# [Retros] How many times could a Diagram be repeated during aproofgame?

Rol, Guus G.A.Rol at umcutrecht.nl
Tue Jun 5 12:12:07 EDT 2007

Olli,

>From the retrocorner-site at http://www.janko.at/Retros/Glossary/Codex1991.htm:

Article 18 - Repetition of Position
A position is considered as a draw if it can be proved that an identical position [21] has occured three times in the proof game combined with the solution.

Article 18 can be viewed as either an addendum to conventional drawing rules or as a replacement for them in the retrograde field. The latter seems more natural as the purpose for the rule is apparently to get rid of arbitrary (do or don't) draw claims in favour of a predictable automaticity. Wolfgang Dittman in "der blick zurück" supports a similar automaticity for the 50-moves rule and I fully agree with him.

PB R309 #7 operates correctly with article 18 in place, whether or not draw claims are permitted as well. If there is anything strange in how R309 plays out, it is not to be blamed on article 18 but on the peculiar properties of retro-strategy/retro-synthesis and on my devious mind.

The crucial position arises after 3. .. Ne6. Black cannot claim a draw and neither can the position be drawn automatically because the triple occurrence of the position cannot be proved. Without recourse to a retro-variant mechanism there is no option but to continue play. The very act of continuing play invalidates all histories in which the triple position occurrence was an option. After all, those would have terminated the game in an automatic draw at this point. In the histories that remain black wasted his right to castle 4 moves ago, so white profits with 4.Qf1!

Deep down, the logic of the solution is identical to that of a mutually exclusive castling problem. In the parameterization, replace the "dominant right for white to castle" by the "dominant obligation to ignore an uncertain repetition draw" and the remaining pieces fall into place.

Guus Rol.

> -----Oorspronkelijk bericht-----

> Van: retros-bounces at janko.at [mailto:retros-bounces at janko.at]

> Namens Olli Heimo

> Verzonden: dinsdag 5 juni 2007 13:38

> Aan: The Retrograde Analysis Mailing List

> Onderwerp: Re: [Retros] How many times could a Diagram be

> repeated during aproofgame?

>

>

> Rol, Guus kirjoitti 05.06.2007 kello 11:56:

> > Roberto indicated in an earlier response that he operates in the

> > context of the retro-conventions in de Codex. The automatic draw

> > indicated in article 18 takes place after the actual 3rd repetition

> > has been completed. As an example, Proca problems may start from a

> > drawn end state.

> Very strange. A few weeks ago there were Guus's problen 7# on

> this forum. Solution was 1.Bg8 Ne6 2.Bh7+ Ng8 3.Bg8 Ne6 4.

> Qf1!, reasoning that Black can no more castle, otherwise

> Black could have claimed for draw instead of 3. - Ne6. But

> think that Black just made that move. Now White has option to

> claim a draw, but he is trying to win. So he tries 4.Qf1, but

> then surprisingly 4. - 000. I have not read the article 18 in

> de Codex. I have read articles 1-2 which tell that problem

> chess rules are due to the on the board chess conventions

> unless for some special reason otherwise stated. And I see no

> reason why article 18 should make any exception. I must read

> that article. Olli.

>

> _______________________________________________

> Retros mailing list

> Retros at janko.at

> http://www.pairlist.net/mailman/listinfo/retros

>