'Find the Worst Possible Move' Puzzles

I wrote an intro in my previous comment! Please read it and don't get mad!

Actually, the naturalness does appeal to you or you wouldn't have liked the problem! But since you are (like me) spun into a theoretical cobweb, you had to find a formal way to make it work! But your intuition is much better than your theory - which is the only reason that decent retro problem were composed in the past 100 years. Not the theories.

Yes, in the model I described you couldn't do a problem without castling right! And of course I thought of this. First of all, my problem does contain "no castling right" and in the nicest way possible - after proof, in the main line and in the main try!

More fundamentally, such a problem would be in a higher category of complexity because you can't manifest "no castling right" in the worst move except by implication. You'd need to change the type into "worst move, worst conditions" which constitute a challenge in themselves. For instance, you might have to take into account 50M or 3R issues. But with proper definition it can be done. All the non-default stuff is concentrated in the "worst move" only in the most basic category!

No the naturalness of "Optimism. Worst Move" appeals to me. I am repelled by your weird imaginings that en passant must be allowed, but castling cannot be forbidden. That makes utterly no sense to me

Of course, that isn't true since they are in different phases. It depends on the acceptance of the phase model. See the paragraphs I just added to the previous post with the critical argument.

Maybe contrarian is the right formalism: it's like if you watch a video in YouTube of someone juggling in their home, there must be something stupid about to happen. Maybe the idea of these problems is like that: it's bound to be unintuitive, otherwise someone wouldn't be making a fuss about quick quick come and look at this game position.

There was a related trope/meme in Magic the Gathering called "So I lost, right?" Here's a random example I just found in a tournament report:

"He played one Verdict, three Jaces, two Last Breaths, a Celestial Flare, an Elspeth, two Detention Spheres and a Sphinx's Revelation for four. So I lost right? Nope, I still had four guys on the board and another in my hand when I attacked for lethal. Like I said, sometimes you just get lucky." There is an inverse trope/meme "So I won, right?"

This is a pretty good argument for contrariness. But my intuition tells me it's horrible and wrong to be able to say "e.p. is definitely permitted", but not "castling is definitely not permitted". Please respond to that thought experiment.

Very much not a fan of you editing previous posts in this way, in the middle of a discussion. Accidental ships-passing-in-the-night can cause the occasional lack of serialization in posts. But please don't do it as a matter of course

We should define phase: I think I have a different notion of phase from you. Yours includes retro and forward as separate "phases". Of course I agree that different conventions can apply to retro and forward "periods" of the solution or whatever we call it. The issue was just the semantics of what is a phase.

Anyway, this is irrelevant to my point. The discovery that omigod I can e.p. is just the same as the discovery that omigod my opponent can't castle. They aren't in different "phases" whatever phase means. Either is what would drive the "So I lost, right?" interest of the problem

My basic unit is the proof game. Anything that might have happened in a proof game, might show up in some problem type. So the e.p. move requires at least that it is not provably illegal. 

For the rest I do not approve of e.p. and (non) castling in the same phase. I need not ratify e.p. captures after the first move which has special status resulting from the storyline. Note that the whole thread started with this story line (post #19, first diagram). By the peculiar rules of chess, the issue of e.p. rights doesn't show up in later stages and that is why it is sort of irrelevant how you define the e.p. permission. But conceptually, I would only allow e.p. in the first move phase. The rest is played by default rules/convention and there is no inconsistenvy or weirdness in it.

The idea of the first move is not to allow just e.p. moves but to maximize things one could do or could have done wrong. For instance, when playing with fuddled men I would permit a worst move that places a king in (possible) check by a (possibly) fuddled piece even if a convention exists that would ordinarily prevent you from doing so. Still, after the first move normal play resumes or one would need another layer of information policies to provide for it.

You introduced "optimism" which I think is a logic by "scarcity" (of the orthodox retro-properties). Btw, isn't that what used to be called Ceriani logic? The thought experiment example I gave for retractions also applies to "optimism". Suppose the execution of an e.p. move could be 1 move delayed then I could make an unprovable e.p. capture on the 2nd move of a "worst move" problem under "optimism". Which opens the question on whether that is desirable and where the justification is. But with chess rules as they are, optimism appears to work well. Just as "dominant e.p." but that apparently requires more explaining.

I only read this one after my last post. I understand your issue with my post-edits. I just hate to add out-of-place-comments in new posts with references to older posts. Sorry.

Another thing I edited in an earlier post was the reference to "AP-after-Keym" a problem type in 2 forward phases - one treated by PRA, the other one by AP logic. In a simple sense, a phase technically relates to a change of addressed domain or a change of rules. For instance PRA in the foregoing example is about assessing both sides on move while the AP part is about playing the content moves and achieving the goal. 

My permission for "worst move" e.p. is not intended as a "discovery" but as a rule/convention just as in optimism. That I didn't add it n the original problem I explained to the OP as a call on the character of "worst move" problems without knowing whether or not any contract on it is in place. With the presentation as a trick/joke I tried to avoid having to explain retro-conventions and deviations. Had I written it to you, I would have packaged it differently.

Still, I am in favor of including "worst move e.p. maximalization" in the worst move type as a default since having to add "optimism" kind of gives away the plot. When it is a feature of the genre, then the challenge to see it remains alive. You criticized my proposal also by comparing it to SPRA but that is quite different. SPRA is one of the logic types which we attempt to keep as a small set with a wide range of applications, while "worst move" is a problem type and composers create an abundance of them when fruitful. Having a "worst move type A" and a "worst move type B" is not a big issue. Also, optimism is not a safe logic when expanding to fairy types as it disturbes the operation of standard conventions after the worst move with unexpected collaterals. So we have a different view here.

Obviously, part of my viewpoint is that the "worst move" and "its resolution" are indeed in different phases and we are not going to agree on that. The difference comes from a different appreciation of the "worst move" story line. Story lines in retro-chess are required because all we create is based on a non existing reality. Diagrams have no actual history beyond that I entered them from the diagram editor. Still we pretend there is (at least) one relevant reference game. The story line justifies our mapping of a diagram onto a believable history. And gives us all the retro logics. They are not purely deductive. They depend on our story of reality. The story line I suggested for "worst move" sets this move apart from the remaining moves but if I have a good story to set moves 1,3, 5... apart from moves 2,4,6... then it is totally OK to create an alternation of 2 phases with different "rules". Story lines are responsible for structure more than "similar looking move series" like in a forward solution. Note for instance that the appreciation of solution moves - good or bad - and problem outcomes are only in the story lines and nowhere in the chess rules; these only terminate with a status word like "mate" or "stalemate"; the valuations of those states in compositions and the conditions are in the story lines associated with the stipulations. Many mathematicians believe that the Gödel sentence is really true (look in the forums). It isn't. It's only true in an appropriate model. And the model is outside the system and based on a story line. A different story line and the Gödel sentence is false. Just as in chess.

Note: this is all apart from the existing reality in the composition field. If there is a consensus about "worst move" rules and there are accepted methods for specifying retro-states (like optimism) then we should of course go with the flow for publication. Can't fight all wars all the time.

For Problem #6, yes

I know that SPRA is a “logic type” and worst move is a stipulation. If we are separating out S (aka O) from PRA then that means we have also a logic type ORS. Your problem is a nice example of ORS and under the (flawed) RS+PRA convention we can just have O in the stipulation. You have a problem type which you want to define a logic type for, and here we have multiple disagreements. Proof games don’t require any conventions since you are driving forward from the game array so I don’t see the relevance. If they are A to B then I guess the default conventions would apply, although I have seen any of this form: an opportunity for composition there

Where I am really at sea is that you haven’t defined “phase”. Please do so, mathematically if possible, but certainly precisely. It seems to be some tagging structure on moves. I am still completely not getting why (in some manner obvious to you) “phases” mean that castling must be handled in a different manner to e.p. Your intuitive model of Alekhine 1933 absolutely leads to the contrarian “mode” (if that’s what we call O,P,D&C).

It’s certainly a question that declaring the mode may make the problem too simple, but that’s no excuse for resorting to an unnecessary joke when the slot in the extended standard model is waiting for it

I must first repeat that I would never have presented the problem this way in a formal context (I didn't e.g. expect that you would read it) but placed it here as a sneaky comment on the foregoing thread. This was not a retro-thread so it's kind of desirable to avoid a discussion on that point - which now obviously failed!

Which means that I agree it needs reformatting in a formal presentation. My feeling is though that the story line for this problem type could be different from the way you see it. The story line precedes all notions of retro-logics and phases. I guarantee you that a whole bunch of people (not including you) would be willing to play e.p. for a worst move "after all you were looking for the really worst move ..." who wouldn't ordinarily do so in a common problem. All the formalisms only exist to fix our story lines into executable algorithms. The ones that do so best for the greatest number of story lines are the best formalisms.

The answer to your phase question is of the same nature. A phase is not something that is, but something you design on the basis of a story line (which is often an archetypal concept). Because it thereafter has an individual identity it comes with its own handling instructions which we label as "rules" or "conventions" or "action types" applied to "object domains". You recognize them by the actions, rules and domains but they are defined by the arbitrary breakdown of the arbitrary story line into manageable entities. Phases exist for any activity, not just moves. For instance when you retro-analyze a diagram you start with a phase where you count the material. You draw conclusions and proceed with analyzing individual unit trajectories which may be perceived as a different phase (you can put away your calculator). An endgame lover definitely separates "introduction", "thematic main line", "try lines" and "transformations" (character changes during solution) and so on as different phases though they are all just moves. It depends on what you need the phases for.

The "worst" thing that could happen to "Alekhine 1933" is that someone would recognize it as a new story line for compositions (not for games of course) and wonder if a formal definition would be required and workable. Which might lead to a "no, don't do it", "yes, we can define a problem type for it", "yes, but we really need a new logic", "only after we remove some disturbing rules like 50M and 3R" etc.. It's a project. I only used it as an illustration of a "different story line" for justifying move choices by placing them in the past.

Nothing is about the math as it is trivially simple in our subject (most of the time), It's all about the handling of concepts.

 

OH

What a bump

Hello

I agree with your choice

https://www.chess.com/forum/view/more-puzzles/find-the-worst-possible-move-puzzles?page=1#comment-51439828

what about Qb3 for your puzzle? or Qh3, or Qxe2, or Bb5? There's a lot of losing moves to choose from.

In post #27, EnCroissantCheckmate (nice name btw) asked why castling is illegal if en passant is legal. How is simple- what was the last thing white played, if black played b7-b5? If you put the pawn on b5, you see that black’s pawns and knight box in the rook on a7- and it could not have been the last piece that white moved. If it were, it would have left an empty square it could return to. Obviously, the pawn on a5 can’t be the last piece he moved, because there is a rook right behind it- same logic as rook, if it moved there would be an empty square. The pawn on c5 could not have moved, because if it was on c4, it would have been checking, so it would have had to be black’s move. Also, both pawns could not have captured, because there are not blank squares diagonally behind them either. Therefore, he had to have moved one of his either pieces- and if the king or rook has been moved, he can’t castle.

In number 10s puzzle, why can blacks pawn not retake? Am I blind?