29 April 2011

And the story continues ...

I kept on playing with these puzzles I mentioned in the previous post. This stuff is really addictive! Alain had listed some unsolved puzzles which I managed to crack. Weeeeeeeeee! Maybe you want to try one of them yourself?!
  1. Find a position with unique realization after a series of 9 white moves ending with the mate of the black king delivered by a knight. (T. Luffingham)
  2. Find a position with unique realization after a series of 10 white moves ending with the mate of the black king by a double check. (Richard Stanley)
  3. Find a position with unique realization after a series of 10 white moves ending with the mate of the black king by a discovered check but no double check. (Richard Stanley)
There's still one left for which I haven't found the solution, yet. It's by Gianni Donati: Find a position with unique realization after a series of 13 white moves ending with the mate of the black king delivered by a promotion to bishop. Can you help me?

Finally, let's return to problems with the theme that the given move determines all the previous moves. An interesting example is this one for which I also managed to find the missing solution: Find the game that ends with the direct mate 6. Bh7# (François Labelle). This is accomplished by the moves 1. e3 g5 2. Qh5 Bg7 3. Qxh7 Nh6 4. Qxh6 0-0 5. Bd3 Bh8 6. Bh7#. If you feed a solving software like Euclide, Natch or Popeye with this position and ask it for the proof game that produces the position after 11 half-moves, its answer might surprise you. The software tells you that the solution is not unique. Hm, what went wrong?! Nothing! This example clearly shows the difference between this type of chess puzzles and the proof games.

In a proof game you start with a final position and there has to be a unique sequence of a given amount of moves to get there. You have no information about any of the moves. On the other hand, this other problem type does not give you the final position but the last move of the game claiming that all the previous moves are uniquely determined thereby. Among all the solutions to the example above that lead to the final position, there's just one (!) that has the move 6. Bh7#, all others end with 6. Bxh7#. Tricky, eh?

These are some easy ones which won't cause much if any headache:
  1. 3. Qxf8+
  2. 3. Qe4+
  3. 3. - Bf6+
  4. 3. - Bxb3+
  5. 3. - Rxe5+
  6. 4. Ra8
  7. 4. - Re2

1 1. Nc3 2. Nd5 3. c3 4. Qc2 5. Qxh7 6. Qxg8 7. Qxg7 8. Qe5 9. Nf6#
2 1. h4 2. Rh3 (or
1. a4 2. Ra3) 3. Rd3 4. Rxd7 5. d4 6. d5 7. d6 8. Qd5 9. Qc6 10. Rxe7
3 1. a4 2. Ra3 3. Rg3 4. b3 5. Bb2 6. Bxg7 7. Bxf8 8. Rxg8 9. Rxh8 10. Bh6 (or 10. Bg7)
4 1. e4 f5 2. Qf3 fxe4 3. Qxf8+
5 1. d4 e5 2. Qd3 exd4 3. Qe4+
6 1. d3 e6 2. Kd2 Be7 3. Kc3 Bf6+
7 1. c3 d6 2. Qb3 Be6 3. Kd1 Bxb3+
8 1. e4 h5 2. Qxh5 Rxh5 3. e5 Rxe5+
9 1. a4 b5 2. axb5 Nc6 3. Rxa7 Rb8 4. Ra8
10 1. e4 h5 2. Qxh5 Rxh5 3. e5 Rxe5+ 4. Kd1 Re2

22 April 2011

Again no diagrams

In previous posts, I wrote about proof games. There, you have a diagram together with the stipulation to reach this position after a given number of moves. Of course, these moves (i.e. their order) are unique.

Lately, I read about chess puzzles dealing with sort of an extension/variation of this — not having a diagram. They ask to find a position that meets a certain goal and for which there is a unique proof game in the given amount of moves. I am sure, you've already seen such puzzles. There are several (sub)types and the French problemist Alain Brobecker has collected/classified lots of them.

The first type are games to be found that are uniquely determined by the last move. An example (puzzle #1): Find the game ending with 4. - b5# (François Labelle). In 2003, the Canadian problemist François Labelle wrote a computer program allowing to find such games having up to 11 half-moves.

In another type, you are only allowed to make moves with the white pieces (a seriesmover). Labelle's program can also handle that. The second puzzle is this: Find the series of White moves ending with 6. exf8=R# (François Labelle).

Finally, there is a third and most challenging type. These tasks just describe a certain theme that is to be realized within the given number of moves by a unique proof game. Here is puzzle #3 for you (author unknown): Find a position with unique realization in 2.5 moves ending with the mate of the black king. Three solutions.

1 1. d4 c6 2. Kd2 Qa5+ 3. Kd3 Qa3+ 4. Kc4 b5#
2 1. d4 2. d5 3. d6 4. dxe7 5. Qd6 6.exf8=R#
3 Originally, Alain only gave one solution, but already mentioned there was another one: 1. e4 f5 2. exf5 g5 3. Qh5#. As you can see, the mate after 2.5 moves as such can be achieved by various ways. But, the challenge is to find a final position which has a unique move order that leads to it. Therefore, e.g., the position after 1. e4 f5 2. e5 g5 3. Qh5# is not allowed, as you also could arrive there after 1. e4 g5 2. e5 f5 3. Qh5#.
I took up the challenge and searched for the second solution. Finally, I managed to crack the puzzle. In fact, I found even two more solutions (though not really different)! 1. e3 e5 2. Qh5 Ke7 3. Qe5# and 1. e4 e5 2. Qh5 Ke7 3. Qe5#.

15 April 2011


Here's the chess problem to which you've already seen the solution:

Comins Mansfield
Chess Amateur, 1926

Some remarks:
The black pawn on c7 can be cancelled without causing any damage. The white queen can be replaced by a third white bishop, therefore, wQb1 and wBa8 can change places. Moreover, there are possible setups with the wRf1 on f2 and the bNe7 on f6 — but you can't combine all possibilities.

Not many days ago, I read about the HAP (Human/Animal/Pawn) notation. What is that? It's a special kind of chess game notation, which can also be used for chess problems only. It was invented by Mark Tilford.

A move is denoted by just giving the type of unit that makes the move. There are three types of units:
  • Human: King (K), Queen (Q) or Bishop (B)
  • Animal: Knight (N) or Rook (R) (in the early days of chess, the rook was the elephant)
  • Pawn (P) (not considered human for whatever reason)
Additionally, the following notation is applied:
  • Castling: "O" (not revealing whether it's short or long castling)
  • Promotion: "P=A" or "P=H"
  • Captures (including en passant): "x" followed by the type captured
  • Checks: "+" for all check types
  • Mate: "#"
  • Stalemate, dead position or draw (50 moves, threefold repetition): "="

Using this HAP notation, every chess game can be written down. But, in most cases, different games give the same HAP translation. In other words, you might derive several different games from an arbitrary HAP notation. For example, take this one: 1. P P 2. P H#, the famous fool's mate. There are eight ways of employing those moves, e.g. 1. g4 e6 2. f3 Qh4# or 1.f4 e5 2.g4 Qh4#. Of course, we are rather interested in such HAP notations that belong to only one game each. With a slight modification of the given example, we can fulfill this request with White mating (puzzle #1): 1. P P 2. PxP P 3. H#. Can you find the proper moves?

Some people started composing HAP problems and there is even a homebrew retroanalysis program written by Mario Richter to test those. This is puzzle #2 for you: 1. P P 2. A A 3. H H 4. AxP AxP 5. O HxP# (composed by Bernd Gräfrath, 2007/11/23).

I hope you played a little bit with this and enjoyed it. These are the solutions to the two HAP problems:
1 1. e4 f5 2. exf5 g5 3. Qh5#
2 1. g4 e5 2. Nf3 Nh6 3. Bg2 Qh4 4. Nxe5 Nxg4 5. O-O Qxh2#

08 April 2011


Yes, I owe you the solution to the joke problem you saw in the last post. Did you succeed in unmasking the "pusher"? Or were you desperately trying to crack the puzzle and finally gave up in despair?

Don't worry, you were not really supposed to solve it. This kind of problems is rather for the people's amusement. Firstly, for the composer to see (well, if I only could) people shaking their heads looking at it and making wild guesses. Secondly, hopefully, for those who bother to have a look at the solution and comprehend the special idea behind.

OK, the pushing piece is the white king! And what can he hustle? Of course, there are his subjects on the h-file. First poke: 1. Kh6+! As a result, all other pieces are being pushed one square downwards: Rh6→h5, Bh5→h4, Bh4→h3. The choices of the black king are irrelevant, White's moves are always the same. Second poke: 2. Kh5+ (Rh5→h4, Bh4→h3, Bh3→h2). Third poke: 3. Kh4# (Rh4→h3, Bh3→h2, Bh2→h1). That's all. I hope you enjoyed it at least a little bit.

Would you like to compose a chess problem yourself? Just try — I give you the solution and you set up the position. The English word for this sort of challenge is synthetics, the Germans say Problemsynthese and the Dutch use the word Reconstructie. In several chess magazines, there were (in fact still are) such reconstruction tasks, e.g. Die Schwalbe, Probleemblad, The Problemist and Schach. Here is such a task for you.

Construct a twomover that has only the following variations:

1. Nd1! (threats 2. Nc3#)
1. - Rb3 2.Kxb3#
1. - Rb2+ 2.Kxb2#
1. - Qa5 2.Kc1#
1. - Qd4 2.Ng5#
1. - Qd3+ 2.exd3#
1. - Qd2+ 2.Kxd2#
1. - Qxd1+ 2.Kxd1#
1. - Qxe6 2.Kc3#

The composer used 18 men, but 17 already suffice. Again, take your time and have fun finding out how this works!

01 April 2011

Keep Pushin'

There are times when, apparently, things are so easy, everything runs smoothly. But there are also many situations when everything seems to be stuck, nothing works, no possibility to get ahead. Then, it's good to have some support, a little bit luck might help, too. Still, it's up to you to exert yourself, to keep pushing. This applies to life as well as to solving chess problems.

And here we arrive — surprise, surprise — at today's theme. I composed a chess puzzle and all I can say: Keep pushin'! I won't give the solution immediately, so take your time and (try to) enjoy. Maybe, you want to motivate yourself by listening to a song by the American rock band Reo Speedwagon. For this occasion, I'd recommend a song called ... exactly (wow, it's already 35 years old).

Gerson Berlinger