[Retros] Ch5: Place of the Retro Logics

Andrew Buchanan andrew at anselan.com
Fri Jun 20 04:53:16 EDT 2014

Orthogonality is a great virtue.


The convention which prioritized PRA over RS (which is what we’re talking about here) was approved by a bunch of practitioners. And at the time I liked the idea that there was some kind of comparison between RS & PRA, which worked through many examples in detail. I hadn’t seen this kind of analysis before. Great work. But the rule itself is a bit ugly and difficult to understand and non-orthogonal, and seems to be provisional. The update of the Retro Corner tutorial pages was only half-done, resulting in chaos in the links. If it was a building job, they would have been called back and told to fix it for free. I hardly ever go to the Retro Corner now. The main article is fine, but fails to give any motivation for why the change was deemed necessary. (Maybe someone who was closer to the decision-making process can opine at to the motivation.)


Saving a few characters is not about ink, it’s about some people’s notion of terseness. But it’s a barrier to entry. Similarly, many people playing Magic the Gathering for a while were unhappy if a creature card with the ability “Flying” also stated “A creature with flying cannot be blocked except by other creatures with either flying or reach...”. But prioritizing removal of barrier to entry of new players is why Magic has thousands at major tournaments while the attendees at a chess problem convention can fit in a single small photo. And these days, superior tutorial methods (i.e. “Duels of the Planeswalkers” intro game have been developed which mean the cards can have reduced text.


I do differ, Guus, on your view of direct/help/self etc.


Chess problems are a “film of the book”. They require adaptation. There are two gaps to be filled: (a) history of the game (b) decisions for the players. We have been focusing mainly on the history, but actually there are a number of decision by the players which need to be addressed:

(1)    what is each player trying to do in the forward play? (force the goal, co-operate, force the opponent to achieve the goal)

(2)    can they resign? (no)

(3)    can they agree a draw? (no)

(4)    assuming the criteria are satisfied, can they claim a draw under 3R?

(5)    assuming the criteria are satisfied, can they claim a draw under 50M?


Q1 is the most important to answer. It’s not a “justification” – it shows how the tree of moves is to be searched.


I liked the itemization of composition elements: diagram, playing rules, basic retro conventions, retro logics, goal logics, remainder stipulation.


By the way that’s the first time I’ve seen an explanation of RV. I had no idea what it was. The meaning of the term had been successfully hidden.




From: Retros [mailto:retros-bounces at janko.at] On Behalf Of Kevin Begley
Sent: 20 June 2014 15:02
To: The Retrograde Analysis Mailing List
Subject: Re: [Retros] Ch5: Place of the Retro Logics




>"Once upon a time in the future, you will receive a message from an authoritive WFCC body announcing a few changes to the problem presentations..."


By far, your best post on the subject... to date... and some of your early posts were fantastic reading!  


This one reads positively Orwellian (!!) ... with a strong underlying current of undeniable Truth to be found in your every sarcastic prophesy. 

Brilliant!!  You have allowed the audience to experience both the horror and the ominous certainty of a tragic future. 


I absolutely love the way in which you have rolled out the anticipation for your coming proposals...




On Thu, Jun 19, 2014 at 11:42 AM, Guus Rol <grol33 at gmail.com> wrote:

Dear Retro-friends,


Once upon a time in the future, you will receive a message from an authoritive WFCC body announcing a few changes to the problem presentations. It is subdivided into the following items:


1. As of today, we will no longer tell the solver what kind of mating problem he is solving, direct mate, helpmate, selfmate or reflexmate.


2. He is to assume it is a direct mate in most cases


3. But he must make an exception if the white units on the board are overwhelmingly outnumbered by black units, in which case he should assume it is a helpmate.


4. If on the other hand the blacks are overwhelmingly outnumbered by the whites and direct mate is too easy, you should take it as a selfmate.


5. In all other cases the author should provide the correct stipulation being "reflex mate" or another form of mate.


6. These new conventions have been carefully prepared by former members of the Codex committee for "retro conventions".and we welcome them in our midst.


I guess you will not only find this extremely strange but also completely unncessary and unfruitful. To save a few cents in printing ink or a few bytes in a digital document (stipulations will be shorter), the independent status of basic "goal logics" has been compromised by mixing them up in positional evaluations. A worse crime to transparency is hard to imagine.


The direct-, help-, self- and relex-forms are required logics in traditional problem solving to complement the primary commands such as "mate in 2" or "stalemate in 5".It is of course OK to nominate the direct-form (or any other form) as a default but this requires no reference to external factors such as board positions. It is unimaginable that the problem solving community would accept the potpourri created by the points 1 - 6.


The role of the global retro logics - mainly pRA, RS, AP and RV (as a "rest" group) - in relation to retro problems is roughly the same as the role of direct- help- self- and reflex- logics in relation to traditional chess problems. They all provide "justifications" for accepting proposed solutions to problems. Examples: You can present 1 mating variation to satisfy a helpmate problem but the same variation will never count as a full solution to a direct mate. You can present a solution with an unproven e.p. capture which is OK in AP-logic but not in the other logics. You can present a solution where white castles to prevent black doing the same (RS-logic) but this would only be half a solution under pRA.


The common components of a retro-active problem are: Diagram, playing rules, basic retro conventions, retro logics, goal logics, remainder stipulation.For some unknown reason the retro logics have been.mixed into positional evaluation in the same way the goal logics were mixed up with board positions in the points 1 - 6 which started this post. The mixup is worse than it seems since (a) it makes people believe you actually should resolve certain positions through a certain logic where at best the applicable logic should be a default value (b) it totally ignores the logical effects created by 3R ad 50M. You may not like these effects but they are nevertheless completely real. I again refer to my R309 in PB which is a good example for almost everything.(c) The beautiful vague term RV with a potential to cover the forgotten and unprivileged states was removed from the retro-vocabulary; how would you now sell a reflex-mate problem where "castling right" provides one solution and "no castling right" another? Oh, and I refuse to mention SPRA (d) it obscures the powerful archetypal and independent nature of the retro logics. 


The whole program was delivered under the pretext that it covered most or all interesting problems found in the retro field. To my knowledge, the purpose of problem rules and conventions is not to cover what is already there but to make space for creativity and innovation. The cause of all this, I believe, is the "scarcity of retro-active issues in orthodox chess" which gives an excuse for treating it on a case by case basis. If such an approach were continued for retro-activity in fairy problems, the Codex committee would produce tons of digital paper for all its different cases!


The meta-concept at stake here is the "Power of Orthogonality".When breaking a subject into individual components one would like to assure maximum combinatorial capability as this delivers maximum freedom of action. E..g. you wouldn't have an arithmetic that only adds numbers without the digit "3" in them. With the retro components listed a few paragraphs ago this means ideally the freedom to combine Any diagram with Any playing rules (plus connected basic conventions) with Any retro logic with Any goal logic and with Any remaining stipulation. I know of quite a few reasons why that ideal cannot be attained but straightjacketing the retro logics as in the current Codex, is not one of them.


Best wishes, Guus Rol.   


Retros mailing list
Retros at janko.at


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://one.pairlist.net/pipermail/retros/attachments/20140620/7186d9e4/attachment-0001.html>

More information about the Retros mailing list