Dead reckoning puzzles

Andrew, I think your post gives all the required supplemental info with regard to the origin of the compositions. I agree that stipulations such as you listed categorize the challenges as "retro". I will not change my comments to #20 because it simply cites the earlier posted diagram in #14 for readability. My comment was directed towards that posted puzzle, not to a (different) original, and is still valid as such. Chess.com members are pretty sloppy as to credits and precise stipulations of the puzzles they post and I do not see it as my mission to track and trace the meta-data of the puzzles. I comment on what they post in order to correct misunderstandings which might arise from reading their posts by others. By the way I didn't rate #14 (which has no stipulation whatsoever) as unsound but as meaningless in a solving context because there is a double argument for castling. That doesn't render it meaningless in a high brow retro environment which addresses DP. By the way, I do not share your love for legality over permittability and I will explain that in our "legality discussion".

Our discussion on the "sillyness" of the meta-choice conventions will continue for a long time I suspect. Your acid test is invalid for the simple reason that the convention which requires you to deliver a "legal" diagram applies to all orthodox composition types. All compositions are retro and require low level handling of retro-uncertainties such as "can white be on move?". What you want all composers (not just the retro specialists) to see is whether they are just "retro" or really "retro!" in a particular challenge. In my discussion with Rocky64 (which you supposedly read) you can precisely read the what and why of my objections. There is a second argument which blows the acid test apart. The idea of having the "same" solutions with or without a history overlooks the fact that these 2 solutions must be evaluated by different rules - because one requires DP and 50M in force and the other one discards them. Someone solving post #7 in this thread would not get 2 different solutions from merely excluding or including the game history because he would think it's an endgame with endgame rules. It's the ultimate example of the self-fullfilling prophecy problem: one solver sees an interesting retro DP problem, another one a boring endgame study and never the twain shall meet (presumably these 2 are seen as different solutions by analytical content, not by score!). Of course that assumes the puzzle as posted and not the original one but such doesn't make the case less relevant. There are more issues with the "same solutions" to identify the retro-type which I may already have addressed in the "Rocky64 papers"

Note: In my email a few weeks ago I wrote that I do not advocate the somewhat clumsy "retro" text in the stipulation but rather the concise "RS" in the context of retro-strategy as a logical default - and only if necessary. I could not simply plunge my preferred notations in the chess.com environment without elaborate and off-topic explanations.

Different question: Are there any assumptions for these compositions for who is on move? Is that part of the stipulation or is it borrowed from the conventions? The way I recall it from long ago is that purely retro-analytical problems (no forward stipulation) do not consult the conventions for missing information. Did anything change?

Note: I changed the stip for #23 to "white wins". It doesn't expose the e.p. move as much as in the #2 though the solution moves are the same! But you are free to repost it with #2. The original publication here is not protected by chess.com. Chances are the post mysteriously disappears in a year or two. I lost quite a number of compositions and I can't even remember precisely which.

Are you alright? You seemed a bit off today.

I want to bring one key remark of yours to the top:

> One solver sees an interesting retro DP problem, another one a boring endgame study, and never the twain shall meet.
Yes I have some other unpublished compositions where you can only see the retro play if DP is in fact switched on. But 17A was meant to protect the innocent endgamers. It was not meant (I hope) as a scourge for retro composers. So if we switch DP on experimentally, and find retro content, then we've broken the vicious circle. Chicken & egg: one did come first.

And now for the rest :-)

> I will not change my comments to #20 because it simply cites the earlier posted diagram in #14 for readability.
#20 cites #7 (not #14) and comments on it extensively, including shouty bold. It's all based on a false premise. Now you know the facts, I would appreciate if you fix it, but it's up to you.

> "solving context"
This whole thread is about DP. That's the context here. When you talk instead Roger Irrelevantly about "solving context", you are just being contrary I think... :)

I think it's quite cool and unexpected that if these two conventions are cleaned up to remove redundancy then they harmonize beautifully with DP rule. Sad you haven't responded on that.

> By the way, I do not share your love for legality over permittability and I will explain that in our "legality discussion".
What I said here was "If it cannot be proved whether the relevant castling right remains, castling is permitted." How is this not liking permittability? (Maybe you are talking about the "two levels of legality", where in my writing I need to balance the valid concerns of purists like yourself, with the equally valid but I suspect unreconcilable concerns of pragmatists like Francois Labelle, sigh.)

I am sure that the acid test can't quite work in all cases.

>All compositions are retro because legality check
It is clearly insufficient in almost all cases to make a problem "retro". So the acid test works fine.

>"which you supposedly read"
Haha you cheeky fellow! Bless you! :) Honestly I can't remember: there's so much.

> The fact that these 2 solutions must be evaluated by different rules - because one requires DP and 50M in force
I would allow those forward rules to be in effect in both cases in order to do this particular thought experiment about retro. Or instead switch them off in both cases? But have to compare like with like. So again the acid test survives.

> "What you want all composers (not just the retro specialists) to see..."
No I'm not. C.f.: https://www.youtube.com/watch?v=nS9W-wlJHPA "So what you are saying is..." :-)

> Note: In my email a few weeks ago I wrote that I do not advocate the somewhat clumsy "retro" text in the stipulation but rather the concise "RS" in the context of retro-strategy as a logical default - and only if necessary.
I am not advocating sticking "retro" or "RS" in the text here, but notating if DP would apply in a pure forward stipulation. No retro.

OK, following your idea because it's nearly the end, I hope that I never have to say in a problem stipulation. "This is a retro" (Unless it obviously can't be... that would be so cool... "Ce n'est pas un pipe." Adrian Storisteanu would love it.)

If it's an orthodox d#, s#, h# then there is a convention (Art 15) for who has the move, and what to do when that's not possible retroanalytically. But your memory is correct that there's no convention for pure retros, and nothing has changed. Even for Type B & C Last move retros. there is no default for who moved last. Maybe for retractors there is, but I don't know much about those yet (maybe some day)

Great I will post it in PDB as #2 thanks - as "White wins" suggests a study. I am honoured to share another composition with you

I'm apparently a bit off. I thought I only cited one of the puzzle diagrams in my own posts but I did so on more. Hence the post number confusions.

You seem to assume that the only reality in existence are the original puzzles. Not so. The people who would read the thread have no knowledge of the originals and try to understand what was posted (not by me, by others). I commented on the reality of those posts and I am in this context completely uninterested in the original puzzles. The puzzles which are here could be solved and evaluated as they are. You seem to have the idea that the existence of original sources somehow should prevent people from discussing perverted versions of them. This discussion is relevant though because the posted puzzles have interesting points themselves, notably on the retro/endgame/problem meta-choice. I am not selling DP-puzzles, I comment on retro-issues.

I could of course go back and discuss the original puzzles as well but they are much less interesting to me than the perverted ones precisely because the perversions are worthy of inspection. So I'll leave it to the interested readers to follow the leads in your post and enjoy the unaltered sources of the DP-puzzles!

That you fell over the boldface text shows you didn't understand the context of my comments. The original posters had left years ago and I did not attempt to blow anyone away in a discussion. I grabbed the opportunity to write my comment in an educational fashion by highlighting the essential parts for later visitors. There is quite a lot of flat text to read without undulations.

Many posters on chess.com simply copy and paste puzzles on a keyword such as "dead reckoning". I don't rate them all highly - and quite correctly as you have seen by the perversions. Most puzzlers here are on the level of "seeing an effect" and expect that effect in the solution of their posted puzzle. In the particular (perverted) DP castling puzzle that effect was absent and I commented on that when referring to "solving context". Obviously those comments do not apply to the original version which was not posted here - as with the other altered puzzles.

My objections to article 17A are completely based on problem portability (as I explained to Rocky64 in our long discussion in his introduction to chess compositions). Hybrid retro-problems are drag queens. They disguise as endgames and orthodox problem types while secretly carrying the retro-virus with changed rules. These problems circulate around the planet (hopefully without perversions) and arrive at all kinds of chess people such as problemists and endgame lovers. How are they to recognize a DP-puzzle in Qi1's #7? I have stated exactly the same about your endgame/retro puzzles though they have more uncontestable retro-content than #7 and are therefore easier to identify. Note that #23 is worst of all. Someone used to solving direct mate problems or endgames will immediately reject the e.p. option and find no retro-content that would ever incite him to look up the retro-section of the codex and include DP. He shrugs his shoulders and goes home. That's why I am dissatisfied with article 17A - not for what it means to retro-analysts.

By the way, if you propose to explicitly specify the use of DP - I am not sure I understand your comments on that - then my objections melt as ice in the sun. Almost any way which makes this information available to the solver is better than hiding it in the dark.

My comment on your love for legality over permittability was based on another post/message/email but my head is spinning to much to find it. Not necessarily about these problems. Of course I know you accept "permittability".

All my comments about the acid test are again directed to the portability condition. Random problem solvers are to know about your acid test and do your acid test and use the adjusted rules en draw the proper conclusions before they can solve the retro-virus infected challenge. I am also sure I can generate a gradient scale of retro conditions where retro-solvers will all find different dividing lines between endgame and retro. The only thing that saves this weird concept is inverted solving. They will draw the line where it appears to work to get the desirable solution to a particular problem. It's the joke domain. You can reread this analysis in the Rocky-papers.

I wrote somewhere "what you want all composers ... to see. ....". That was a typo or thinko. I intended "solvers", not "composers". I suppose you do not want to deny that all solvers must figure out what type of problem they are solving in order to discover some of the rules. 

Note that revealing only the presence of the DP rule would leave undisclosed the status of the 50M rule. It is equally volatile and likely to cause issues in some problems. That is why I see a role for "retro". There are also the yet undiscussed issues of RA-compositions without conventions. The "RS" spec (better than "retro") could be useful to separate the drag queen retro's from the purely retro-analytical type when such is unclear. Note for instance that "white starts" is not in your stipulation for the original of #7 which makes me wonder if the composer perhaps wanted it presented as an RS-type with a default for the starting side.

Finally, I am extremely sceptical about all meta-choice articles such as 17A. They are subjective and block all future chances of algorithmic implementation. These and other issues guarantee us just one future: trouble, trouble, trouble.

@anselan: in case you read my last post earlier please revisit it. I clarified some matters this morning!

Sorry sorry sorry, I don't have time to address all of your very long post, but I wanted to kind of sign off on this topic for now by picking up one recurrent theme, which I agree with. That "retro" is not defined really kicked the ball into the long grass.

The meta-choice articles include 16.3 (PRAvRS), 17 (50M) & leastly 17A (DP). Although they may be practical for 90% of orthodox compositions, it's in our natures to examine the boundary cases.

Your vision is to replace all of this with an all-embracing model. My vision is more limited, to find a way of operating within the existing frame, but fixing the grosser problems. I am sure that political revolutionaries have struggled with an analogous dialectic: abrupt vs gradual change since Karl Marx got his membership ticket for the British Library in 1867.

I do feel that the notion of Cursed Problems (which I've mentioned before) is essential and fascinating, and is underpinned by a requirement to identify our stakeholders. That's kind of where my journey might take me next.

One recent gradualist victory over the status quo is: http://matplus.net/start.php?px=1593849749&app=forum&act=posts&fid=gen&tid=2504&pid=19306#n19306. The fact that it's taken so long for such trivial changes is indicative, but of exactly what I'm not sure.

Ah, great! I'll read the revised articles later. I have encountered the term "cursed problem" and I suppose it was in the context of the golden age classification which I didn't really dive into.

Note that my approach to meta-data for compositions is not fixed. I stated to Rocky that missing meta-data may be provided on any level that is readily identifiable and available to the solver, e.g. in rules, conventions, stipulations, captions, footnotes, reference links, publication context and so on.

The reason that this has become an issue in the past decades is because compositions which used to be published in a few fixed places (like problemist magazines and newspapers) in known frameworks for specific audiences have now started to roam the world through social media and the web in general. Therefore they must be made portable and carry the relevant meta-data inside the portable packages! Though of course, you'd have to assume the accessibility of public digital resources. We can't include the general rules and conventions in every package.

Note: I am not referring to existing digital formats like FEN and PGN but to a culture of specifying meta-data in the awareness that particular publication contexts will be lost when the compositions start roaming. How that would look in a digital format is an implementation issue, I only addresed it informationwise.

Absolutely! It's my motivation for striving for maximal formalization. When there is a border, chess composers will visit it. It's the job of rulemakers to assure the boundaries are clearly marked when they arrive!

Note that in the older versions of the FIDE laws (I verified it for 1997), article 5 referenced article 9 on multiple occasions entangling basic with competition laws. Because the basic rules of play have now (2017/2018) been absolutely separated from the competition rules, the explicit inclusion of relevant article 9 paragraphs is required - and they are indeed explicitly identified in footnote 12 of the codex as FIDE article 9.2 and 9.3. I don't think explicit inclusion was required for the 1997 handbook since the laws linked by article 5 to complete its operation can never be excluded. Which means that "positions" were always defined!

Note that there are now more issues with the repetition and 50M/75M rules/conventions as was predictable. They need to be fixed one day. Also there appears to be an arbitrary distribution of termination rules over article 5 and article 9 - from a composition perspective. Things would be simpler if the articles 5.1.2 and 5.2.3 were declared invalid and explicitly excluded from codex implementation. Not by diagnosing them as irrelevant since they are not. They are unnecessary and (may) destroy compositions.

Thanks for the clarity on the "positions" status. My next major goal is to resolve the 50M rules. Thanks for calling out 5.1.2 & 5.2.3 - I agree that these need to be "nobbled". My idea here, since they are a bit trivial by themselves, is to loop them in with the decision parts of 50M & 3Rep. Yes this means that 50M splits into two aspects: move history + automating the decision to claim. This also allows us to dispense cleanly with clocks, move sheet, timing. Similarly for 3Rep.

The simpler ways the game can end are listed in 5, while those that require more machinery are classified in 9, I think.

The idea was probably that tournament policies might affect the situations in article 9. From the perspective of composers that distinction is useless. There is only one tournament in the universe which is the composition tournament .

It's also worth considering to disable all competition rules and replace them in the Codex. Unfortunately some of the "human intervention and arbitration" stuff is also in article 5 which we need to prune as well. Article 4 is already redundant in digital interfaces and no less in compositions. It can always be reinstated for special composition types like Osorio's MDR (minimum deviations of the rules) which is based on it and for joke problems.

I predict the handling of 50M is tricky, even if they make you the master of ceremonies with unlimited powers. The challenge is to retain as much composition stuff as possible without golden age rule. Possibly you will have to permit more than 1 flavor.

Yes: there are problems based on Article 4, so it seems kind to keep it. Not going to hurt anyone.

I have zero ambition to be master of ceremonies. I don't even want to be on the Codex Committee. I just want to be a petitioner. I agree 50M may be a challenge though

Got it!

Since this is a DP-thread and since the Codex has now lined up with the 2018 FIDE laws, there is 1 "new" issue which begs to be resolved.

We discussed it before in terms of the conventions but I don't raise it there - only for the FIDE laws which cover its actual use in chess games. Since the Codex conventions primarily depend on input from the FIDE laws we want them to be clear.

Also I do not attempt to get agreement on what the laws "appear" to say/intend but merely on the viewpoint that the laws fall short on unambiguous formulation giving rise to disagreement and confusion. At the same time I am aware that this is very much a niche issue and FIDE may judge it insufficiently relevant for game play to resolve ever. Which is probably OK for them but a pity for us composers since it forces us to resolve it in the conventions eventually. We are in the niches a lot!

The issue is not the playing of legal moves beyond DP but the adjacent one of playing them beyond the hard termination boundaries for 5REP and 75M which are now in the FIDE rules. You may believe they are not hard since there are still "legal moves" while I believe that termination ends every option to play a "legal move" or an "illegal move" - Elvis has totally left the building. That would be great since it demonstrates we have a non-trivial disagreement on the FIDE laws and that is a valid motive to request clarification. As I wrote earlier, I had a long discussion with 2 individuals who arrived at my conclusion independently - which is to say before we exchanged viewpoints and arguments on it. That discussion is still somewhere on chess.com. I presume you may find as many lawyers supporting the opposite view. Frankly, I don't care what the final FIDE verdict is, though I obviously believe that my view is closest to what the text currently says. Or it's all by chance and FIDE never even considered the interaction of the laws involved!

The core question is whether the hard 5REP/75M terminations in the 2018 FIDE laws are transparent to DP-evaluations or prison doors with the same solidity of stalemate (or checkmate for that matter though irrelevant). Note that the 5rep and 75M laws have clever formulations to make one another invisible but those are missing in the DP formulation. Before 2017/2018 there were no hard terminations on these drawing properties which made it natural that DP would ignore them.

I just put it up here as a point to resolve by FIDE. I don't expect anything from you except perhaps mention it when in stakeholder company. Or will a FIDE official honor this thread with a visit plus follow it up? Clearly the issue can be a bug to retro-composers since FIDE should resolve it before we do. I predict that whatever FIDE decides will be taken as a cue for the same approach to 3REP, 50M and DP in the conventions.

(Note: there is also an unavoidable fraction of our "legal move" debate here which I will resume by email - wrong thread here)

I think DP would indisputably have visibility over 5Rep & 75M in the Laws. However the latter two are excluded from Codex right now by default - I think the idea from a conservative point of view is to let them settle in to OTB chess for a few more years and achieve their final form. Does DP have visibility over whether Codex has excluded these rules? 🤣Madness

INCONSISTENCIES:

I hope for you that they do. If not, then DP will continue to draw on the basis of the 5REP and 75M visibility because that is what it says in the FIDE laws. Excluding these laws on a higher plane does not affect the interaction of FIDE laws on the FIDE plane. DP follows those by your wish.

You actually want to have everything both ways. One day you like the Codex to replace operations on the FIDE plane, the next day you want it as our private play thing. IMO, it's all fictional. Parts of the Codex function on a separate level because they have no counterpart in the laws, notably the unprovable conditions. The others are simply replacements in the FIDE laws to make compositions more powerful and effective. Do you think that a composer of endgames looks at the missing 50M/75M laws as conventions on a different plane than the game rules? If he did, he couldn't even produce a legal proof game for his 87 move endgame study. Proof games, by definition, operate on game rules and on game rules only, and without proof games retro-analysis does not exist. Once we recognize there is only 1 active game domain for any particular composition governed by its Active Game Rules, then it is clear that all the generic laws in it (like DP) operate on all the game conditions.

If you'd wish to have conventions with a special status, you'd have to carefully design explicit meta systems like the ones on unprovability. Because the setup with active game rules and proof games is untouchable you'll need to respect those at all times. Note for instance that my theory on unprovabilities striclty separates queries and rules in an orthogonal construction which assures that both systems can operate independently.

FAIRIES:

Addendum: The last post is actually more or less equivalent to what we discussed on fairy retro's and chess960. Once you change a rule (whether by stipulation or by codex) you must do it across the board to avoid inconsistencies. We solve a fairy retro with the applicable conventions, with active fairy game rules and with fairy proof games. And we apply DP to the active fairy rules - if we want it. The situation where a game rule only appears on a distinct convention plane disappears in this process.

RECENT CHANGES:

Addendum2: In the old days we had yet unfinished discussions about the status of 3reps and 50ms in the conventions. These conventions were and are quite unclear and, as isolated cases, were hard to approach through a generic concept, I referred to that situation as scarcity. But recently things have drastically changed. Now there are several exclusions of FIDE laws which demonstrably affect consistency over the complete rule/convention domains forcing us to rethink their relationship. That obviously includes the revisit of the legal status of the 3rep and 50m conventions - on the agenda anyway. They are all in the same "law-change" group.

Note that I think it is still possible  to have lawlike meta-conventions on a different plane but the decision to place them there needs to be carefully scrutinized for consistency and demands justification since there is the attractive alternative of treating them as active game rules as well. With the meta conventions always invisible in the proof games we need to feel that their meta-decisions are psychologically "real". They will for instance abort solution lines by random "draw" interventions - from the viewpoint of the proof game players following their game rules.

CLOSING THE CIRCLE:

Addendum3: Busy or out of breath? The codex is not a place where you can simply swing the rope on one end to get the required response at the other end. There is a leash on that end to keep it grounded and that leash is the proof game leash. The conventions are for the solvers and composers of the retro problems and may be manipulated to support them. But the problem solutions still need to be checked against proof games with zero visibility of the convention space. They only conform to whatever game rules they are fed. Those rules come from the FIDE laws and from conventions intended as games rule changes (and from fairy rules) but they are projected in a distinct game space where all rules are levelled for the game players who inhabit them. 'Closing the circle' refers to the act of validating that the same solution moves generated by the solver in accordance with the conventions can also be legally played by a player pair in the game space. In society we would refer to this process as a reality check.

The illusion is that reasonable, fair sounding and good looking conventions will provide uncontestable guidelines for composers and solvers. Not so. You can never skip the proof game check as we have known for a long time. The old issue of ending in stalemate when the conventions disallow playing either of 2 mutually exclusive e.p. moves (RS-logic) is an example of proof game power. There is no proof game with stalemate - therefore there can be no stalemate outcome irrespective of what the conventions imply.