[Retros] Massacre K+K proof games once again

Norbert Geissler norbert.geissler at ng37.de
Wed Nov 16 18:28:10 EST 2005

Hello Mario,

Answers to your questions:
1) You are right, at that time we had calculated 33 and 34 plies
But two years later even 35 plies were calculated.
Result: no sound K+K proof game ...
2) I didn't publish anything, this was always done by Gerd.
So it's up to him to answer this question.
3) As far as I remember:
33 plies: 2 hours
34 plies: 4 weeks
35 plies: 9 months
Of course, for 35 plies we had better hardware at hand ...

Hope that's what you wanted to know.

All the best,

Mario Richter wrote:

>Dear retro friends,


>in the award for the retros section of the year 1994,

>published in "Die Schwalbe", october 2004, Gerd Wilts

>and Norbert Geissler received a special award

>(congratulations!) for the following problem:




>White: Kc2

>Black: Kg8


>PG 17.0, Black captured on move 2


>Spezialpreis: Nr. 8491 von Gerd Wilts & Norbert Geissler


>1.e4 d5 2.exd5 Qxd5 3.Qh5 Qxa2 4.Qxh7 Qxb2

>5.Rxa7 g5! 6.Qxf7+ Kxf7 7.Rxb7 Rxh2 8.Rxc7 Rxg2

>9.Rxc8 Rxf2 10.Rxb8 Rxd2 11.Kxd2 Qxc2 12.Kxc2 Rxb8

>13.Bg5 Rxb1 14.Bxe7 Rxf1 15.Bxf8 Rg1 16.Rxg1 Kxf8

>17.Rxg8+ Kxg8


>The jugde (Werner Keym) wrote:


>"Die Autoren haben herausgefunden, dass es keine exakte

>Beweispartie mit nur 2 Steinen gibt und mittels einer

>Zusatzbedingung die erste dargestellt..."


>"The authors detected that there is no exact two-men proof

>game [mri: aka K+K massacre proof game], and realized the

>first exact one with an additional constraint ..."

>[mri: black captured on move 2].



>My questions are:


>(1) The correct statement with respect to the K+K proof games

> should be:

> ..."The authors detected that there is no exact two-men

> proof game *with 33 or 34 plies*".

> Right?


>(2) Have the methods and results of that research been

> published somewhere?

> I was only able to find the following article by Torsten

> Linss, which refers to the research of Gerd and Norbert:

> www.math.tu-dresden.de/~torsten/problem/coco_txt.html

> (and an older one by Karlheinz Bachmann in "Die Schwalbe",

> Nr.141, Juni 1993).


>(3) With today's (home) computers it is possible to do the

> complete search for the 33-plies-K+K-proof games in

> less than 2 seconds, and that for 34 plies in approx.

> 10 minutes.

> It would be nice to compare that times to those needed

> 11 years ago. (Especially to Gerd+Norbert, if they

> are reading this) Is any info about that still available?


>Thanks for any clarification,


> mario





>Retros mailing list

>Retros at janko.at





More information about the Retros mailing list