But on Yet Another Chess Chess Problem Database, it doesn’t seem to have anyway to list or lookup retrograde chess problems. Looking at the database page, it seems to just list it as a mate in one. Plus, looking at the Solution page with the other link, it say’s white’s last move was castling, but looking at the diagram, it doesn’t seem like it’s even necessary for white to have castle or to take back the castling for white to legally take black’s pawn en passant. White wouldn’t be able to castle if it was in check, so to me it only seems necessary to determine that black’s last move was pawn g7 to g5.
Seems to me the only reason white has to have castled in the problem is so that the problem can be presented as a retrograde or “Proca Retractor” problem. Since Yet Another Chess Problem Database doesn’t seem to have a way to look up those types of problems, it might make coming across a retrograde problem in that database as unsolvable or much easier to solve since a person coming across a retrograde problem wouldn’t know to take back moves.
Anybody want to chime in? The question is really about using the online database. Am I missing something in using it, or is Yet Another Chess Problem Database just have a blind eye to Proca Retractor problems?
If white’s last move was not 0-0-0 then black’s last move may have been with the rook and e.p. would not be possible. It also requires the second black bishop to have been created on c1 and the rook to have reached f2 prior to exf3, meaning that the white pawn was already on f4 and thus could not have been a recent move that would have allowed exf4+ Kh4, 0-0-0 g6-g5. It also means that cxd3 was played prior to bxc3 (so the pawn could promote on c1 and the bishop could leave without capturing a pawn.
I’m trying to figure out what black’s move was prior to 0-0-0 since every piece other than the g-pawn would have been nailed to their squares to keep the position legal (having the last moves be exf4+ Kh4, 0-0-0 would have required five pawn captures and the black a-pawn wouldn’t have been available to be captured).
All I can think of is something on the order of:
Bh7 Nc6, Bg8 Ne5, Bh7 Nf3+, exf3 g6, 0-0-0 g5 (no e.p.). The sequence near the h1 corner would have been:
white cxd3 black c2, white bxc3 Ba3-c5-e3.
white rook out and white h5, white Nh8 and Bh7-g8, white Nh1, black Bc4-g1, black Re3-f3-f2, white Qh2, black Re3-h3, black Bh4-g3, black Kf5-g4, white Bg5, black h6 and hxg5 , white Rh6, black Kh4 and g4, white’s bishop and black’s knight marking time until the above sequence. Seeing as the composition won a first prize I’m sure I’m missing something, but I can’t figure out what.
I wasn’t really asking about that particular problem, but rather trying to figure out if it’s possible to filter Yet Another Chess Problem Database for retrograde problems.
I wasn’t familiar with retrograde problems and was trying learn about them. That was a sample problem for a retrograde problem. I was curious if the problem was listed on Yet Another Chess Problem Database, and it was. But when I looked at it on the database website, it didn’t seem like it was marked as a retrograde problem, just a mate in 1 from the starting position.
I tried looking up retrograde problems on the database, but there didn’t seem like there was any function on the website to allow that type of filtering.
First let’s get a diagram in here so everybody can see what we’re talking about:
JW, please feel free to edit this diagram into your post above.
Well, you certainly got a lot farther than I did.
One thing you may have missed (or at least glossed over slightly) is that, in proving that e.p. is legal, you must prove not only that black’s g-pawn was the last piece moved, but also that it came from g7 rather than from g6. The proof of this appears to be that, if the pawn had been on g6, white would have had no previous legal move (the g6 pawn would have prevented white’s Nh8, Bh7, and Rh6 from having just moved) other than O-O-O – or something along those lines.
The B, N and R could have reached the h file prior to g7-g6. I’m not merely trying to figure out how to prove that g7-g5 happened, I’m trying to figure out what black’s move was before 0-0-0 if it wasn’t g7-g6. I’m also trying to figure out why the sequence couldn’t have been:
Bh7 Nc6
Bg8 Ne5
Bh7 Nf3+
exf3 g6
0-0-0 g5 (no e.p.).
Sounds like there is a lot more going on than I realized. It probably isn’t the best retrograde problem to highlight for someone that might be curious about, and not very knowledgeable, on the subject of retrograde chess problems.
I’ve dabbled in chess compositions before. My interests waxes and wains. I’m trying to stoke my interest more, but there doesn’t really seem to be much out there for programs or even websites that tackle the issue. A while back, there was an app on google play just called “chess tactics” that had pretty much a database of most of the compositions out there to solve. That program stopped appearing on Google Play years ago, and the interface was exceedingly primitive for such a grand endeavor of having 130,000+ chess compositions. -You had 2 chances to get the problem correct or it spit out the answer and then moved to a new problem. No chance of looking up the last problem you did, nor did the program have anything, in the “expert” section, other than the 130,000+ chess compositions randomly spit out. The compositions weren’t indexed or searchable either.
I’m guessing someone was able to get the data set from one of the online chess problem databases, stripped out anything other than the basic compositions. That is no retrograde, helpmate, self-help mate, or fairy problems. Then stuck it in the Chess Tactics program with all the header information stripped out. What I did was plug in a few of the problems into Just Another Chess Problem Database, and all the setups I input into the database were in that. After testing a few of them, I figured all the problems were probably in that database, or some other free online database.
I’m curious about compositions because most of the hard or expert level compositions are really works of art. Even if I can’t get most of really hard ones, when I look up the answer, the solutions are often quite elegant and beautiful. And when I do solve them, I get a real sense of accomplishment.
There’s an app called Jopo’s Chess Puzzle on Google Play I’m going to try out. I doesn’t have a trial version, which is rare, but looks to be what I’m looking for. Been a while since I’d looked on Google Play for chess puzzles, but this is the first app that even remotely looks like what I’m wanting. There is a chess problem reader for Yet Another Chess Problem Database, but I couldn’t find anyway to mass download files off the Yet Another Chess Problem Database website.
I don’t know if chess compositions is what I’d really consider as training material for playing chess. At least, not if your doing the expert level puzzles. But if one can appreciate the beauty of chess compositions, then it’s a fun past time.
I went ahead and bit the bullet and took a chance. For five bucks, I figured it was worth the risk. It’s pretty simple design, only marginally better than the Chess Tactics I used years ago. It’s pretty similar to the Chess Tactics in that it serves up random problems, but at least it doesn’t just serve up the answer and go to the next problem if you get it wrong. If you close it out and launch it again, a new random problem is served. It has several sections. The one I’m most interested in of course is the direct mates, but it has tactics and stuff for lower level and higher level players, endgame stuff, traditional mate in X problems, blah, blah, blah. Twelve sections all together.
Considering the price range, I’m satisfied, but maybe it will get updates for more functionality someday. As it is, the functionality is pretty basic, mostly just serving up random problems in each section.
With that constraint, what was black’s last move? Neither bishop, no pawn and one rook have any square they could have moved from. If the other rook moved then it came from f1 or e2 and would have meant that white was illegally in check before black moved. That leaves just the black king to have moved from g5. That means white last move was an exf4+ capture. If that pawn started on e2 then the f3 pawn came from f2 and there was no way for the black rook to get to f2 without having checked the white king from f1. So the exf4 pawn must have originally come from f2 (after fxe3 the f2 square can be occupied by the black rook and then subsequently exf3 and exf4 would put the pawns where they are). That requires five white pawn captures (c3, d3, e3, f3 and f4). There are six black pieces missing (Q, N, N, light squared bishop, 2 pawns) plus one more pawn that promoted on c2 to a dark squared bishop. There is one white piece missing that was captured by the black h-pawn, leaving no way for the black a-pawn to be captured on one of the five capture squares.
The black e-pawn was captured on e3. The black bishop was captured on d3 or f3. The queen and two knights were captured on c3, d3/f3 and f4.
There is one little problem with that. The black light-squared bishop never moved (pawns are still on b7 and d7 and was thus unavailable to be captured by a pawn. That means the only way to get the five captures by white’s pawns is to include the black a-pawn in the captures. The only way to do that is for the black a-pawn to have captured a piece, and the only white piece missing was captured by black’s h-pawn to double up on the g-file. With the black king eliminated from being the piece that moved before 0-0-0 that leaves g7-g6 as the only remaining move and that means g7-g5 was not played following castling and e.p. is not a valid move.
The retrograde analysis mate-in-one problem becomes a retrograde analysis mate-in-two problem (Ng6+ and hxg6#).
I finally looked at the solution and it changes the problem. It says that 0-0-0 gets retracted and hxg6e.p.# is the solution (g7-g5 was played prior to 0-0-0, not after). The link on the problem did not mention that.
It did point out that the h7 bishop had to get out before exf3 and that the black e-pawn could not be captured by a white pawn, so the c3, d3 and f3 captures had to be captures of the Q, N and N, drastically reducing the possible moves available after exf3.
The first link had a solution with only a minimal explanation, and did not state that the problem included retracting the 0-0-0 move. Since I already had the solution from that page (or at least I thought I did) I didn’t bother with the other page.
You only half read the original post. The entire post was specifically about Proca Retractor problems, and not being able to filter for those problems in Yet Another Chess Problem Database.
The original post specifically mentions that the only reason long castling exists is to make it a Proca Retractor problem. That’s the crux of the post:
Yet Another Chess Problem Database can’t filter for Proca Retractor problems, and does not correctly define that a problem needs to have moves retracted
Proca Retractor problems just come across as being listed as either a very easy direct mate, or a problem with an answer that makes no sense.
