Two years ago, we had a collection of “random” endgame studies, a small kaleidoscope, in January, and then again in July. With the many things happening in endgame studies that wouldn’t fit into a coherent article, let us take this approach again.
As the readers know, Yours Truly is on a special Discord server for chess composition. Several interesting projects started there, such as the “Hopper” magazine, where recently Andrew Buchanan (Singapore) finished a collection of “help-dead position” problems, aimed at showing different combinations of pawn promotions that lead to a position that can’t be won by either side by any legal series of moves (usually this happens by forcing stalemate, or by reaching an insufficient material distribution such as bishops of the same square color or even bare kings in one case).
In the Tech Mahindra Global Chess League, an absolutely insane endgame happened between two former World Chess Champions in June 2023:
The ending of that game inspired one of the Discord server users to play around, and he found an interesting positional draw after a battle that uses the same knight sacrifice.
The solution for this endgame study is very complicated, as it has many sidelines, so instead of presenting it here, readers should look at the end of this article where it can be replayed.
Recently I saw that in the “Queen’s Gambit” video series that a study by Henri Rinck was used. There is an earlier version of the study that I found interesting. Alois Mouterde in 1921 used a different but similar domination with a diagonal twist that was shown a few more times, and in my opinion, Grigory Slepyan’s 1978 study is among the best that use Mouterde’s idea.
Interestingly, after 1.Nd5+ Kd7 2.B:g7 the only safe square for now is on h4. But after 2.-Rh4 3.g3 the fourth rank turns out to be as trapped as the eighth. With 3.-Rh3 4.Kg2 White wins.
For Black, this study, which gained the 10th place at the Belarusian Championship 1977-1980, is an exercise in pain.
After 1.Rh7+! the play splits, and the author may have wanted the side variations as other main variations: 1.-Kg6 2.K:c2 Re2+ 3.Kd1 Re3 4.Rh6+! K:h6 5.Bd2 wins, or here 3.-Re5 4.Rg7+! K:g7 5.Bc3 wins.
So the better defence is 1.-Kg8 2.K:c2 Re2+! 3.Kd1 Re3 (after Re5 both sacrifices on g7 and h8 would work) 4.Nc4! Rd3+ 5.Kc2 Rd4. The rook is in the perfect position to make the defence plan work, so White can’t sacrifice on g7, as we will see. 6.Rh8+! K:h8 7.Bc3 e5! It seems that Black reached a draw, as taking on d4 leaves White without sufficient material to enforce a checkmate. 8.N:e5 surprisingly dominates the rook. 8.-Ra4 9.Kb3 Re4 10.Ng4/Nd7+ and 11.Nf6+ wins. It would be an incredible logical study if somehow the starting position was reached once with and once without Pa6 in different lines, as without the pawn the entire game would be a draw. This would be what composers nowadays try to achieve.
Ukraine, a now very troubled country, has a rich history of chess composition. The father and son Kapusta are a fine example of that. Both composed checkmates in three moves with unguarded queen sacrifices next to the king in the second move, as seen on their biographies on the Chess Composers blog. Only a few years after his son Viktor (* 1948, IM for Chess Composition) was born, Lev (19 October 1919 - 1990) won the first place in the Ukraine - Uzbekistan match 1954.
Lev Kapusta, Match Ukraine - Uzbekistan 1954, 1st place. White to move and draw
Black has an easy strategy: Control b1, use the pieces to move White’s pieces away and either win both knight and bishop for the pawn or enforce the promotion. There isn’t much White can do. Exchanging pieces only helps Black currently. The only hope is to give at most one piece for the pawn.
The solution manages to do this after sacrificing the bishop: 1.Nc2! Nc1 2.Bb1 Kd3 3.Kb5! Kc3+ — a very interesting moment. White refused to give a discovered check and instead ran into a discovered check himself. A beginner’s mistake often is to overestimate the power of a discovered check, so a small sense of paradox is added by this sequence. 4.Ka4 Bd3 5.Ne3! B:b1 6.Nd5+ Kc2 7.Ne3+ Kd2 8.Nc4+ Kc3 9.Ka3! This move is only possible because the king would fork both pieces after 9.-K:c4 10.K:b2. 9.-Nd3 10.N:b2! N:b2 stalemate
Small fun fact: If you add white pawns on d4 and e7 in the final position, only the move 1.e8N would draw (Charles Michael Bent, EG 1968, mirrored). In fact, without Pd4 it would be the same, but Bent used that additional pawn for his introduction. You can replay his study at the end of the article.
Steffen Slumstrup Nielsen is one of the modern endgame study composers we talked about before in this series. In 2020, he held his jubilee tourney for his 45th birthday. The first prize went to an interesting struggle for a draw by Black that was crushed at the end.
White is far up in material, but surprisingly lines like 1.Kh3? c3 2.Nb5 c2 3.Nd4 c1N, etc. result in a draw. After 1.h7! it seems that Black is lost, but he can play for stalemate(!). 1.-c3 2.Nc6! Nf3+! 3.Kh3 Ng5+! 4.h:g5 c2 5.Nd4! c1Q 6.Ne2+ K:g5 when 7.N:c1? Kh6 is a draw. Similarly, after 7.Nf7+ Kg6 White draws with 8.N:c1? K:h7. So 8.h8N+! would lead to the trivially won endgame of king and three knights against the bare king if it wasn’t for the stalemate defence 8.-Kh5. Taking on c1 is stalemate, so is the game a draw? No, the knight on e2 just moves into the opposite direction: 9.Ng3 mate. It is known that four knights win against a bare queen. Three knights only do under special circumstances like here where the opposing king or queen is quickly lost.
A good method to find unusual endgame studies is to look for special awards, Many of them show something that judges didn’t want to consider for the normal tourney awards, but felt should nevertheless be awarded. Some of those endgame studies are contributions to endgame theory that will never matter in practical play, such as the following endgame of four bishops versus two pawns at and near their starting square. “Wait! That is an easily won material advantage!” Indeed it would be… if the bishops weren’t on the same square colour.
Alexey Troitzky, 64 1932, special prize. White to move and win
This is left for readers to figure out. Troitzky’s solution ends with:
31.d4 b5 32.Bcb2 b4 33.Ba1 b:a3 34.Bh8 a2 35.Kc3 K:a1 36.Kc2 mate
The idea 1.e6 Kg6 2.Kc6! Kf6 3.Kd7 Ng3 4.e7 Ne4 5.e8Q or 2.-Kg7 3.e7! Kf7 4.Kd7 winning must look very familiar. Wait, isn’t this one of the most famous blunders in chess history that led to such a combination? So was the special prize awarded for adding the small second variation that likely was present in the original game? Did Mees just copy the game combination and make it economical?
Well, maybe he did. But if he did, he used a time machine. You see, there is no typo here, the study is from 1941 and not 1951…
David Bronstein, the brilliant out-of-the-box thinker, probably saw that the knight will easily stop the pawn after 57.Ka4 Kf3 58.Ne6 e2 59.Nd4+, but 57.-e2 would be devastating. So the king needs to be pulled closer to protect e1 after the pawn moves, even if no win is possible.
Now 57.Kc2?? seems to draw easily, as 57.-Kf3 58.Ne6 e2 59.Nd4+ stops the pawn. The Soviets were on top of chess knowledge, so it might be that Botvinnik knew Mees’ study, but he also might have easily found on the board that 57.-Kg3!! avoids the “trapped” square f3. Disillusioned, Bronstein resigned.
Certainly an interesting coincidence, but it is not impossible that such things happen in chess. They happened already in other contexts. The story of the Titanic is widely known, an “unsinkable” ship that hit an iceberg and sank on its maiden voyage on 15 April 1912, leading to a great loss of life. In 1898, Morgan Robertson wrote a novella called “Futility”. It was about an “unsinkable” ship that hits an iceberg on its maiden voyage and sinks, leading to a great loss of life. There are other parallels between the novella and the real events that can be attributed to Robertson’s knowledge in the field, but one certainly can’t: the name of his ship was Titan.
The coincidence of Edgar Allan Poe’s prediction of a cannibalism in “The Narrative of Arthur Gordon Pym of Nantucket” is widely known, in which Richard Parker is eaten by his companions after they shipwrecked, a story that came true in 1884 when a 17-year-old boy named Richard Parker suffered that fate.
But you might not have heard of a story called “Eureka” that Poe wrote in the final year of his life. Here he describes a primordial particle that causes an instantaneous flash to create an ever-expanding universe. This crazy idea certainly would explain the mysteries that baffled scientists at the time, but it couldn’t be taken seriously… until scientists arrived at the conclusion that the universe must have been condensed at the beginning and then instantaneously expanded. The originally ironically used term “Big Bang” stuck to this. The theory also explains why the night sky is so dark, as it postulates that the light of many stars just hasn’t reached us yet. Edward Robert Harrison published his findings in 1987 in his “Darkness at Night”, but admitted there that his conclusion was anticipated by Poe.
I was shocked to see one of my favourite studies cooked, but it turned out that only a bad line was cooked. Black has a better defence there, so the study is correct.
The solution is the beautiful 1.g7+! B:g7 2.Ba3+ Kf7 3.Bd5+ Kg6 4.Kf2 Nh1+ 5.B:h1 B:d4+ 6.Kg3 Be5+ 7.Kh4! Bf6+ 8.K:h3 h5 9.Be4 Kh6 10.g5+! B:g5 11.Bf8 mate or 10.-K:g5 11.Bc1 mate
The issue is that after 5.Ke3 B:d4+ 6.K:d4 Nf2 7.Ke3! the endgame of two bishops against knight is reached eventually, which is known to be won thanks to computer analysis. It was one of the first endgames where computers changed the human verdict, which held that one fortress exists. For all intents and purposes, that fortress is likely to be practically a draw but theoretically won.
Obviously Black doesn’t have to see if he can reach that fortress, but rather play 5.Ke3 Ng3! 6.Kf3 Nf1 and the endgame is actually drawn. Of course, here it is necessary to point out that White wins with 6.Kf2, turning the supposed destructive dual into only a loss of time by repeating two moves…
Let us close this month with two special prizes that were awarded to rook endgames. First White will win against one pawn, then draw against two.
E. Kolesnikov, Shakhmaty v SSSR 1989, special prize. White to move and win
What? 1.Kf7 K:f2 2.Ke6 and 3.K:d5 wins easily? Yes, but not 1.Kf7? K:d4! 2.Ke6 Ke3. So White can play 1.Rf8! K:d4 2.Kf7 Ke3 3.Ke6 d4 4.Kd5 d3 5.Kc4 d2 but this also only leads to a draw. The important point is that 3.Re8+! Kf3 4.Rd8 Ke4 5.Ke6 d4 6.Kd6 d3 7.Kc5 Ke3 8.Kc4 d2 9.Kc3 wins
Here the solution is straightforward but still interesting:
1.Rd1+! Kf2 2.Rh1 Kg2 3.Re1 Kf2 4.Rh1 Kg3 5.Re1 Kf3 6.Rh1 Kg3 7.Re1 Kf4(!) 8.Re2! Kf3 9.Rh2 Kg3 10.Re2 draws
One last puzzle for readers: Would this still be a draw if White had the king on a8? You can find the solution as the last replayable entry.
