Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Sergei S. Markoff
Posts: 224
Joined: Mon Sep 12, 2011 9:27 pm
Location: Moscow, Russia
Contact:

Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Sergei S. Markoff » Thu Jun 06, 2019 9:52 pm

Turochamp, Machiavelli, SOMA, SOMAC, Shannon's programs — we know a lot about them, even there are several reconstruction attempts especially by Chessbase team (Frederic Friedel, Martin Feist). However with a German attempts it is known a little. Zuse's first works on Plankalkül provides us with some move-generation routines, but I can't find any sources for his ideas on move selection. Guys from PyTuroChamp team (https://mdoege.github.io/PyTuroChamp/engine.html) made their reconstruction of Zuse's engine (https://github.com/mdoege/PyTuroChamp/b ... er/plan.py) and they even digged piece values from somewhere: 1 for a pawn, 2 for a knight or a bishop, 3 for a rook and 4 for a queen. I hope they don't invented these strange numbers themselves but I'm still unable to find any clue in Zuse's online archive.
But the paper machine of Günter Sсhliebs from DDR was never referenced in Western sources. In Soviet books and articles it was referenced multiple times (especially in the works of legendary Anatoly Kitov — [1], [2], [3]) among with Shannon's work. The original source for Schliebs' work is [4]. I never seen [4] (but started some attempts to obtain the scan of this article), but Kitov's works provides several details, at least evaluation function structure and some feature scores and also the point that Schliebs' "program" was fixed-depth searcher.

Evaluation features of Schliebs' paper machine:

queen — 9 points
rook — 5
knight, bishop — 3
pawn — 1
backward pawn penalty — 0.5
isolated pawn penalty — 0.4
doubled pawn penalty — 0.3
mobility (for each field for "most strong pieces" (?))

One work of Kitov occasionally assign 0.5 score for the pawn instead of 1.0 (mistake?)

1. Китов А.И. (1956). Электронные цифровые машины // http://www.kitov-anatoly.ru/naucnye-tru ... vye-masiny
2. Китов А.И., Криницкий Н.А. (1959). Электронные цифровые машины и программирование // http://www.computer-museum.ru/books/ecm_i_prog.pdf
3. Китов А.И., Криницкий Н.А. (1958). Электронные вычислительные машины и программирование // http://elib.ict.nsc.ru/jspui/bitstream/ ... ov1958.pdf
4. Sсhliebs G. (1953). Über die Gründzuge eines Programms für eine Schachspielende Rechenmaschine / Funk und Ton, 1953, vol. 7, pp. 257—265.
The Force Be With You!

Dann Corbit
Posts: 9849
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Dann Corbit » Thu Jun 06, 2019 10:26 pm

How long must we wait until someone makes a Winboard or UCI port?
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Sergei S. Markoff
Posts: 224
Joined: Mon Sep 12, 2011 9:27 pm
Location: Moscow, Russia
Contact:

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Sergei S. Markoff » Thu Jun 06, 2019 10:34 pm

Dann Corbit wrote:
Thu Jun 06, 2019 10:26 pm
How long must we wait until someone makes a Winboard or UCI port?
I think it's easy to do using PyTuroChamp Shannon's program as the base, it will take not more than few minutes, but I want to examine the original Schliebs's article to be sure in details. For example mobility terms and search routine are not 100% clear in Kitov's works.
The Force Be With You!

Dokterchen
Posts: 101
Joined: Wed Aug 15, 2007 10:18 am
Location: Munich

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Dokterchen » Fri Jun 07, 2019 7:35 am

Thanks a lot Sergei! A very nice and interesting finding.

KR
Torsten
Torsten Schoop

Sergei S. Markoff
Posts: 224
Joined: Mon Sep 12, 2011 9:27 pm
Location: Moscow, Russia
Contact:

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Sergei S. Markoff » Tue Jun 11, 2019 9:36 pm

Sergei S. Markoff wrote:
Thu Jun 06, 2019 9:52 pm
I never seen [4] (but started some attempts to obtain the scan of this article)

<...>

4. Sсhliebs G. (1953). Über die Gründzuge eines Programms für eine Schachspielende Rechenmaschine / Funk und Ton, 1953, vol. 7, pp. 257—265.
Hooray!!

Here is the paper: https://yadi.sk/d/my7VqYaYaaLc1w

And it looks really interesting :)
The Force Be With You!

Dokterchen
Posts: 101
Joined: Wed Aug 15, 2007 10:18 am
Location: Munich

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Dokterchen » Wed Jun 12, 2019 6:18 am

Sergei S. Markoff wrote:
Tue Jun 11, 2019 9:36 pm
Sergei S. Markoff wrote:
Thu Jun 06, 2019 9:52 pm
I never seen [4] (but started some attempts to obtain the scan of this article)

<...>

4. Sсhliebs G. (1953). Über die Gründzuge eines Programms für eine Schachspielende Rechenmaschine / Funk und Ton, 1953, vol. 7, pp. 257—265.
Hooray!!

Here is the paper: https://yadi.sk/d/my7VqYaYaaLc1w

And it looks really interesting :)
For 1953 a really remarkable paper. In the summary ("5. Zusammenfassung") there is even a concept of an automated tuning mentioned.

Thanks a lot Sergei!
Torsten Schoop

Rom77
Posts: 3
Joined: Wed Oct 24, 2018 5:37 am
Full name: Roman Zhukov

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Rom77 » Wed Jun 12, 2019 6:20 am

In order not to create a separate topic for the sake of one message, I will write here.

There is an interesting article by David Levy "Alan Turing on Computer Chess" from S. Barry Cooper, J. van Leeuwen "Alan Turing: His Work and Impact" about Turing's "paper machine":
googlebooks

Gerd Isenberg
Posts: 2125
Joined: Wed Mar 08, 2006 7:47 pm
Location: Hattingen, Germany

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Gerd Isenberg » Wed Jun 12, 2019 12:20 pm

Rom77 wrote:
Wed Jun 12, 2019 6:20 am
In order not to create a separate topic for the sake of one message, I will write here.

There is an interesting article by David Levy "Alan Turing on Computer Chess" from S. Barry Cooper, J. van Leeuwen "Alan Turing: His Work and Impact" about Turing's "paper machine":
googlebooks
Thanks! I missed that.

Gerd Isenberg
Posts: 2125
Joined: Wed Mar 08, 2006 7:47 pm
Location: Hattingen, Germany

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by Gerd Isenberg » Thu Jun 13, 2019 3:27 pm

Sergei S. Markoff wrote:
Tue Jun 11, 2019 9:36 pm
Sergei S. Markoff wrote:
Thu Jun 06, 2019 9:52 pm
I never seen [4] (but started some attempts to obtain the scan of this article)

<...>

4. Sсhliebs G. (1953). Über die Gründzuge eines Programms für eine Schachspielende Rechenmaschine / Funk und Ton, 1953, vol. 7, pp. 257—265.
Hooray!!

Here is the paper: https://yadi.sk/d/my7VqYaYaaLc1w

And it looks really interesting :)
G. Schlieb referred Shannon's groundbreaking paper, which is surely the base his paper. It has the same board represention (same piece codes), evaluation function and similar search ideas concerning selectivity and quiescence.

Schlieb

Code: Select all

f(P) : 200(K—K’) + 9(D—D') + 5(T—T') + 3(L—L' + s—s') + (B—B') 
              —0.5(B.—B.’ + B,——B.' + B,—B,‘) + 0,1(M—M’) + ...
Shannon
https://www.computerhistory.org/chess/d ... 14f453dde/

Code: Select all

f(P) = 200(K-K') + 9(Q-Q') + 5(R-R') + 3(B-B'+N-N') + (P-P') - 0.5(D-D'+S-S'+I-I') + 0.1(M-M') + ...

syzygy
Posts: 4442
Joined: Tue Feb 28, 2012 10:56 pm

Re: Widely unknown pioneering chess "paper machine" by Gunter Sсhliebs

Post by syzygy » Thu Jun 13, 2019 10:57 pm

Gerd Isenberg wrote:
Thu Jun 13, 2019 3:27 pm
G. Schlieb referred Shannon's groundbreaking paper, which is surely the base his paper. It has the same board represention (same piece codes), evaluation function and similar search ideas concerning selectivity and quiescence.
And these are by far not the only similarities between the two papers. Ouch...

Sure, Schlieb did include a reference to Shannon's paper, but he borrowed considerably more than is justified by that reference.

Schlieb made some subtle mistakes:
Shannon wrote:A chess "position" may be defined to include the following data:
...
(4)A statement of, say, the last move. This will determine whether a possible en passant capture is legal, since this privilege is forfeited after one move.
...

For simplicity, we will ignore the rule of draw after three repetitions of a position.
Schlieb wrote:Eine Problemstellung ist durch den Stand des Spieles (beim Zuge u) gegeben, und dieser ist durch folgende Angaben eindeutig bestimmt:
...
3. die Vergangheit hinsichtlich der Feststellung,
...
b) welche Bauern bereits ein Feld vorgegangen sind (Enpassant-Schlagen),
...
d) ob Wiederholungen stattfanden (Remis nach dreimaliger Wiederholung).
Clearly 3b) ("which pawns have already advanced a square") is not what you need to determine whether an en passant capture is possible. And to correctly detect 3-fold repetitions, you need more than 3d) ("whether repetitions have taken place").
Shannon wrote:In chess there is no chance element apart from the original choice of which player has the first move. This is in contrast with card games, backgammon, etc. Furthermore, in chess each of the two opponents has "perfect information" at each move as to all previous moves (in contrast with Kriegspiel, for example). These two facts imply (von Neumann and Morgenstern, 1944) that any given position of the chess pieces must be either:

(1) A won position for White. That is, White can force a win, however Black defends.

(2) A draw position. White can force at least a draw, however Black plays, andlikewise Black can force at least a draw, however White plays. If both sides playcorrectly the game will end in a draw.

(3)A won position for Black. Black can force a win, however White plays.
Schlieb wrote:Aufgabe und Lösung sind ferner scharf definiert: Schachmatt und nach den Regeln erlaubte Züge. Im Schach gibt es - anders als bei Kartenspielen - nur einen Freiheitsgrad, die Verteilung der Farben. Jeder Spieler hat jedoch jederzeit die "volle Information" über die Brettposition. Aus diesen beiden Eigenarten folgt, daß jede Problemstellung nur drei Möglichkeiten in sich birgt [3]: sie ist
A. eine Gewinnstellung für Weiß,
B. eine Remisstellung für Weiß und Schwarz,
C. eine Gewinnstellung für Schwarz,
wie der Gegner auch immer spielt.
The similarity is again striking, and where Schlieb deviates, he is inaccurate. "Wie der Gegner auch immer spielt" (however the opponent plays) is far too simplistic in the case of a position that is drawn or won for the opponent.

Schlieb also doesn't quite seem to understand Shannon's argument that chess would be at least a draw for white if players could pass instead of being forced to move. He incorrectly seems to think that this argument works only if (regular) chess is a draw. And I don't understand why such a rule change would make an evaluation function unnecessary.

Post Reply