'Find the Worst Possible Move' Puzzles

Worries me I found these so easy

Most of them are really easy, but not all of them

In puzzle 9 a bishop on c1 is missing. Diagram 10 is illegal which makes it even less likely to appear in a game than RewanDemontay's diagram

Puzzle 9: Fixed

Puzzle 10: It is funny how Puzzle 10 is illegal but puzzle 8 and @RewanDemontay's puzzle are legal

Puzzle 10 appears to be easily fixed with a knight on h3.

Here's a tricky one. Worst possible move for white? As in fastest loss?

Great puzzle, but why is castling illegal after the en passant move?

Last black move must have been b7-b5. What played white the move before?

True. Castling is legal

Also, this seems like a joke puzzle as you never said that en passant was legal

Yes!! That's why I said it was tricky. But it is kind of fit for the genre "worst move" which inverts standard chess logic and thereby invites inverted assumptions. But it's definitely borderline! Like me.

Btw, there are quite a lot of ways in retrograde chess to define how to fill in unknown conditions like castling and e.p. rights (and others). Since they are not simple to relay to chess .com puzzlers you can look at them as "jokes" and use your intuition to figure them out.

The one I used is PbyD (prove it by doing it). You prove your right to castling or e.p. simply by playing it in some variation. It is sort of appropriate in problem types pretending to "look at what happened to me in a game" (see your own post #19). Obviously an e.p. capture could have been valid in a game even when you can't prove that from the diagram!

I object to the use of non-standard conventions in standard problem types but not in jokes or problem types where these conventions appear over-restrictive. For instance, there are problem types where you must take back a move and provide a different one. In all of those types you may legally take back an e.p. move without need to prove it. They ue PbyD logic by default just like I did in #26.

This is really cool. It's already been posted in PDB as https://pdb.dieschwalbe.de/search.jsp#P1378929. I have commented on it to align it with existing conventions. I am not a fan of the term PbyD which was invented by Ronald Turnbull, because it doesn't align helpfully with other notation. I would call this Optimistic RS. So we can have Optimistic/[Default]/Pessimistic PRA/RS. The qualifier is separable so we can stipulate a problem (e.g. this one) being simply "Optimistic" and allow the PRA/RS question to work itself out separately.

We can say that retractors are by definition Optimistic RS. Conventional Retros are by definition Pessimistic RS. Default is optimism for castling and pessimism for en passant. This extends to any other fairy uncertainty (e.g. Fuddled Men). It does *not* cover last move, which is so important that it is handled by a separate convention.

"Worst move?" is also a known stipulation in adversarial (= non-help) context. So bottom line, I think with the stipulation "Optimistic. Worst move?" this problem might be considered totally orthodox in a future world.

Ah, we are in the preamble to our "legal move" discussion! Didn't think you would pick up on it here! Actually I used the term PbyD because if was easiest to communicate to a non-expert. "Dominant e.p." would for instance do as well.

But more important is to recognize the different phases. Finding the worst move is a different phase with different logic (just as would apply to a retraction) while normal conventions are resumed for the remainder of the solution. I'll revisit that when discussing "legal moves" and "logical phases".

We've been in accord too much recently, fortunately there's much to disagree with here

“Worst move?” is just a forward aim like "s#5", or "Mate in how many moves?" and can best be handled like that. There’s no reason to expect it to differ from other aims in handling of forward play. Yes it may require some elaboration, but then so does other orthodox stuff like #n with the ridiculous contortions to self-justify vast swamps of dualization in a problem one is otherwise too fond of

You were complaining about SPRA but you are making a similar kind of conflation error by insisting that "worst move" somehow has its own retro conventions. It's just a rare forward aim.

I don't think forward and retro are "phases" (which to me is main line variants, set play and tries, and possibly forward play *subsequent* to retractions). Conventions would be constant across phases. If you want conventions to vary, set up twins. There isn’t a word for retro & forward "modes" afaik but we can invent one: maybe “direction”. But the conventions apply differently according to the stipulation type: retractor and last move are different for fairly obvious reasons. That leads to three main families which are normally handled as optimistic, default and pessimistic for meta-convention purposes. One might add a fourth: "contrarian", for completeness purposes, which is the inverse of "default" (analogous to contrarian investor).

Dominant/recessive is not the happiest metaphor since the biological sense is almost always that you’ve got exactly two genes in competition. Also there is the genotype/phenotype concept for repeating position which is such a hard concept to get over that I prefer to play the genetic metaphor card there. So I rationally prefer optimistic/pessimistic at the level of an individual move right. I’m also over-loading these terms to talk about the overall perspective but could use e.g. Panglossian/Eeyorish instead for these global terms, however I don’t think new words would add anything.

The thing is, your problem is very nearly orthodox if we think clearly. Certainly the idea shouldn’t be wasted by being positioned as a joke, which introduces vagueness & doubt in even the staunchest solver's mind. “Optimistic. Worst move?” describes it exactly. One day this fine problem will be canon if I have my way

What is this "almost impossible" notion? Why are you saying "theoretically" like it's a bad thing, and "realistically" like it's a good thing?
I think you meant:
"It can!"

Chess problem composers have avoided aiming for "realistic" positions for about 150 years. E.g. adding pieces just to make the forces look more balanced is called "dressing the board", and is an absolute no-no.

Thanks for taking the time to post these actual positions though - they are fun.

...

@anselan:

Intro: rereading your message only adds to the confusion. You state that "convention should be consistent across the phases". But also that they are different for obvious reasons by the stipulation type such as for retractors and last move problems. I'm afraid I won't get that in a 1000 years. It's the stipulation type that creates the phases. How can the conventions then be both consistent and different across the phases? You will properly reply that "different" is not the same as "inconsistent" but that won't work. Since we are dealing with the same retro-properties (e.g. e.p. right in the diagram) the different handling of them is automatically contradictory unless you absolutely identify the phase it pertains to. In fact, I do not understand a single word of what you write on phases and conventions so I'll just ignore it in my further reply. Just write my own view. 

Just sticking to 1 issue. There is no necessity to see the worst move type as one for special conventions but a possibility - which I explored here. Note that retractions are always under a different logic than the forward play. It's not even a choice, it's mandatory. Otherwise you couldn't have different logic assignments (RS, PRA, AP) for problems with a retraction - none of which permit arbitrary e.p. moves. By chance, the ensuing issues do not quickly show up in orthodox chess because of the scarcity phenomenon and the general unclarity of the information status of retracted moves. But if you follow this thought experiment things become clear:

Suppose as a rule the execution of an e.p. move could be 1 move delayed. I retract the e.p. move b5xa6 and replace it with another move. Could that move now be gxh6 e.p. because blacks pre-preceding move might have been h7-h5?

The answer is of course "no" since the unconditional retraction of an e.p. move doesn't change the convention that forward e.p. moves may only be played if provable. Just to show that the principle of "one problem, one logic" is an illusion. Don't tell that to the chinese government. That is also why pure retro-analysis is not permiited to rely on any of our standard conventions. They go out the window when searching for provable truths.

It is possible to tie problem types to different logics or different combinations of logic types. Note for instance that what used to be named "AP after Keym" is a combination of PRA and AP-logic - different logics for different phases  It's not weird or inconsistent since the logics only reflect justifiable ways in dealing with certain information gaps dependent on the character of the problem type. So the "worst move' type could be standardly associated with retraction logic for the "worst move" - or not - or for worst move type A and not for worst move type B. Like I indicated in my earlier messages it all depends on the story line. There is a critical difference between the following stipulations: "Checkmate in 2 moves" and "Alekhine played a game in 1933 where he checkmated his opponent in 2 moves though his oppopent played the best defense; how the heck did he do that?". The 2nd phrasing not just takes the question out of the composition domain but it opens the gate to the assumption that any game condition might have existed. So why not play e.p. especially when nothing else works? "Worst move" could be framed in a similar way. "What is the worst move a player presumably played in a game that would lead to the shortest mate you would find in a problem?" This creates a "game history" phase and a "problem" phase. Note that my mind did this automatically when I first saw this type. It feels sort of natural to look at worst move type problems in this way.

That "naturalness" doesn't appear to me. Frankly, I think this is your idea so it's a bit precious to you But honestly if it was anyone else who'd had that idea, especially on the list of people you disapprove of, you would be scorning it. Why should "worst move" which is a cool idea be enmeshed for eternity with the notion of e.p. optimism?

Here's the key thought experiment: suppose I wanted to do a "worst move", which has no e.p. but hinges upon someone *not* being able to castle, although it looks as if they can. Perfectly reasonable scenario from the Alekhine 1933 perspective. Under your convention, it would be impossible to create this problem. This is why there should be no special relationship between this stipulation and the optimistic/default/pessimistic/contrarian variable. Except default is the default, for simplicity

worst move is to resign

Heh that's true over the board - problems don't allow resignation, or disqualification or agreeing draws.