. #The positions are duplicated with without a black pawn on f5. #Some programs may only recognise insufficient material when one of the #parts have a lone king and is fooled by the black pawn on f5. #The pawn on f5 may also delay a deep search (not allowed!) to understand #that the white pawn gets promoted. 8/8/8/5p2/5K2/8/3NkP2/8 b Ke2xf2! 8/8/8/8/5K2/8/3NkP2/8 b Ke2xf2! 8/8/8/5p2/5K2/8/3BkP2/8 b Ke2xf2! 8/8/8/8/5K2/8/3BkP2/8 b Ke2xf2! # #A special case with the same material as above. A long search will show, that #it is correct to capture the knight here, as black promotes first and wins. #So this is only a test of insufficient material rules, that they do #consider the result of a long search. #At least 11 ply should be searched in this position. 8/p7/8/8/7K/8/5NkP/8 b Kg2xf2! # #Now two positions each with 3 moves, where 2 knights is the theme. 8/8/8/8/3N3B/2N5/3k2K1/4r3 w Bh4xe1+! Gives an easy win. 8/8/8/8/3N4/2N5/3k2K1/4B3 b Kd2xe1? Don't play this and claim draw. 8/8/8/8/3N4/2N5/6K1/4k3 w Nd4-f3+! Don't claim draw, but mate. # #Here both parts have insufficient material, but white wins with a quiet move. 8/8/8/6N1/6b1/8/5K2/5N1k w Ng5-e4! Don't claim draw, but mate. 8/8/8/8/4N1b1/8/5K2/5N1k b Don't claim draw, but mate. 8/8/8/8/4N3/5b2/5K2/5N1k w Ne4-g3+! Don't claim draw, but mate. # #Now the theme is pawns on the a-file or h-file combined with a bishop on the #wrong square-colour. #Don't capture the bishop in this position although it is more centralised #than the knight and it is more worth than the knight. #This may take a little search to solve, if the program is constructed like #Dabbaba, that only recognises the position as draw when the black king is on #the four squares nearest the corner. #A search to a depth of 11 ply is allowed here, so the king can reach h8. #But a very, very deep search showing the knight can help the pawn to #promote is not allowed. 8/8/2NkB3/8/8/8/7P/7K b Kd6xc6! # #Black must keep control of the corner and not 'escape out in the free' with #-,Kf8 as white wins with Bg6-h7. 6k1/8/6BP/8/8/8/7K/8 b Kg8-h8! 6k1/8/6BP/8/8/7P/7K/8 b Kg8-h8! An extra pawn on h3 may confuse some programs # #After -,Kxh6? Kg8,Kg5 (-,Kg6 h4,Kh6 h5,Kg5 Kg7) Kg7,Kh4 Kh6,Kg3 Kg5 white wins 8/5K1k/7P/8/6B1/7P/8/8 b Kh7-h8! # #Here comes a position in 4 versions: #With without pawn h2 and with without bishop e8. #Fx will Genius2 only play -,Bh3! if there is a pawn on h2 and no bishop on #e8! Genius5 manages it if there is no bishop on e8 (h2 does not matter) B1b4k/8/8/8/8/8/6P1/5K2 b Bc8-h3! B1b1B2k/8/8/8/8/8/6P1/5K2 b Bc8-h3! B1b4k/8/8/8/8/8/6PP/5K2 b Bc8-h3! B1b1B2k/8/8/8/8/8/6PP/5K2 b Bc8-h3! # #In the above positions, a new theme was introduced: #Two bishops on the same square-colour. They cannot mate a lone king. # #Fx Genius5 does not know immediately, that it should capture on b8. For 2 #minutes it considers -,Kxc6?? on a 50 Mhz 486. 1BB5/2k5/K1B5/8/8/8/8/8 b Kc7xb8! # #This example is taken from PLY 1996/2 page 16 and is the only example that I #know from practical play. #White cannot move his bishop: fx Bd5,Bc6 Bf7?,Bc6 e8Q,Bxg2+ mate. #Promotion to anything else than a knight is answered by -,Bxg2+! #Genius5 finds the right move here. 8/4P3/2B5/8/8/6p1/5kP1/5b1K w e7-e8N! # #x=exit