Killer draughts

[Juri]

Благодаря этому разговору я вспомнил о письме Питера Михаельсена к Эрику ван Дюссельдорпу:

Dear Eric van Dusseldorp,

I have read your description of Thai checkers and your "pseudo-plädoyer" for transferring a rule from Thai checkers to International checkers. I found both very interesting!
Your proposal of using the "king-halt"-rule in the International game is not new. According to the rules of "Central-South German checkers" the king has to stop at the square immediately behind the last captured piece, see J.Boyer & V.R.Parton: LES JEUX DE DAMES NON-ORTHODOXES, Paris 1956, ch.11. In that same book it is mentioned that the former Dutch champion A.-K.-W.Damme proposed this rule in 1922 for the game on 100 squares. Boyer/Parton remarks that 2 kings against 1 always win. Damme's proposal was rejected by B. Springer and other masters (controverses in Le feu de Dames, 1922). A.C.W. Wageningen remarked that in positions in the normal game lost for 3 pieces against 1 king, the 3 pieces would win easily. After these controverses Maxime Fayet came up with a new proposal preserving the essence of the project. The king-halt rule proposed by Mr. Damme should only be used in case of multiple captures. This idea was likewise abandoned.
None of these two proposals are exactly identical with yours. The king-halt rule used in Thai checkers appears to be the same as that described in a brochure, published in Hamburg 1951, as that used in Germany nowadays: i.e. that after ALL captures the king must be placed on the square immediately behind the piece(s) captured, not only after the last piece in case of multiple captures.
This rule has also been dominant in Denmark since the mid 20th century. I found the earliest evidence in a children's encyclopedia, published in 1941. According to Louis van Deven, Illinois, U.S.A. this rule is also used in Finnish checkers. An excange student, Tuomo Kasanen demonstrated to him how they played checkers in Finland. In the Danish/Finnish/German game 2x12 pieces are used on an 8x8 board, ordinary pieces do not capture backwards and there is no obligation of capturing the most, if there is a choice. L.v.Deven gives some examples, and thells that "one odd thing about this game is that 3 kings always win against 1 king, but 2 kings never do." (Het Nieuwe Damspel 1981). You may compare this statement with that of Boyer/Parton referred to above.
The king-halt rule was used in Germany already in the early 19th century. It is mentioned in ARCHIV DER SPIELE, Berlin 1819: I quote: "Deutsches Damspiel. Die einfachen Steine gehen nur einen Diagonal-Schritt vorwärts; aber niemals rückwärts, sondern bloss rechts und links. Eben so schlagen sie auch nur. Die Damen gehen vor- und rückwärts über die ganze Diagonal-Linie, so weit sie kommen können, und schlagen auch so, jedoch nur dergestalt, dass sie wie ein einfacher Stein die Dame unmittelbar hinter den geschlagenen Stein setzen müssen." Several contempory German game book authors describe the rules of German checkers without this rule, so it was apparently not universally accepted.
I have downloaded several Zillions-implementations, which play varous checkers variants, including German checkers. The two German checkers programs I have tried follow different rules, one with and one without the king-halt rule. Both are made by German programmers. Thus, some Germans do still use that rule, when playing checkers.
Arie van der Stoep in his A HISTORY OF DRAUGHTS, Rockanje 1984, knows examples of this rule also from Austria and Argentina (p.164).
It is a very good question, how this rule was introduced in Thai checkers.
As you know, Jean-Bernard Alemanni has recently got some more detailed rules of the Thai game from a Thai player.
We must conclude that there is no evidence of Frisian checkers being played in Thailand, cf. the article "Fries(?) dammen in Thailand" door Arie van der Stoep, HOOFDLIJN mei 1997 nr.54, p.24-26.
Recently I have come across some rules of Czech draughts/checkers. I have already sent these to Jean-Bernard and now send them to you, too.
In an article by A.W. de Haan from 1957, reprinted in the newest issue of OER ALLES, I found some interesting pieces of information concerning the distribution of ortho-diagonal, or - in my opinion: rather in orthogonal checkers. The source of information was the former world champion, Herman Hoogland, Utrecht: "In vroeger tijden werd er in Duitsland ook "dwars en recht" gedamd. Maar ook nu nog gebeurt dit in Turkije, Griekenland, Bulgarije, Hongarije, Albanië, Egypte, Perzië, in een deel van Israël, in Abessinië, India en Indonesië, maar dan op een bord met 64 velden."
From other sources I know about orthogonal checkers in Turkey, Greece, Egypt, Israel/Palestine, and also in Kuwait, Bahrain and at the East African coast near the border between Kenya and Tanzania. Finally an orthogonal, perhaps ortho-diagonal variant is played in Armenia. Several countries on H. Hoogland's list are not mentioned in A.v.d.Stoep's 1984 book, or in K.W. Kruijswijk's from 1966. I wonder, if Hoogland's sources of information have been preserved in an evt. archive. According to A.v.d. Stoep (1984, p.162) Hoogland kept up a correspondence with a player in Ankara who made problems in Turkish checkers. It seems thus that the former world champion was really interested in that kind of checkers.
Best wishes,
Peter Michaelsen.

В начале этого письма говорится о том, что впервые подобные правила пытался ещё внедрить в сто клетки Арнолд Дамме в 1922 году, но большинство мастеров с Бенедиктом Шпрингером отклонили это предложение...


@Tot@ - http://shashki.com/PNphpBB2-viewtopic-t-2362.html

[Juri]

In this letter there is one interesting idea - to use the Killer-rule only for Multiple capturing (and when last captured one is a king).
In this case 2 can not win against 1.
But 3 - can.

In this case less differences comparing normal rules, but amount of victory also will grow.

In this situation Single king on the main line will not lose against One king and one (two) peaces cutted by this main line.
So main line still important.

[MLWi]

Juri, do you know why both Damme's and Fayet's proposals were dismissed? What's wrong with the killer rule?

Mats

[Juri]

Juri, do you know why both Damme's and Fayet's proposals were dismissed? What's wrong with the killer rule?

Mats

@In that same book it is mentioned that the former Dutch champion A.-K.-W.Damme proposed this rule in 1922 for the game on 100 squares. Boyer/Parton remarks that 2 kings against 1 always win. Damme's proposal was rejected by B. Springer and other masters (controverses in Le feu de Dames, 1922). A.C.W. Wageningen remarked that in positions in the normal game lost for 3 pieces against 1 king, the 3 pieces would win easily. After these controverses Maxime Fayet came up with a new proposal preserving the essence of the project. The king-halt rule proposed by Mr. Damme should only be used in case of multiple captures. This idea was likewise abandoned.@

You mean these words?
I guess they told about king-hult using Not only in a sutuation when the king capturing an opponent King (killer-draughts), but in each situation (Thai draughts), even when the king capturing peaces (Damme's proposition). It was clear refusing, because it's really another game.


for example here white will win...

And even restriction of Fayet (use the king-hult in case of muliple captures) didnt help.

[Juri]

But I found the Fayet's proposition has a sense - @use the king-hult in case of muliple captures@.

It is possible to combine the Killer-rules which are proposed now (to use king-hult in a sutuation when in last moment of taking process the king capturing an opponent king) and ADD Fayet's restriction (to use the king-hult ONLY in a case of muliple captures)

In this case less differences with normal game, but 3 kings win against 1 always.
Also some ideas in end-game from the Normal rules could be saved.





when only 1 peace reached 15 - it's a draw (so the disposition of peaces and main line still is important)


and anyway the number of victories will grow too!

in the Killer-rules which are proposed now this situation will be draw - 1...50-11 =

[Juri]

But I aggree, for now the main target is to show that the number of victrories will grow significantly and the game will not get a lot of changes, to attract more people for the idea of killer-draughts.
Next step - to discuss and find the best variant of it, which gives minimum changes and the maximum effects.

http://shashki.com/PNphpBB2-viewtopic-t-2362.html
on the russian Forum the autor of 64 draughts programme @Aurora@ A. Svirin posted how the number of victories will grow in end-games. You can translate with google:

AlexanderS
Тема сообщения: СообщениеОтправлено: Янв 31, 2016 - 05:43 AM

@Немного статистики, если кому интересно будет:

По эндшпилям киллер-шашки в 100 имеют больше в 1.7 раз выигрышных позиций (по всем 6-шашечным позициям; для 5 шашек - 1.8, для 4 шашек - 1.9).
По разным классам разброс большой. Например, если взять 3х3 только простые то в 100 количество результативных позиций 59 млн, в киллер - 138 млн.
По всем классам 3х3 (с дамками) количество результативных позиций 3.9 млрд против 12.5 млрд.@

[Juri]

If to use King-hult Only for Multiple capturing than we will get end-games similar as in 64 draughts.

Interesting. This one is it possible to win?

[MLWi]

Juri, if I understand you correctly, you propose that the king-halt rule is used only when the captured king is last in a multiple sequence. I don't think that draught players can object against this rule change, because it is minimal and does not affect the character of the game.

So this means that three kings can always win against one king?

What do you think is a good name for this variant? "Dame Halt Draughts"? "Modern Polish Draughts"?

Mats Winther
http://two-paths.com/bg

[Rein Halbersma]

But I aggree, for now the main target is to show that the number of victrories will grow significantly and the game will not get a lot of changes, to attract more people for the idea of killer-draughts.
Next step - to discuss and find the best variant of it, which gives minimum changes and the maximum effects.

I don't understand the need for "Killer-light". The endgame rules will change so the opponents who want to preserve the endgame literature will not accept this anyway. But furthermore, the drawing margin is quite a bit bigger. At first glance, the endgame becomes similar to those with +/- draws. In any case, what is really being gained? In my opinion, there is not viable place for Killer-light.

I think those who want change should embrace Killer and get the low drawing percentage. OTOH, the Thai rule gives very unnatural endgames compared to Killer. They essentially kill all line-endgames with a king holding several man behind a diagonal. This seems like a strategic loss for the game.

[Rein Halbersma]

Juri, if I understand you correctly, you propose that the king-halt rule is used only when the captured king is last in a multiple sequence. It sounds logical, because it means that the capturing rules of the game are retained, i.e., that one must always capture the longest sequence. I don't think that draught players can object against this rule change, because it is minimal and does not affect the character of the game.

So this means that three kings can always win against one king?

What do you think is a good name for this variant? "Dame Halt Draughts"? "Modern Polish Draughts"?

Mats Winther
http://two-paths.com/bg

The main objection comes from the obsolence of endgame theory. This happens regardless of how big the change is. Given a choice between having 3 vs 1 endgames being won, versus 2 vs 1 endgames, the choice seems easy. Why tolerate draws at all? Besides, the Killer endgames play quite naturally, and don't take a lot to get used to.

[MLWi]

An important defensive property of draughts is to enter a drawn endgame. This is no longer possible in killer draughts. However, if only 3 against 1 is a win, this defensive property is retained, to a degree. Arguably, it is important to retain such a defensive stratagem, because it makes the game deeper and more interesting. Otherwise, to lose a piece means that the game is theoretically lost. But today, should you lose a piece, you can continue and play an interesting game. This is an important argument.

Mats

[Rein Halbersma]

An important defensive property of draughts is to enter a drawn endgame.

I agree that when neither player makes a major mistake (such as losing material), a draw is a fair result. But entering a drawn endgame by sacrificing one or two pieces has been the death of competitive draughts for a long time. Many great strategies don't get rewarded by a drawing percentage >90%. It's like a 90% income tax, there's no proper reward.

This is no longer possible in killer draughts.

This is not true. Six years ago, I analyzed 42 endgames with 4 pieces each by former world champion Chegolew (nicknamed "Slawa") viewtopic.php?f=65&t=2599 out of these 42 games, 41 ended in a draw. Now this world champion was renowned for his endgame technique, but he hardly could win such endgames under the current rules. In Killer on the other hand, I estimate that the drawing percentage for these same games would have been around 45%. So Killer does not force a decisive result no matter what, it just rewards small but recognizable advantages (tempo, opposition) in a generally fair way.

However, if only 3 against 1 is a win, this defensive property is retained, to a degree. Arguably, it is important to retain such a defensive stratagem, because it makes the game deeper and more interesting. Otherwise, to lose a piece means that the game is theoretically lost. But today, should you lose a piece, you can continue and play an interesting game. This is an important argument.

I don't think that a large drawing margin adds anything to the strategic depth of a game. In draughts, being a piece behind you can only get two results: lose or draw, but hardly ever a win. A truly interesting game will have many different aspects that need to be counterbalanced: tempo, opposition, left/right balance, freedom of movement etc. Being ahead in some and behind in others will add uncertainty and interest.

But being a piece behind, should just lose, unless there is a positional compensation, and the Killer-light proposal does not add anything in that respect.

[Juri]

Juri, if I understand you correctly, you propose that the king-halt rule is used only when the captured king is last in a multiple sequence. I don't think that draught players can object against this rule change, because it is minimal and does not affect the character of the game.

So this means that three kings can always win against one king?

What do you think is a good name for this variant? "Dame Halt Draughts"? "Modern Polish Draughts"?

Mats Winther
http://two-paths.com/bg

Yes, I mean this. @the king-halt rule is used only when the captured king is last in a multiple sequence@

Any name is ok for me.
But the first one is hard to use in other languages. Second one - not clear why polish?)

Modern international 100 (64) draughts?

[Juri]

I don't understand the need for "Killer-light". The endgame rules will change so the opponents who want to preserve the endgame literature will not accept this anyway. But furthermore, the drawing margin is quite a bit bigger. At first glance, the endgame becomes similar to those with +/- draws. In any case, what is really being gained? In my opinion, there is not viable place for Killer-light.

Difference is that in the ranking you get 2 points and your opponent get - 0. And everybody has a winner today!)
the "light killer" it is much more than +/- draws.

The main idea don't change to much a logical course of a game.
If some one have advantage like 1 king + 3 pieces against 1 king + 1 piece it would be winning situation.
Very common situation nowadays, when someone have a big advantage and then the opponent sacrifice 2 pieces to escape.. Even in a "light-killer" it is impossible, you will lose anyway.

Very important to save a natural, logical transition from second middle-game to end-game.
And I suppose that in a "full killer" it can brings some chaos.

I played two games with more or less strong player. I played fast, try to change more pieces in orderto go faster to end-game. Finally 2 times we got result (1 time lost, 1 time won), but with situation in a game before (in second middle-game) it didnt have any connection. Something happend and someone lost, even hard to explane what happend and why someone lost) Nobody had an huge advantage before.

In a "full-killer" you can start to play from this position:

and if the level is different - have good chances to win)
--------

Second idea - to save at least small connection with endgames in normal draughts.


Look like here to win you need to use the well-known way. The king-hult rule cannot help do it easier.

the situation when you use king-hult in a multiple sequence more rare. So if you get the equal end-games when both have kings, to catch the opponent's king you need to spend 1 king + 1 piece (some kind of a catching-king construction).
In a "full killer" main line in this case means nothing. I can simply change king to king.


this end-game is winning.


and what about this one?

already have some theory and knowledge.

continue.

this is one also possible draw? weak peace 25...

To win end-game become more difficult. So if you get some advantage you still need to play precisely.
For example.


To win the follow position you need to know some useful idea, in full-killer nothing need to know. Easly can win.

Conclusion:
- in a full killer you need to calculate more and work till the deep end-game, but a thinking process in the end is quite different. Of course amount of winning positions is higher. 2 kings win 1.
- in a light version an transition to end-game is more logical and natural. 3 kings win 1 no matter who controls main line. So amount of winning positions will grow too, but less if compare "full killer", but probably enough.

[Rein Halbersma]

To win end-game become more difficult. So if you get some advantage you still need to play precisely.
For example.


To win the follow position you need to know some useful idea, in full-killer nothing need to know. Easly can win.

I don't understand: this engame is won in regular draught and killer-light because after (16-11-7-1 and 25-30-34-39) the sacrifice 50-44 6-1 wins, but in killer 50-11 gives a draw. What was your point?

In any case, your examples changed my mind about the beauty of killer-light, it requires quite some skill to think of new catching positions!

The main worry remains: my estimate of draws in top-level killer would be 40-50%. I think in killer light it would be a lot higher, although of course less than with regular draughts. I think endgame databases for killer light should be made as soon as possible to determine its merits.

Maybe Alex Moiseyev can enter some opinion here, eg about the parallels with checkers opening ballots (first 2-moive, then 3-move, now 11-men). Changing the rules every time human skill evolution gives 90% draws does not seem very attractive. I'd prefer a clean break and have many decades of unchanged rules after that.