[Retros] The paradigmic divide on (retro) Rules and Conventions

Andrew Buchanan andrew at anselan.com
Wed Jun 18 15:16:30 EDT 2014

Dear Guus,


Ok thanks for the pointers to the relevant emails. That will save a vast amount of reading. Hopefully these will explain the mass of terminology used in the later paragraphs.


“My” e.p./mate “issue” is just that you seem to be using my position (whatever that was, I forget) to bolster your own case! :) I don’t think I have any problem with the diagram itself, but please remind me what it was.


I will read these emails and revert to you, but nothing I see in *this* email gives me any enthusiasm, I’m afraid.


As Andrey Frolkin observed, almost all most problems don’t need the conventions. Either the position contains no conditional move, or has no forward play, or if it has forward play it can be resolved through use of the en passant *rule* and castling *rule*. And these days PRA would take priority anyway over RS for the remaining cases. Nevertheless, we do want to ensure that the RS is finally cleaned up. But given the context, the case for a massive paradigm shift which destroys the behaviour of any current compositions is poor. And how can we know the need is for applying conventions, until after we have exhausted all retroanalytic reasoning?


I think you will find that the notion of valid moves was developed in a different context, but it was a long time ago. It concerns me that you feel the need to capture kings in order to validate the theory. But let’s see the diagram and use it as a test case.


It also concerns me that you find “positions” inconvenient to handle, when they encapsulate so much (but not all) of the relevant game history. Ignoring them makes so much of the argument unnecessarily infinite.




From: Retros [mailto:retros-bounces at janko.at] On Behalf Of Guus Rol
Sent: 18 June 2014 21:33
To: The Retrograde Analysis Mailing List
Subject: Re: [Retros] The paradigmic divide on (retro) Rules and Conventions


Hi Andrew,


For the "results" I adressed in my mail to Kevin, all you need to read are the 3 or 4 mails beginning with the 2nd version of "The basics of the relationship between Laws and Conventions". This is where I started from scratch to treat the subject of "retro-activity" (as opposed to retro-analysis).


Obviously, whether or not you consider the points adequately handled depends on whether you go along with my paradigm for a while. At least I believe they are consistent within my theory.


.Your e.p./mate issue dissolves already in the first article. Since my theory changes the order of execution - first the conventions/selections, then the FIDE laws - the games with e.p. rights are already filtered out before FIDE law is applied. And so it is "Mate".


There are other ways to get to the same result e.g.by continuing play after "possible mate". You will find there is no "permission" for e.p. in any variation and so the King is always captured. In retrospect: "possible mate" was "actual mate" in the first place. You will probably object that it is illegal to have your King captured but I do recall that one Andrew Buchanan developed the concept of "valid moves" which are not necessarily legal. "Valid moves" are the perfect means to investigate and define terminating states such as "mate" and "stalemate". Works as well with retro-issues as with standard orthodox chess laws.


Yes, my model skips over the first class concept of "positions", at least so far. These are not so important in my theory as the retro-solver will have to accept that he is not playing a position but a diagram representing many positions and countlessly many more proof games. "Position being defined as a diagram or chess board setup where all retro-states are known". The DGC (diagram game cluster) may or may not deliver "positions" during some stage of the reduction process in the solution, but there is no guarentee for it. It all depends on what the solver encounters in his journey through the variations. In a diagram with mutual castling rights - representing at least 2 positions - one can still play 1. Ne5 mate!


May be I can help to recapitulate the key notes of older mails. Most were on or related to the subject of FIDE redefining "castling right" in the context of position repetition.(3R) as a right that could be executed in the future. That is, nobody was sure whether or not this was the right interpretation of FIDE law, but it might be. In my share of the discussion I argued on both sides (1) by giving reasons why going dynamic on castling rights would not be a good idea (2) by trying to define what rule/convention changes would be required if it were adopted anyway (3) by assessing what additional issues could arise from this interpretation in retro-problems with uncertain states. Quite a good mess altogether. May be better to first find out the correct interpretation of the Law before reading up on the "what if" scenarios in that mail thread.


Best wishes, Guus Rol.


On Tue, Jun 17, 2014 at 8:23 PM, Andrew Buchanan <andrew at anselan.com> wrote:

Hi Guus,


Since 5-May there have been at least 120 mails, many of them very long, on this subject in the Retro Analysis Mailing List. I am still trying to catch up, following my business trip to USA when this dialogue opened. Would it be possible for you to identify the ones that properly make all the points you summarize below? Or pull together the key paragraphs into a single note right away?


In particular, what was this e.p. mate that I apparently made which is causing me so much pain & trouble now?


I did read one email “The basics of the relations​hip between Laws and Convention​s”, which began in the right way, by saying that the we start with the game, and look at how to adapt it to compositions. But your model seemed to be missing the concept of position as a first class object, although you mentioned it in an aside. You are just dealing with piles of billions of partial games. It may be more abstract, but it’s messy and inefficient.


You and I really have different aesthetics, it is becoming increasingly clear:

(1)    I *like* what you term the “classical indeterminacies” under RS. They are really neat. They are the logical consequence of dealing with a composition defined for a diagram rather than a game. That’s the way the world is, and it’s beautiful.

(2)    And conversely, it is wrong that a convention should cause an open position to become a stalemate, or some other pathological interaction with the Laws. It’s no longer chess.

(3)    I don’t want to “marginalize” conventions, but in my perspective they are the big guns, to be used sparingly. I want chess retro analysis to proceed as far as I can until we need the convention for the critical final steps in reasoning.

(4)    I agree that there has been a dearth of hard logical reasoning and sensible abstraction applied to the rules/convention space over the decades. I do not see evidence that your own approach is the only way to do it cleanly. I’ve started to do some mathematical modelling, but I am at least 80 emails behind you.




From: Retros [mailto:retros-bounces at janko.at] On Behalf Of Guus Rol
Sent: 18 June 2014 01:25
To: The Retrograde Analysis Mailing List
Subject: Re: [Retros] The paradigmic divide on (retro) Rules and Conventions


Hi Kevin,


I am glad you understood the message I tried to get across even when you consider it badly constructed. Some (but not most) of my language is lighthearted and not designed to withstand semantic disection. As I do not hold the work of the Codex committee in great esteem, I set out to describe it as produced by  "scribes". Why I consider "selection rules" more appropriate than "conventions" you can read in my first article on the Rules and Conventions. Since the retro conventions are really about "selecting games" rather than about "permitting moves" in my paradigm there is nothing arbitrary about the choice of this term. But alternative suggestions are always welcome. None of my terminology is designed to be definitive, it only serves to separate distinct ideas and operations. But there will be many rewrites before all this is in my book..


You have not noticed what my first 3 articles have delivered but it is really quite a lot. All that is seen are the straight lines from cause to effect but the collateral damage caused to crooked competing theories is easily overlooked. To name just a few:


1. It makes an end to all attempts to marginalize the retro-conventions in relation to laws and rules.

2. It states that the conventions are not "sweets for the kids" but a mandatory requirement in every form of chess supporting the retro-field both orthodox and fairy. If you read discussions e.g. on "disparate chess" you can see them all rendered futile under this dictate. Just add the required conventions to any fairy type as was done to orthodox chess.

3. By reordering the time-relation between conventions and laws - conventions before the laws instead of after - all classical indeterminicies are gone. Famously, Andrews example with a "possible" e.p. move plus mate is resolved instantly, and with it many similar situations.

4. These principles delivered a step by step approach to the status of 3R (conventional) automation and 3R claimed draws. Plus its final resolution. More importantly, it poses serious questions to anyone choosing to deny its conclusions.

5. The demand for a retro Rule Book was established and some of its content defined. This need not worry anyone as a theory claiming the power to handle fairies, could be expected to easily manage a few amendments to orthodox chess laws.

6. Anybody with a mathematical mind can see the portential of an approach that replaces case based static conventions by generic and dynamic first order logic. It eradicates the influence of uncomputable opinion and authority based "expertise". Finally there is a window towards transparent and provable conclusions.


You may not agree to all this but I suggest that these few points already comprise bigger changes to the retro field than ever proposed before. You complain that I always promise to write more later and such is true. But I also leave no post without setting a little concrete step forward. That it may take some time to reach my destination is all due to Albert Einstein: "Everything should be made as simple as possible, but no simpler".


Best wishes, Guus Rol.



On Tue, Jun 17, 2014 at 1:19 AM, Kevin Begley <kevinjbegley at gmail.com> wrote:



This is beginning to read like a deliberately bad parody.


"Had  ['the scribes'] done their job, the conventions would have never been named 'conventions' but probably something more appropriate like 'selection rules'."  


Let's break that down... 

1) The function of "the scribes" was to copy, read amend, explain, and protect the law.

2) Because they did not perform their function, your disfavored jargon ("retro conventions") gained the upper hand over your preferred jargon ("retro selection rules"). and 

3) The reader is given absolutely zero cause to care about either term. 


In fact, the reader has probably lost all reason to care, because this author has been carelessly inventing terminology (e.g., the Digital Boards Theory, where the word "digital" reads like the worst assault on logic ever performed, in the interest of obtaining a meaningless acronym). 


You keep describing the bread, in terms that no reader need care, as if to avoid discussing what's between the slices. 


Forget the scribes, and worry about your function:  explain what is the significant difference in your approach, and persuade readers why they should want to consider it. 


Lovecraft managed to describe Cthulhu, without any need to waste time developing an aimless jargon. Steinbeck would not have lamented so long about the hard row that the tumbleweeds of his own jargon must hoe. 


If your approach might actually constitute an improvement, why have you deliberately hidden it away in useless jargon? I can endure the fog only for a purpose, and you have yet to show any product.


Let's cut to the quick...  

1) pour what you have into a shot glass. 

2) let the reader sample it, and  

3) only then, will anyone care to read about your distillation process. 



On Mon, Jun 16, 2014 at 11:23 AM, Guus Rol <grol33 at gmail.com> wrote:

Dear retro-friends,


You will have noticed that there exist no uniform agreement on the understanding of the relationship between Rules (Laws) and Conventions in the retro-field. I will give my view on where the schism originates and what it entails primarily. On this level of abstraction I do not expect to deliver irrefutable proof or conceptual superiority. I only invite you to evalute both approaches on the basis of their effectiveness and expansive capability.


The "other" paradigm is based on a common understanding of the concept "Convention". An appropriate definition may be (Mirriam Websters): A custom or a way of acting or doing things that is widely accepted and followed. May be a similar description will be found somewhere in the rule books and explanations by FIDE and WFCC. In such an environment, the conventions will obviously give way to everything published under the more formal headings of approved "Laws", "Rules" and "Regulations". It is also predictable that such a starting point would generate the ideas that are currently prevalent in the retro community.


"My paradigm" is fundamentally different. It didn't come from reading every bit of law and jurisprudence to be found but from studying the subject of retro-analysis and more in particular of retro-activity (retro uncertainties) through its architecture. I found that - without the conventions - the subject lacked an essential command & control structure necessary to bridge the gap between retro-problem and "FIDE law". I concluded that a rigid formal decisioning system was needed, operating not in contention with "FIDE law" but performing a distinct reduction task all of its own. Strange enough, the commands in this control system were more or less the same as could be found in what regulatory bodies had baptized "Conventions". This left me with no other option than to redress the underprivileged conventions to their mandatory role of  "controllers of the retro-active universe". In my view, this is what should have been done by the scribes of the these conventions in the first place from a true understanding of the nature of the retro-field. Had they done their job, the conventions would have never been named "conventions" but probably something more appropriate like "selection rules". And may be they would have written some conventions alongside of them on subjects that truly begged for "widely accepted customs"..


There is much more to say about this paradigmic schism as the gap gets wider when the application scope gets larger. If you read my first post on "the basics of the relationship between rules and conventions" (actually the 2nd one by that name) you can find the difference on the chess board already on the first page. It is up to you to chose. And you can always change your mind, of course.


Best wishes, Guus Rol.


Retros mailing list
Retros at janko.at


Retros mailing list
Retros at janko.at


Retros mailing list
Retros at janko.at


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://one.pairlist.net/pipermail/retros/attachments/20140619/a72c1757/attachment-0001.html>

More information about the Retros mailing list