Breakthrough draughts

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Breakthrough draughts

Post by TAILLE » Wed Apr 04, 2018 19:51

Hi,

Do you know the "breakthrough" draughts variant (I do not know if this is the correct wording)? You take the standard rules of the international draughts game but you decide that the first side with a king wins the game. With this rule the game becomes extremely difficult because you can see that the draw does not exist.

For your information Damy has solved this game for a 8x8 board.
In the intitial position the move 22-18!! is the only winning move; the seven other moves are losing moves.

Position after 22-18!!
Image
Gérard

GuidoB
Posts: 2325
Joined: Sun Sep 21, 2003 21:43

Re: Breakthrough draughts

Post by GuidoB » Wed Apr 04, 2018 19:58

:lol:

You really made my day. This is an absolutely brilliant discovery. Congratulations to Damy!

User avatar
Marcel Kosters
Posts: 2244
Joined: Mon Jan 20, 2003 16:13
Location: Haarlem, The Netherlands

Re: Breakthrough draughts

Post by Marcel Kosters » Wed Apr 04, 2018 20:10

Cool, Gérard!

Is it also possible to research other game rule varieties with Damy? Such as for example the 'Killer King' in 10x10 international draughts? (If a king takes another king, he must stop on the square directly behind the taken king).
Nu: 28.200 DamZ! DamSets. Doel: 100.000 DamZ! DamSets zo snel mogelijk. Doe mee met 1, 10 of 100 DamZ! DamSets. De strategie is onbeperkt schaalbaar. Met club of vrienden snel besteld via marcelkosters@gmail.com.

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Breakthrough draughts

Post by TAILLE » Wed Apr 04, 2018 20:21

Marcel Kosters wrote:Cool, Gérard!

Is it also possible to research other game rule varieties with Damy? Such as for example the 'Killer King' in 10x10 international draughts? (If a king takes another king, he must stop on the square directly behind the taken king).
I have programmed only the 10x10 international draugths game and the 8x8 and 10x10 breakthrough draugths games.
As an addition information, for solving the 8x8 breakthrough draughts game I had to generate the 16 pieces endgame database which needed a lot of time, firstly to find an efficient compression mechanism and secondly to generate the database itself.
Gérard

User avatar
Marcel Kosters
Posts: 2244
Joined: Mon Jan 20, 2003 16:13
Location: Haarlem, The Netherlands

Re: Breakthrough draughts

Post by Marcel Kosters » Wed Apr 04, 2018 22:03

TAILLE wrote:
Marcel Kosters wrote:Cool, Gérard!

Is it also possible to research other game rule varieties with Damy? Such as for example the 'Killer King' in 10x10 international draughts? (If a king takes another king, he must stop on the square directly behind the taken king).
I have programmed only the 10x10 international draugths game and the 8x8 and 10x10 breakthrough draugths games.
As an addition information, for solving the 8x8 breakthrough draughts game I had to generate the 16 pieces endgame database which needed a lot of time, firstly to find an efficient compression mechanism and secondly to generate the database itself.
16 pieces database? Wow! Although the board of course is 8x8 that still must be an incredibly big database! How does the size in terms of GB compare to the state of the art in 10x10 endgame databases?

To motivate my earlier question: as a sports consumer and not as a scientist, I'm particularly interested in computer research into 10x10 varieties that do not touch the attractiveness of the current main objectives (to take all your opponents or to block all your opponents) of the 10x10 game, but that with simple and targeted refinement do drastically decrease the likelihood of a draw between absolute top level players.
Nu: 28.200 DamZ! DamSets. Doel: 100.000 DamZ! DamSets zo snel mogelijk. Doe mee met 1, 10 of 100 DamZ! DamSets. De strategie is onbeperkt schaalbaar. Met club of vrienden snel besteld via marcelkosters@gmail.com.

Wieger Wesselink
Posts: 1157
Joined: Sat Jun 28, 2003 13:22
Location: Eindhoven, The Netherlands
Contact:

Re: Breakthrough draughts

Post by Wieger Wesselink » Wed Apr 04, 2018 22:47

TAILLE wrote:Hi,

Do you know the "breakthrough" draughts variant (I do not know if this is the correct wording)? You take the standard rules of the international draughts game but you decide that the first side with a king wins the game. With this rule the game becomes extremely difficult because you can see that the draw does not exist.

For your information Damy has solved this game for a 8x8 board.
In the intitial position the move 22-18!! is the only winning move; the seven other moves are losing moves.
Congratulations Gérard with this impressive result!

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Breakthrough draughts

Post by TAILLE » Wed Apr 04, 2018 23:42

Marcel Kosters wrote:
TAILLE wrote:
Marcel Kosters wrote:Cool, Gérard!

Is it also possible to research other game rule varieties with Damy? Such as for example the 'Killer King' in 10x10 international draughts? (If a king takes another king, he must stop on the square directly behind the taken king).
I have programmed only the 10x10 international draugths game and the 8x8 and 10x10 breakthrough draugths games.
As an addition information, for solving the 8x8 breakthrough draughts game I had to generate the 16 pieces endgame database which needed a lot of time, firstly to find an efficient compression mechanism and secondly to generate the database itself.
16 pieces database? Wow! Although the board of course is 8x8 that still must be an incredibly big database! How does the size in terms of GB compare to the state of the art in 10x10 endgame databases?

To motivate my earlier question: as a sports consumer and not as a scientist, I'm particularly interested in computer research into 10x10 varieties that do not touch the attractiveness of the current main objectives (to take all your opponents or to block all your opponents) of the 10x10 game, but that with simple and targeted refinement do drastically decrease the likelihood of a draw between absolute top level players.
My 16 pieces end game db for 8x8 breakthrough draughts takes "only" 160 Gb which is not so big comparing to 470 Gb of my 8 pieces egdb for the 10x10 international draughts game but I have to say that, due to the particularities of the breakthrough draughts I was able to have a far better compression!

Yes, my view is that the current international draughts game is certainly a drawish game. I do not have a statistics but I will not surprise to see that 90% of the games between GMI will be a draw. Between the best programs the number of draws is really very high and in this context there no great motivation to try and improve a program for this game. That is why I changed for this breakthrough draughts game which appears very interesting and very difficult.

Let me give you an exemple of play in breakthrough draughts:

Image
White to play and win

1.34-30! (the only winning move) 17-21 2.28-22! (still the only winning move) 14-20 3.30-25 (still the only winning move) 21-26 4.25x14 19x10 (diagram)

Image
White to play

and now white as a typical combinaison in this kind of game
5.22-17 12x21 6.33-28 13-19 7.35-30 24x35 8.27-22 18x27 9.38-33 27x29 10.39-34 23x41 34x5 W+

Isn't it a beautiful and highly tactical game?
Gérard

User avatar
Marcel Kosters
Posts: 2244
Joined: Mon Jan 20, 2003 16:13
Location: Haarlem, The Netherlands

Re: Breakthrough draughts

Post by Marcel Kosters » Thu Apr 05, 2018 00:41

TAILLE wrote:My 16 pieces end game db for 8x8 breakthrough draughts takes "only" 160 Gb which is not so big comparing to 470 Gb of my 8 pieces egdb for the 10x10 international draughts game but I have to say that, due to the particularities of the breakthrough draughts I was able to have a far better compression!
Indeed, as Wieger expressed, congratulations for this achievement are in order!
TAILLE wrote:Yes, my view is that the current international draughts game is certainly a drawish game. I do not have a statistics but I will not surprise to see that 90% of the games between GMI will be a draw. Between the best programs the number of draws is really very high and in this context there no great motivation to try and improve a program for this game.
Last month, I spoke with Jaap Bus who as you know is one of the two motors behind the organisation of the computer draughts competitions. According to Jaap, the strongest 4 programmes nowadays literally never beat each other anymore. And indeed, the strongest grandmasters with the games between them are not far behind those 4 programmes.
TAILLE wrote:That is why I changed for this breakthrough draughts game which appears very interesting and very difficult.
Cool that you searched for an alternative.
TAILLE wrote:Let me give you an exemple of play in breakthrough draughts:

Image
White to play and win

1.34-30! (the only winning move) 17-21 2.28-22! (still the only winning move) 14-20 3.30-25 (still the only winning move) 21-26 4.25x14 19x10 (diagram)

Image
White to play

and now white as a typical combinaison in this kind of game
5.22-17 12x21 6.33-28 13-19 7.35-30 24x35 8.27-22 18x27 9.38-33 27x29 10.39-34 23x41 34x5 W+

Isn't it a beautiful and highly tactical game?
A strong aspect of 'your' rule in my opinion is, that it is simple and just added on top of the existing rules. Frankly though, I'm having some trouble with the premise that scoring a 'touch down' at all cost should supersede any strategic or material principle that does apply to the rest of the game. It also eliminates the in my opinion quite elegant goal of the current game, which is capturing or blocking all pieces of your opponent. Maybe we should call this variety the Pyrrus King. But that is my (first) impression and from the perspective of a (mind)sports consumer of limited draughts strength. And also, based on just one example.
Nu: 28.200 DamZ! DamSets. Doel: 100.000 DamZ! DamSets zo snel mogelijk. Doe mee met 1, 10 of 100 DamZ! DamSets. De strategie is onbeperkt schaalbaar. Met club of vrienden snel besteld via marcelkosters@gmail.com.

User avatar
Jan Pieter
Posts: 2432
Joined: Sat Sep 27, 2003 05:37
Real name: Jan Pieter Drost

Re: Breakthrough draughts

Post by Jan Pieter » Thu Apr 05, 2018 11:24

In 2003 we already played an experimental tournament with this rule, but with one difference: after a breakthrough the game continues, and the player with the king only wins when he survives. To put it in other words: a regular game is played and only when there is no winner, the victory goes to the player who first got a king. In this way draws are also impossible but the game stands much closer to draughts as we know it.
I think it also is more of a challenge for programmers. The breakthrough draughts of Gérard merely seems a matter of brute force, since games won't last long.

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Breakthrough draughts

Post by TAILLE » Thu Apr 05, 2018 11:38

Marcel Kosters wrote:Last month, I spoke with Jaap Bus who as you know is one of the two motors behind the organisation of the computer draughts competitions. According to Jaap, the strongest 4 programmes nowadays literally never beat each other anymore. And indeed, the strongest grandmasters with the games between them are not far behind those 4 programmes.
Of course this is a terrible observation but because the great majority of players are not super GMI the international draughts game looks still very interesting. Even the possibility of a draw is not a problem because, as a consequence, the best payer will try to take risks and even be in a losing position before turning the table and win anyway after a mistake of their opponent.
This context is difficult : at the very top level (the 4 strongest programs or the 4 strongest GMI) the international draughts game is quit dead because almost all games will be a draw but in all other situations the game keeps attractive.
As far as I am concerned, as programmer of the quite strong Damy program, I prefer to switch to a different version.
Marcel Kosters wrote:A strong aspect of 'your' rule in my opinion is, that it is simple and just added on top of the existing rules. Frankly though, I'm having some trouble with the premise that scoring a 'touch down' at all cost should supersede any strategic or material principle that does apply to the rest of the game. It also eliminates the in my opinion quite elegant goal of the current game, which is capturing or blocking all pieces of your opponent. Maybe we should call this variety the Pyrrus King. But that is my (first) impression and from the perspective of a (mind)sports consumer of limited draughts strength. And also, based on just one example.
There are some similarities between breakthrough draughts and international draughts games but obviously, because there are no draw, the strategy and the tactics are necessarily different. Anyway I am quite sure of one point: I you take a top GMI and a weaker player in the international draughts game and you ask them to play a breakthrough draughts game the GMI will win due to their better analysis capacity and to due to their tactics ability.

Let me show you another typical example:

Image
White to play

In international draughts game this position is uninteresting because you know that each side will manage to have a king and because you know that 3 kings cannot win against 1 king the game will end with a draw.
In breakthrough game white can win by the quite elegant sequence:
1.32-27! (the only winning move) 19-23 2.34-30 (still the only winning move) 17-22 3.33-28!! and white wins the race to the king row.
Gérard

Casper van der Tak
Posts: 734
Joined: Fri Dec 19, 2003 02:05
Location: Beijing / Laveno-Mombello

Re: Breakthrough draughts

Post by Casper van der Tak » Thu Apr 05, 2018 12:14

First of all, Gerard, congratulations on the result, which is quite a feat! We can take it as an indicative result that 32-28 in our classical game is the best opening move, something I have suspected for a long time. :-)

I think the question of reducing draws is important. The simple rules, board and pieces of draughts can be a strength. We can explain our rules within 5 minutes, and after that, nobody should forget. If we can create a complex game from such simple rules, this is elegant and attractive. The problem is that if at the highest level games almost always end in draws, the simple rules work against us - it will just appear we are playing a simple game. Worse, a simple game for simple souls. With such a bad image it will be difficult to promote the game, because eventually, the negative feedback will cause most people to lose interest.

This is a reason I have been for quite some time in favor of alternative rules that would lower the percentage of draws (and I believe it is quite similar to the reasons why Henk promoted the Delft rules). There are different ways to go about it, including breakthrough, killer, alternative counting rules (resulting is small and big victories depending on material at the end), and different board sizes & forms (e.g. 9x11, 10x11, etc). One can even consider reducing time available, although to me this does not look like the right approach to make the game structurally more attractive: reducing the time to think seems a poor way to create a game of thought (there are other reasons for limiting time which may be more valid, for example that it could be more attractive for the spectator and more "modern", but I digress).

Within the "breakthrough" family, I like to keep the possibility of combinations that involve giving the opponent a king and then using the king during the combination. With some of the others, I also like keeping some of our old strategic principles, so that if you get a king at a high expense, which is then captured at low cost, you will in the end lose. One way of doing this is use the rule that the player who obtained the first king still on the board in the final position wins. Compact, simple rule.

I see draughts as a family of games with common principles and rule sets, with details differing between the family members. Tournaments and championships can then be organized in different members of the family. Similar say to billiards or bridge.

What I think needs to be done now is experimenting in actual GMI-only tournaments with these alternative rules, and find out what attracts the most media attention, is preferred by spectators, and can draw in the most funding. How to finance these tournaments? If there is interest here and people to work on it, I believe it should be easy to crowdfund an experimental tournament cycle, and I would be happy to work on that together with others.

I see this as complementing actions such as those by Marcel on the distribution of low cost game materials. Materials is important to popularize at the bottom, and image is important, to make the game more attractive and keep those we get interested.

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Breakthrough draughts

Post by TAILLE » Thu Apr 05, 2018 12:57

Casper van der Tak wrote:One way of doing this is use the rule that the player who obtained the first king still on the board in the final position wins. Compact, simple rule.
I am not sure of my understanding of your proposal.
In the following position, which is an obvious draw in international draugths, who is winning with your rule ?

Image
White to play
Gérard

Casper van der Tak
Posts: 734
Joined: Fri Dec 19, 2003 02:05
Location: Beijing / Laveno-Mombello

Re: Breakthrough draughts

Post by Casper van der Tak » Thu Apr 05, 2018 15:05

TAILLE wrote:
Casper van der Tak wrote:One way of doing this is use the rule that the player who obtained the first king still on the board in the final position wins. Compact, simple rule.
I am not sure of my understanding of your proposal.
In the following position, which is an obvious draw in international draugths, who is winning with your rule ?

Image
White to play
Black: 1) 10-5 41-46 2) 11/7 26-31 3) 7-1 31-37 4) 5x41 46x5 no further kings can be captured, and the king on 5 was earlier on the board than the one on 1 (move 1 vs move 3).

MenO
Posts: 1211
Joined: Thu Mar 17, 2005 16:32
Location: Phnom Penh

Re: Breakthrough draughts

Post by MenO » Thu Apr 05, 2018 15:16

Casper van der Tak wrote:
TAILLE wrote:
Casper van der Tak wrote:One way of doing this is use the rule that the player who obtained the first king still on the board in the final position wins. Compact, simple rule.
I am not sure of my understanding of your proposal.
In the following position, which is an obvious draw in international draugths, who is winning with your rule ?

Image
White to play
Black: 1) 10-5 41-46 2) 11/7 26-31 3) 7-1 31-37 4) 5x41 46x5 no further kings can be captured, and the king on 5 was earlier on the board than the one on 1 (move 1 vs move 3).
That seems quite complicated for long (macro-)endgames, since you'd have to remember which king is which. Jan-Pieter's rule is much simpler: when the final position is a draw with the classical rules, then the player who first made a king wins. This rule is just as compact, but far easier in practice.
(In Gérard's example, white would win.)

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Breakthrough draughts

Post by TAILLE » Thu Apr 05, 2018 16:09

MenO wrote:
Casper van der Tak wrote:
TAILLE wrote: I am not sure of my understanding of your proposal.
In the following position, which is an obvious draw in international draugths, who is winning with your rule ?

Image
White to play
Black: 1) 10-5 41-46 2) 11/7 26-31 3) 7-1 31-37 4) 5x41 46x5 no further kings can be captured, and the king on 5 was earlier on the board than the one on 1 (move 1 vs move 3).
That seems quite complicated for long (macro-)endgames, since you'd have to remember which king is which. Jan-Pieter's rule is much simpler: when the final position is a draw with the classical rules, then the player who first made a king wins. This rule is just as compact, but far easier in practice.
(In Gérard's example, white would win.)
I agree with MenO it is too complicated to remenber the order of kings.
Let's take for proof this other example:
Image
White to play

After 9-4 41-46 11-6 44-50 6-1 42-47 etc. you reach a position in which the different kings have not the same value. The king on 4 has very high value and black has to try to capture this king, even against 3 kings. But you have also to keep trace of kings on 46 and 50 because black can win only by capturing white king 5 but without losing kings 46 and 50.
BTW I guess nobody knows how this endgame will be finished!
Gérard

Post Reply