Valentín's website update thread - comments and appreciation

[Valentin Albillo]

.
Hi, pier4r and all of you chess lovers,

(03-17-2023 11:06 PM)pier4r Wrote:  Interesting the chess section with tests again the chess engines of the time (with the hw of the time too) https://albillo.hpcalc.org/files/misc/chesstests

Thanks for your appreciation, pier4r, here are some highlights from my site's Chess Tests Section:

Among the Basic Suite test positions, Position 38.- A. Herbstmann, 1954, White to play and win: 1. Ka3-b2 !!, was most difficult at the time (ca. 1998,) taking Crafty 12.9 running on a Pentium 100 some 34 hours to find (and 50 hours to evaluate) this unique winning move.

Out of curiosity, I did try this position using more modern hardware/software (namely Stockfish running on a 12-year-old iPad 2) and it found the correct move Kb2 at 24 plies in 32 seconds (i.e. 3,800x faster,) evaluating it as +11.9, i.e. clearly seeing it as a winning move. By the time 28 plies are reached the evaluation improves to +19.5.

Other remarkable Sections and exciting particular test positions include:

Bizarre positions

Bizarre Positions is an amazing Suite Extension where you'll find only absolutely weird positions nowhere else to be found ! For us Computer Chess fans, you'll see that some of these positions stress the chess engines to their limits, and even can help discover latent bugs.


74.- "Melee", proposed by Valentin Albillo, 1997, White to play and mate in 8: h7xg8=N !!

75.- "The Ring", proposed by Valentin Albillo, 1998, White to play and mate in 12: 1. f7-f8=Q


Heiner Marxen tried this position on his own special mate-searcher, CHEST, and was able to solve it, demonstrating that it is a mate in 12 (and no less) in some 31 hours.

The Never Concept

The Never Concept refers to a situation or action that a human chess player can perfectly see and understand that it will never happen, yet a computer chess program is absolutely unaware of this and can easily misevaluate the position completely, thus committing serious and even fatal errors, even if searching to extreme depths.

82.- V. Chekhover, 1952, White to play and draw

Unsolved positions

91.- "Mate the Royal Couple" - Proposed by V. Albillo, White to play and mate in 12

I thought (but couldn't prove) it was a mate in 12 but as you'll see in the Addendum 2020, chess fan Vincent Lejeune used ChessMaster 9000 in 2003 to try and solve this position and found that e4 is a mate in 11 !

92.- "Surrounded" - Proposed by V. Albillo, White to play and mate in 12 (?)


Still unsolved, AFAIK.

Chess Book Reviews

Over the decades I've built quite a decent library of chess books and more than 20 years ago I did review some of the best, with an emphasis in (but not limited to) Computer Chess, uploading the reviews to a chess-related website of mine ("Chess Tests"). Each review includes a position taken from the particular book and tackled with the programs of that time period, with comments.

Acknowledgements

Here's a sample from the Acknowledgements Section:

"Robert Hyatt, renowned author of world-class chess programs Cray Blitz and freeware Crafty, whose versions 12.6, 12.7, 12.8, 12.9 and 13.3 are extensively tested on this site, sent me kind e-mails, pointing several interesting things about some of my test positions, as well as listings with Crafty 13.3 analysis for some of them.

You can see them in the Addenda at Test 01, 05, 12, 13, 16, 17, 19, 22, 29, 30, 33 and B03. Also, all the results he submitted are identified in the tables with this marker [HYATT]. Many thanks, Robert."

pier4r Wrote:Further related to that, the program Mater, mentioned in the chess page was also programmed for the 71B (although in a more compact way): https://albillo.hpcalc.org/articles/HP%2...0Mater.pdf

Yes, these are the highlights of the full MATER program, as featured in my chess site:

MATER - A simple Mate Searching Program

MATER is a simple mate searching program, implemented as an MS-DOS command. Given a legal position in FEN notation, the side to move, and the maximum number of moves to give mate, it will search and output either a move which gives mate in that number of moves or less, or the fact that there's no such move.

Main features:

● Full legal move generation, including full legal castling, en passant pawn captures and underpromotions.

● Can find mates in any number of moves, up to 64, subject to available memory and time.

● Can find general mates or mates in which the mating side only gives checks. This is useful for finding much quicker long sequences of checks that end in mate.

● Accepts positions in FEN notation, and checks them for legality and syntax.

● Searches for mates recursively, and thus always finds the shortest posible mate.

● The search can be halted inmediately at any moment, simply pressing any key.

● Outputs the first move that gives the required mate, or the fact that one doesn't exist.

● Shows progress for each iteration, including time taken.

● Outputs final total time, as well as nodes examined.

Documentation and 8 awesome examples

Turbo Pascal source code, fully commented, includes a move generator which allows for castling, en passant pawn captures and underpromotions.

A sample of the Turbo Pascal source code:

PROCEDURE GenerateMoves

(*
generates moves for a given side. Options permit generation of all moves or a single move, full legal moves or pseudo-legal (own king may be left under check), etc.

All chess rules are implemented, including all five rules for legal castling, promotions, underpromotions, and en passant captures
*)

Note to readers: Please post any comments to any of this in a separate thread. Thank you.

Enough. Thanks for your appreciation, pier4r, and for posting about this Section.

Best regards.
V.