Do you believe in endgame studies? By that question I don’t mean to ask whether you believe in their existence, but rather whether you believe that they work. Since I started this column it has been mentioned to me several times that the trouble with studies is that often (1) no analysis is given of what happens after Black plays better moves than in the composer’s main line and (2) no mention is made of alternative White moves. This can often lead to suspicion that a particular study is unsound. Sadly of course, some studies are indeed unsound, such unsoundness sometimes being discovered many years after publication.
For an endgame to be considered sound, two criteria need to be met. First, White must be able to achieve his aim (win or draw) against any and all defences that Black may care to throw at him. Secondly, there must be one and only one way for White to achieve his aim (win or draw) on each and every move of the composer’s intended main line, which may branch into sub-lines. An alternative on the first move is called a cook and an alternative at a later move is called a dual. There is no necessity for the play outside of the main line to be dual-free unless such play forms part of the composer’s intended theme. The point at which the composer’s intended main line ends is up to the composer, but normally this will be at the last unique white move of the solution.
The issue that causes the confusion is the composer’s intended main line, as quite often this does not follow black moves that could be considered best. The composer’s intention (hopefully an interesting and unique winning or drawing manoeuvre) is the reason for the existence of the study and if the composer can arrange for it to follow Black’s absolutely best moves, then he will do so. However, as any composer will tell you, sometimes that isn’t possible and so compromises have to be reached.
The way to investigate a study’s soundness is by analysis. The composer will have done a lot of this before publication, but quite often, because of space requirements, or not wishing to teach grandmothers to suck eggs, or because it just isn’t very interesting, it will be edited out at publication. One of the examples shown to me as a 'suspicious' study is our first, which was re-published some years ago with just the main line and no notes. The particular question being asked was What happens after 1...Nc3? I am pleased to say that computer analysis has answered that question and has shown the study to be sound apart from a dual at move 9 that could be considered unimportant. Unlike material that I have presented here previously, this is a didactic study.
Marcel van Herck
Het Belgisch Schaakbord, 1984
White to play and draw
1.a4! 1.Kc7? Nc3 2.a3 (2.Kc6 Nxa2 3.Kc5 Ke4 4.Kc4 Kf3 5.Kd3 Kg2 6.Ke3 Kxh2 7.Kf2 Nb4 8.Kf1 Kg3 9.Kg1 Nd3 10.Kf1 h2 0-1) 2...Ke4 3.Kc6 Kf3 4.Kc5 Kg2 5.Kd4 Kxh2 6.Ke3 (6.Kxc3 Kg3 7.a4 h2 0-1) 6...Kg3 0-1 1...Nd2 1...Nc3 2.a5 Nd5 (2...Kd6 3.a6 Kc6 4.a7 Kb7 5.Ke7 =) 3.Kc8 Kd6 4.a6 Nb6+ 5.Kb7 Kc5 6.a7 Kb5 7.a8Q Nxa8 8.Kxa8 Kc6 9.Ka7 Kd5 10.Kb6 Ke4 11.Kc5 Kf3 12.Kd4 Kg2 13.Ke3 Kxh2 14.Kf2=; 1...Kd6 2.a5 Nc3 3.a6 Nb5 4.Ke8 Ke6 5.Kd8 Kd6 6.Ke8 Nc7+ 7.Kf7 Nxa6 8.Kf6 = 2.a5 2.Kc7? Nb3 3.Kb6 Ke4 4.Kb5 Kf3 5.Kc4 Kg2 6.Kd3 Kxh2 0-1 2...Kd6 3.Kc8 Kc6 4.Kb8 Kb5 5.Kb7 Kxa5 6.Kc6 Nf1 6...Nf3 leads to the same thing 7.Kd5 Nxh2 = 7.Kd5 Nxh2 8.Ke4 Kb4 9.Kf4 There is a dual here - 9.Ke3 Ng4+ (9...Nf1+ 10.Kf3 transposes to the main line.) 10.Kf3 Kc3 11.Kg3 h2 12.Kg2 = 9...Nf1 10.Kf3 Kc3 11.Kf2 h2 12.Kg2 =
This study, composed before endgame databases, is a considerable achievement. Its Belgian composer, Marcel van Herck is a long-time member of the Belgian solving team and is the man who takes subscriptions for EG, the premier international endgame study magazine, published by ARVES. If you enjoy endgame studies you should read that magazine.
Our study for solving is also by Marcel.
Marcel van Herck
White to play and draw/p>
1.Bd2+! Other moves are soon eliminated - 1.Kxa2? Nc1+! (1...c1Q? 2.Bd2+ Nxd2 allows White to give perpetual check, unless Black elects to stalemate White. 3.Rf3+ Ke5 4.Rf5+ Kd6 5.Rd5+ =) 2.Ka3 Nxd3 3.Bd2+ Kf3 4.Bh6 Ke2 5.Ka2 Kd1 6.Ka3 Nf3 7.Kb3 Nd2+ 0-1; 1.Rxd4+? Ke3 2.Bf2+ Kxf2 3.Kxb3 c1Q 4.Kxa2 Qc2+ 5.Ka3 Qc3+ 0-1 1...Nxd2 2.Kxa2 2.Rxd4+? Ke3 3.Kxa2 Kxd4 4.Kb2 Kd3 5.Ka2 Nc4 0-1 2...c1N+ This underpromotion is necessary as 2...c1Q? leads to the perpetual check we have already seen in an earlier note. 3.Kb2! Nxd3+ 4.Kc3 White attacks all three knights ... 4...Ke3 = ... and Black defends them all, but White is stalemated.