# Unusual and Challenging Chess Problems

I've been through both of Smullyan's books, and they're good introductory material to work through, as a source of simpler problems. I would still cite Beluhov's text that I linked to earlier as the best primer to solving though - learning to retracting moves is the simplest way to solve retros.

Quickly going through Hale's page, the problems are all pitched at a good level, good exercises in piece inventory and pawn capture counting. I do however dislike the use of the phrasing "beginning with black's second last move", even in an entry-level text - it encourages a messy way of thinking in terms of playing forward, rather than the cleaner way of thinking in terms of retractions. ("Last 4 single moves?" is the usual way to phrase it.) It's much simpler to solve when thinking with retractions; in my own experience at least, and for a friend I introduced to retros.

Also, I believe Hale's 7th problem is cooked - I see no problem with the black king retracting two light bishop captures (many ways of doing this).

--

Good solution for the original I posted, and well detailed. So the "point" of the problem is that white isn't in mate, but is forced to deliver mate in 1. ("Mate?" would be a humorous stipulation.) If you liked that one, then this one is more of the same (from 4 years back):

@Bob No worries, if you haven't looked into retros much, maybe check out proof games first as they are relatively accessible and nowadays the most popular retro sub-genre.

@Remellion Okay, I just downloaded Beluhov's intro and it looks pretty comprehensive...

Looks like you're right about Hale's Problem 7 being cooked. I should let him know.

Re your original, it's funny how I didn't notice White is mated in the diagram if not for the ep capture! A witty idea and I agree that "Mate?" would be a fitting stipulation.

Here's my attempt at your earlier retro. The black pawns took 8 captures to reach their diagram positions, accounting for all missing white units. Black is missing 5 units, 2 of which were captured by W's q-side Ps. Black's last move was either c6-c5# or c7-c5+. If we retract c6-c5#, White has two potential previous moves that don't create impossible checks:

(1) b2xc3 implies that the original c1-B was captured at home, but that's impossible because as noted the BPs had captured all 8 missing white units.

(2) c7xb8(B) means that this promoted B came from the e-file, requiring 3 captures. These 3 captures plus the other 2 mentioned above account for all 5 missing black units. However, all of these white and black P captures took place on the a- to e-files. In order for White and Black's k-side Ps to reach those files to be captured (or to replace original units that were captured), they must have promoted, but with no more spare units for these Ps to capture, they couldn't have got around each other to promote.

So c6-c5# is an illegal retraction, while c7-c5+ is possible because White could have just played Nc6-a7 or Kd5-d4 (legal check from e4-B). Therefore c7-c5+ was the last move and White can mate with bxc6 e.p.

I'm learning the ins and outs of retrograde analysis by going through your solutions, Peter.

By the way, Remellion, you write "I would still cite Beluhov's text that I linked to earlier as the best primer to solving though - learning to retracting moves is the simplest way to solve retros." but your first post in this thread is No. 33 and you reference Beluhov for the first time in No. 41.  That's my retrograde analysis.

Good job Peter - not missing anything at all. (Even the notation for +/# is correct!) Once again a "Mate?", with slightly more interesting retroplay and mating position. Hope you enjoyed it.

@fightingbob : In post 37 I gave a link that looked like this. (Click the "this".) That gives a link to a post by Beluhov himself, I just didn't give his name in mine. I highly recommend the pdf (indeed a small book) - explanations accompanied by a host of simple exercises that gradually grow in difficulty to full-blown classical retros. Over 150 diagrams in the text.

An alternate primer is another thread on this forum, less organised and with just a couple exercises. Do check out the profile of the thread starter BigDoggProblem - him sharing a ton of 20+ move PGs was pretty much what started me out solving (and then composing) retros.

Thanks much, Remellion.  Apparently the light blue "this" blended in and I missed it as I skimmed Post #37 looking for the name Beluhov.  I went to bolded brown after I received a few complaints in my early posts.

I'm off to download the Beluhov's Introduction to this interesting, and to me unusual, discipline.  Thanks, again.

Here is a puzzle (I'm not sure of its type) with two parts, based on the position below. Part 1 is to identify all legal (non-checking) squares on the board where a white knight can be placed so that White to play is winning. Part 2 is to give a concise statement of White's winning procedure that covers every winning position.

White to place a white knight on the board, play, and win

n9531l wrote:

Here is a puzzle (I'm not sure of its type) with two parts, based on the position below. Part 1 is to identify all legal (non-checking) squares on the board where a white knight can be placed so that White to play is winning. Part 2 is to give a concise statement of White's winning procedure that covers every winning position.

White to place a white knight on the board, play, and win

Original Post 6/11/18

Hi Bob.  Good to hear from you again after all this time.

First, the knight must begin on a dark square, in fact, it can begin on any dark square except a1 (stalemate), b4, c3, c1 (all checking the king) and h8 (unable to reach b3 in three moves).

It is vital for White's knight to reach b3 in three moves in case Black attempts stalemate by moving his pawn instead of his king.  Eventually, White will play the knight to c1 while the Black king is on a1, forcing the pawn to move to a2.

Ultimately, the White knight will end up on b3 mating Black either by Black's choice or by force.

Except for those squares where the White knight reaches c1 in two moves (i.e. c5, e5, d4, f4, g3 and g1), the knight must be maneuvered to c5 while Black's king is on a1 to answer either ...a2 or ...Ka2 with mate or preparation for mate respectively.

Arceus123123 wrote:

Thanks, Arceus123123, but a hate using ISSUU.  I guess it's okay if you are using it for reference, but using ISSUU to read the book from cover to cover would drive me crazy.  Frankly, I like owning my own PDFs.

n9531l wrote:

Here is a puzzle (I'm not sure of its type) with two parts, based on the position below. Part 1 is to identify all legal (non-checking) squares on the board where a white knight can be placed so that White to play is winning. Part 2 is to give a concise statement of White's winning procedure that covers every winning position.

White to place a white knight on the board, play, and win

For a tiny hint this old Daily Puzzle:  https://www.chess.com/forum/view/daily-puzzles/1-27-2018-giving-to-get?page=1 may help a bit.

fightingbob はこう書きました：

First, the knight must begin on a dark square, in fact, it can begin on any dark square except a1 (stalemate), b4, c3, c1 (all checking the king) and h8 (unable to reach b3 in three moves).

It is vital for White's knight to reach b3 in three moves in case Black attempts stalemate by moving his pawn instead of his king.  Eventually, White will play the knight to c1 while the Black king is on a1, forcing the pawn to move to a2.

Ultimately, the White knight will end up on b3 mating Black either by Black's choice or by force.

Except for those squares where the White knight reaches c1 in two moves (i.e. c5, e5, d4, f4, g3 and g1), the knight must be maneuvered to c5 while Black's king is on a1 to answer either ...a2 or ...Ka2 with mate or preparation for mate respectively.

Almost a complete solution.

Just missing 4 light squares (a6/c6/d5/d3) for the mate 1. Nb4+ Ka1 2. Kc1 a2 3. Nc2#

The mate is known as Stamma's mate.

Yes, I missed the variation where the king clears the c2 square instead of requiring the knight to move.  Though I know the name Philipp Stamma in the history of chess, I should have known Stamma's mate; I didn't.

fightingbob wrote:

Yes, I missed the variation where the king clears the c2 square instead of requiring the knight to move.  Though I know the name Philipp Stamma in the history of chess, I should have known Stamma's mate; I didn't.

The above posts give the solution to Part 1, namely the white squares from which the knight can move to b4, and any legal black square except the two corner squares. For Part 2, here's my version of the winning procedure. From a white square, mate in 3 by checking with the knight and moving the king to c1 and the knight to c2. From a black square, there are two steps. 1. If possible, move the knight to c1 in two moves without stalemating Black, and mate with Nb3. 2. If that's not possible, move the knight to c5 or d4 in two moves, and go back to step 1.