Self-solving puzzles

Basically, these are puzzles where all the moves are absolutely forced, eventually leading to checkmate (or stalemate possibly).

Here is a well known example:

White checkmates in 6:

We used to have a thread full of these. Here are a couple of examples from the thread:

White checkmates in 3:

White checkmates in 3:

White Checkmates in 3:

White Checkmates in 5:

Black Checkmates in 5:

White Checkmates in 7:

Challenge:

Can you create a self-solving puzzle where white checkmates in 8? the position has to be legal. For clarification, there must always be only one legal move possible.

oh, i love these forums that actually discuss chess, we havn't gotten many good forums since the spammers and trolls and people who aren't here to learn chess took over.

anyway heres my first try:

I give up :( i will try again later.

Yes, for beginners, forget about 8 moves. Just try to do 2-3 moves. These puzzles can be quite tricky to compose, but fun too 

watcha, your puzzle idea is interesting. But of course from a solver perspective, it cannot be proven that black's last move was not b6-b5+. Still, I like the innovation. We had a couple of those in the last thread as well :)

Hah, this one is great (prob #5 from post 1).

Thanks BDP.

That was an extension to a puzzle I composed with only 4 pawns:

It took me a while to add another pawn while keeping it legal, but was well worth it 

In the original thread, at some point we started challenging each other to create self-solving puzzles with specific requirements. It was really fun and challenging!

Surely we can be more ambitious than 2-3 moves. Even cheesy stuff like this gets you to move 6:

I can extend that to 7 (if you are allowed to have 2 possible moves that both deliver mate)

This thread reminds me of maximummers, where the only legal move (for both sides) is the geometrically LONGEST one...

Mate in 50, nowhere is there a choice

Yes it is true, the easiest way to create a long problem is with a consecutive string of checks. We tried to spice things up from time to time by creating sort of zugzwang and such. While it was nice, it did not make for very long problems.

DefinitelyNotGm:

That is true, although we tried to avoid the possibility of more than one legal move. I believe that if it is allowed, the world record is 13 moves or so (a ton of promoted knights all jumping to the same square if I remember correctly).

rooperi:

Ah yes, I know that one. I wonder how long it took to compose that.

In the mean time, here is an easier challenge:

Create a self-solving problem where one side makes two consecutive knight moves.

mate in 3

I don't understand how Bxf5 on move 24 is the longest move available.

@kromhawk_31:

Because it is on a diagonal.

According to the Pythagorean theorem in any right triangle, the square of the length of the hypotenuse equals the sum of the squares of the lengths of the two other sides. If the hypotenuse has length c, and the legs have lengths a and b, then the theorem states that

Here the bishop makes a four square move along the diagonal:

4*4 + 4*4 = 32

The rook could make a move 5 squares long.

Since 5*5 = 25 < 32, the bishop move is longer.

Good job. Here is the next challenge:

Try to create a self-solving puzzle with as many moves as you can, without using queens at all.

Here is an example of an illegal 8 move self-solving puzzle made by doodinthemood:

'Machine gun rooks' can do the same trick as 'machine gun queens'.

mate in 5:

Some restriction would be needed to avoid such solutions. May be that the starting position should be legal and you can't have more than two diagonally moving piece on the same diagonal or straight line moving piece on the same file/rank by each side.

improved version, mate in 6:

It is a puzzle in itself how to find objective restricitions for puzzle creation which if observed guarantee positions 'reasonably probable' in a game. Certainly the minimum requirement should be that the starting position of the puzzle is legal. Apart from that it is not that obvious.