Rebuild-it puzzles

Here is a series of puzzles where you must place the given pieces on the board to get to the desired position (i.e. White is mated or it is White to play from this position and Black must mate White on the next move). You must indicate your solution (where you put the pieces). Both White's and Black's moves played prior to the position, in the position and after the position must be legal moves and all positions must be legal.

Also, you can't place a piece in an occupied square.

Well done, they all appear correct! The last one is the trickiest. I always get confused by pawn directions in messy positions. In standard compositions there are no coordinates and white is always on the bottom. But chess.com often challenges us with "black on the bottom" which we may overlook.

By the way, the "rules of play" part of your puzzle is completely standard chess. Only the setups are somewhat unusual but there is no need to change the standard conventions. You might for instance make a puzzle where you must place a piece somewhere to enable castling! You need not prove the right to castle since the castling convention permits you to castle if not absolutely impossible!

Great idea! I will try to find a way to incorporate it.

I tried to add a puzzle and discovered the instructions were not consistent with the puzzles. Notably you demanded a unique (captureless) last black move which isn't there for #1 and #6. So is that a requirement or not?

Oh, I understand what you mean! I must clarify. In puzzle 5, the bishop has only one square he could have come from to give check in a way to solve the puzzle. But in puzzle 6, the bishop could have gone from many squares to the square he's on to give check. Puzzle 2-5 fall into one category, and 1 & 6 into another.

The notation used for the Black's last move in all of the (1-6) puzzles is unique in algebraic and figurine algebraic notation, but is not in other notations such as the long algebraic notation for puzzles 1 and 6. Puzzles 2,3,4,5's move to find is unique in all notations.

I must admit I am used to think only in algebraic notation (which >99% of people use), so that must have been specified from my part in the puzzle requirement (it now is). Black's last move must be unique in algebraic notation. Well, from the beggining, I intended for one only move in algebraic notation to be right for any puzzle here, it is a requirement. So what is the point if all of Black's last moves resulted in checks in puzzles 1-6, would that be too obvious, you say? Well, I intended to make puzzles with other requirements than checkmate requirements (like drawing requirements) to be made (while still having a unique solution in algebraic notation). But I mostly want the user solving the puzzle to think retroactively and to show their understanding of the previous position in legal chess logic and that was the main purpose of this indication.

But when I think about it, making puzzles with unique moves according to all chess notations (like most of what I did so far, #2-5) could be more interesting, the previous move requirement could have been an outright flaw. 

What is your honest opinion about it (flaw from my part?/what works better(keeping, letting go or reworking puzzles 1 and 6)?)? Algebraic notation is less restrictive for this type of puzzle making , but opening it to any notation could make the process of making these puzzles more challenging/fun.

I.e. I wanted to present a puzzle like nonexistent-a where the answer can't be right without the indication of capture, and the solver must have had to show his understanding of the capture and of continuity in the notation of his answer. The bishop could have come from a8 or b7, the correct algebric notation answer would have absolutely contained an indication of a capture:

Pieces to place: .
Resulting position: White is mated by Black.

It gave me an idea where Black's last move solution is unique in all chess notations! This could have been a way to force the solver an understanding of continuity in legal positions if I had removed the «the answer is never a capture unless necessary» rule, and the answer would have been unique.

Pieces to place: .
Resulting position: White is mated by Black.

Ah, I see. The way the composer community (and I think most chess players) view uniqueness is not through any of the established notation systems. It's purely visual geometry (like the pictures in the FIDE laws) which is more refined and complete than even the long algebraic notation. Chess is a discrete game in which the smallest change is a full move. The change of a full move can be reflected by full board snapshots before and after the move. One of the snapshots is commonly known, either the one before the move or the one after it. The latter is the case when asking for a last move as in the Rebuild-it puzzles. The move is only unique when just one snapshot exists that meets all the requirements for the unknown snapshot.

Of course, the snapshots here too represent an abstract geometrical type with an abstract board and abstract pieces.. But it is awfully easy to understand as you can translate it directly to the physical game as we play it. Even easier for a digital screen which is still closer to the mathematical form.

Interestingly, everybody knows this intuitively - until they start messing with notations! The purpose of the algebraic notations is to implement the geometrical events and not the other way round. The common long algebraic notation is insufficient to express "last moves" but it can be extended to support them. 50 Years ago I developed a system to notate backward moves but also to mix forward and backward moves without confusion. Which was not adopted by the retro-world. Instead they extended the forward algebraic notation to include captured piece types, like Pe7xBd8=Q.

That's where we are now. Possibly the notation systems will continue to evolve but that will not change the concept of move unicity derived from the physical geometric origins of chess.

Puzzle #8 incoming tomorrow - making final modifications

Puzzle 8: I cannot reply to your puzzle since I have no clue what you mean with continuity. If it exist it probably goes under a different name in the composers environment. As you can read in my last post the composers take the physical geometric reality of the game as a point of reference and the notation is always irrelevant. Of course you can create puzzle types based on notation but then they are a challenge in themselves, not part of a chess reality. Though I must admit, a long time ago I saw some composer doing something weird with a mix of reality and notation. I criticized it but never received a response.

I just erased puzzle #8. I will update it tomorrow.

I see, so you advocate strongly for me to update puzzles 1 and 6 then, the philosophy being that notation must not determine move uniqueness .

Not at all! Unicity only affects puzzles when the core solutions would be wrong (or dualed) without it. Most of your puzzle solutions stay the same without the unicity requirement for the last move. In fact, unicity only makes sense when there are at least 2 different solutions with just one delivering last move unicity.

So you would rather advocate for eliminating the requirement to find the last move Black played resulting in the position, like for puzzles 1 and 6?