A famous classical position

[Luteijn]

You are right, but Yesterday I discussed the position with Anton Kosior. He pointed out the possibility 14.37-32 39-43 15.15-10 33-39 16.32-27 21x32 17.10-5 and the piece of black on 32 gets lost. After 17...32-38 18.36-47 =
...

After 14.37-32 39-43 15.15-10 33-39 16.32-27 21x32 17.10-5 32-38 18.36-47 black wins by the small combinaison 18...43-49 19.47x06 49-27B+

It is a pity, that it seems to be, that I can't proof, that the position is draw. May be stronger players or computers do a better job. In the other way it is quiet an event. More than 40 years theoretici thought, that Dussaut was always remise except some very special cases. The same position with piece 11 is possible on other fields. Simular conclusion are possible. A lot of players try to avoid this kind positions because of Dussaut. May be playing this kind of classical positions become more popular.

[TAILLE]

You are right, but Yesterday I discussed the position with Anton Kosior. He pointed out the possibility 14.37-32 39-43 15.15-10 33-39 16.32-27 21x32 17.10-5 and the piece of black on 32 gets lost. After 17...32-38 18.36-47 =
...

After 14.37-32 39-43 15.15-10 33-39 16.32-27 21x32 17.10-5 32-38 18.36-47 black wins by the small combinaison 18...43-49 19.47x06 49-27B+

It is a pity, that it seems to be, that I can't proof, that the position is draw. May be stronger players or computers do a better job. In the other way it is quiet an event. More than 40 years theoretici thought, that Dussaut was always remise except some very special cases. The same position with piece 11 is possible on other fields. Simular conclusion are possible. A lot of players try to avoid this kind positions because of Dussaut. May be playing this kind of classical positions become more popular.

Why is it a pity ?
A lot of players think there are too many draws in draugths and this is due in particular to the reputation of the classic game ! In this sense it would be a good news but the demonstration has to be complete !
It is a pity of course that Otgonbayar did not play Dussault gambit. I would have been very interested by Podolski moves !!

[Rein Halbersma]

After 14.37-32 39-43 15.15-10 33-39 16.32-27 21x32 17.10-5 32-38 18.36-47 black wins by the small combinaison 18...43-49 19.47x06 49-27B+

It is a pity, that it seems to be, that I can't proof, that the position is draw. May be stronger players or computers do a better job. In the other way it is quiet an event. More than 40 years theoretici thought, that Dussaut was always remise except some very special cases. The same position with piece 11 is possible on other fields. Simular conclusion are possible. A lot of players try to avoid this kind positions because of Dussaut. May be playing this kind of classical positions become more popular.

Why is it a pity ?
A lot of players think there are too many draws in draugths and this is due in particular to the reputation of the classic game ! In this sense it would be a good news but the demonstration has to be complete !
It is a pity of course that Otgonbayar did not play Dussault gambit. I would have been very interested by Podolski moves !!

The Chinook team (for the 8x8 checkers) variant currently has 39 trillion (39 x 10^12) positions which includes all 10-piece endgames. They are also busy computing the 6 vs 5 and 6 vs 5 checkers only endgames with a new technique called partial information databases. It is expected that they will solve the game of 8x8 checkers by the end of this year.

Solving complicated 10x10 draughts endgames is within range of current desktop computers. In 1998 or so, Stef Keetman already constructed 7-piece endgame databases. With modern equipment, one could generate up to the 8-piece endgames. It's only about 23 x 10^12 positions (slightly more than half Chinook has).

So all 5 vs 3 and 4 vs 4 endgames should be solvable! This would probably put an end to the discussion of all the classical 10 vs 10 positions (Woldouby, De Haas-Fabre, Otgonbayar-Podolski)

[TAILLE]

Hi Gerard,

I just discovered today your post on this BBS about the 'famous classical position'. It caught my attention because I am working on a program that plays 10x10 draughts, and I started analyzing the position with my program. Allow me to suggest an alternative approach to draw, which is different from the line being discussed at the very first move from the famous position.

[FEN "W:W25,27,28,30,32,33,34,35,37,38:B11,12,13,14,16,18,19,21,23,24."]
1. 27-22 18-27 2. 34-29 23-34 3. 30-39 12-18(A) 4. 39-34 18-23 5. 35-30 24-35 6. 33-29 13-18 7. 28-22 35-40 8. 22-24 40-44 9. 29-18 44-49 10. 24-20 14-19 11. 18-12 21-26 12. 32-21 26-8 13. 37-32 19-23 14. 38-33 49-27 15. 20-15 27-4 16. 33-29 23-28 17. 25-20 28-32 18. 20-14 32-38 19. 34-30 38-43 20. 14-10 43-48 21. 30-24 11-17 22. 10-5 48-39 23. 5-46 39-25 24. 29-23 25-48 25. 23-19 48-43 draw

(A) 21-26 4. 32-21 26-17 5. 39-34 24-30 6. 35-24 19-39 7. 33-44 16-21 8. 37-31 12-18 9. 38-33 18-22 10. 31-26 21-27 11. 44-40 13-19 12. 40-35 draw

I tried to post this as a reply in the thread on the Multilingual forum, but for some reason I am not permitted to post replies there even though I have a valid login for the BBS. So forgive me for sending this to you directly.

Do you have any comments about these moves, and would you be interested to explore some alternatives along these lines?

-- Ed Gilbert

I agree entirely with you. In the following position

I think white can obtain the draw by 14.38-33 49x27 15.20-15 and if 27-36 16.33-29 16-21 17.29x18 36x04 34-29/30 =

For the moment my own synthesis is thus the following :

The theory may become now :
1.34-29 23x34 2.30x39 18-23 3.39-34 12-18 4.35-30 24x35 5.27-22 18x27 6.33-29 13-18 7.28-22!! (7.29-24? 19x39 8.28x10 35-40!! 9.25-20 18-22!! with good winning chances) 35-40 8.22x24 40-44 9.29x18 44-49 10.24-20 14-19 11.18-12 21-26 12.32x21 26x08 13.37-32 19-23 14.38-33!! 49x27 15.20-15 27-36 16.33-29 16-21 17.29x18 36x04 34-29/30 =

[Wieger Wesselink]

I finally managed to have a look at this position again. And I must say I'm impressed by the new insights found using draughts computers. I always considered the variant 7.28-22 as inferior to 7.29-24. But the variant given by Ed Gilbert is very convincing. After 15... 27-4 16.33-29 23-28 17.25-20 28-32 18.20-14 32-38 19.14-10 38-43 20.34-30 43-48 21.30-24 the position indeed looks like a draw. The white plan is to play 10-5-41 followed by 24-19 and if needed 41-47, 15-10. I don't see what black can do about it. Neither do I see an alternative earlier in the variant.

And the unexpected move 8... 35-40! that was noted by Gérard Taille seems to refute the move 7.29-24. I'm not 100% sure that it is lost for white, but it is no longer clear to me how white should continue to make a draw.

So where I used to believe that 7.28-22 is lost and 7.29-24 is good for a draw, it's now exactly the other way around

[Rein Halbersma]

I finally managed to have a look at this position again. And I must say I'm impressed by the new insights found using draughts computers. I always considered the variant 7.28-22 as inferior to 7.29-24. But the variant given by Ed Gilbert is very convincing. After 15... 27-4 16.33-29 23-28 17.25-20 28-32 18.20-14 32-38 19.14-10 38-43 20.34-30 43-48 21.30-24 the position indeed looks like a draw. The white plan is to play 10-5-41 followed by 24-19 and if needed 41-47, 15-10. I don't see what black can do about it. Neither do I see an alternative earlier in the variant.

And the unexpected move 8... 35-40! that was noted by Gérard Taille seems to refute the move 7.29-24. I'm not 100% sure that it is lost for white, but it is no longer clear to me how white should continue to make a draw.

So where I used to believe that 7.28-22 is lost and 7.29-24 is good for a draw, it's now exactly the other way around

Perhaps Gerard and Ed could run their programs on the following two classical 10x10 positions? In particular it would be nice to know if the partial information endgame databases can be of added value.


Woldouby position: white to play (narrow draw for white??)


De Haas-Fabre position: white to play (narrow draw for black??)

[TAILLE]

Perhaps Gerard and Ed could run their programs on the following two classical 10x10 positions? In particular it would be nice to know if the partial information endgame databases can be of added value.


Woldouby position: white to play (probably lost for white)


De Haas-Fabre position: white to play (probably lost for black)

Hello Rein,
Your question is of course very interesting. My view is that these two positions are draw positions so, I you have a variant that seems to be a winning one I will be happy to discuss with you.
In any case these positions are so difficult that it is totally impossible for me to be sure of the result. Only discussion may improve the theorie. I am ready for that.
Gérard

[ildjarn]

De Haas-Fabre position: white to play (probably lost for black)

Paul Nitsch made an extensive study of this position. Unfortunately I can't remember his conclusion anymore.

[Vraagje]

His conclusion was draw (although I believe it was written by Frank Verdel).

I must have the article somewhere in hard copy format but I do not know if I am allowed to publish it and besides I do not have the possibility to scan it.

For examining Woldouby G. Jansens idea against H. Meijer might be chosen?

[Rein Halbersma]

But the variant given by Ed Gilbert is very convincing. After 15... 27-4 16.33-29 23-28 17.25-20 28-32 18.20-14 32-38 19.14-10 38-43 20.34-30 43-48 21.30-24 the position indeed looks like a draw.

I think Ed meant that the draw was an exact database draw. Not just a drawish evaluation.

So where I used to believe that 7.28-22 is lost and 7.29-24 is good for a draw, it's now exactly the other way around

Can anyone confirm the exact value of both defenses?

[Rein Halbersma]

Hello Rein,
Your question is of course very interesting. My view is that these two positions are draw positions so, I you have a variant that seems to be a winning one I will be happy to discuss with you.
In any case these positions are so difficult that it is totally impossible for me to be sure of the result. Only discussion may improve the theorie. I am ready for that.
Gérard

Well, I am optimistic and hope both position won't be a draw! For Woldouby, I would indeed take Gerard Jansen's win against Hein Meijer. For De Haas-Fabre, I have seen a 1994 analysis of Lex den Doop (correspondence association magazine) where he concludes to a win.

[Rein Halbersma]

Gérard Jansen werpt nieuw licht op Woldouby
de Volkskrant, Sport, 11 februari 2006 (pagina 44)
Ton Sijbrands



Zie diagram 1

Zo stond het na Jansen 44ste zet 42-38. Omdat zwart na 44...12-17?! 45.38-33! 17-21?? (veel meer verdediging biedt 45...26-31) 46.39-34 kansloos ten onder gaat, moet Meijer iets bijzonders ondernemen:

44...24-29 45.39-33 12-17 46.33x24 17-21! 47.38-33! 23-29 48.28-23! 19x39 49.24x44 18-23! 50.27-22 13-18! 51.22x13 14-19!! (conform de theoretische aanbevelingen) 52.13x24 21-27 53.32x21 16x27 54.24-20(!) 23-28 55.20-14 27-31 56.14-10! 31x42 57.10-5! 28-33 58.25-20!

De aanzet tot een stoutmoedig plan: wit gaat - uiteraard met het nodige beleid - 'gewoon' vijf dammen halen!

58...42-47 59.20-15! 33-38 60.44-39! 38-42 61.30-25! 47-36 62.39-34! 36-18 63.34-30 42-48 64.5-28 18-45 65.28-46 48-39 66.30-24 39-50 67.24-20 26-31 68.20-14 31-36 69.14-10 50-6 70.10-5 6-22 71.5-32 22-50 72.15-10 50-6 73.10-4 6-50 74.4-15 50-6 75.15-47 6-50 76.25-20 50-6 77.20-15 45-34 78.15-10 34-45 79.10-4 6-50 80.35-30 45-1 81.30-25 50-6 82.25-20 6-50 83.20-15 50-6 84.15-10 6-50 85.10-5

Dat het eindspel van vijf dammen tegen twee altijd wint, is inmiddels genoegzaam bekend. Maar naar een '5x2' met een extra schijf voor de minderheidspartij - een type eindspel dat zich tot dusver uitsluitend in Scholma-Bies, Westerhaar 1985, heeft voorgedaan - is bij mijn weten nog nimmer onderzoek verricht.

85...50-6 86.32-38! 6-50 87.5-37 50-6 88.4-27! 6-50 89.37-26! 50-6 90.46-5



Zie diagram 2

Omdat wit nu zowel na 90...6-50 als na 90...1-45 de beide zwarte dammen zou uitvangen (in het eerste geval doet hij 91.5-23!, 92.47x33! en 93.26x37 +, in het tweede 91.5-28! en 92.47x29! enz.), had Meijer niet beter dan met 90...36-41 in de overgang naar een zuivere 5x2 te berusten.

Dat eindspel zou Jansen (die de partij overigens fors had kunnen bekorten door niet 90.46-5 maar 90.46-41! te spelen...) zo'n 25 zetten later in winst omzetten. Vanuit de stelling D2,D3 versus D6,D29,D36,D37,D49 volgde namelijk 112.37-48!! (zonder vrees voor 112...2-19 113.29-20! + of 112...3-14 113.6-11! +) 112...3-26 113.6-11! 2x16 114.29-38! 16x43 115.49x32, waarna zwart het opgaf zonder 115...26-3 116.32-21 + af te wachten.

De fase tussen de 55ste en 90ste zet laat evenwel diverse vragen onbeantwoord. Zoals: a) had zwart nog kunnen ontsnappen door tijdig de 'tric-trac' te verlaten?; b) zou wit óók hebben gewonnen wanneer zwart schijf 26 níet naar veld 36 had overgebracht?; c) had zwart kunnen voorkomen dat wit al zijn schijven tot dam kroont?; d) was het na 55...27-32 (in plaats van het gespeelde 55...27-31) wél remise geworden? En de veruit belangrijkste vraag luidt natuurlijk: e) betekent dit heel misschien dat de Woldouby - of anders op z'n minst de vertakking die met 44...24-29 van start gaat - tóch gewonnen is?

Copyright: Sijbrands, T.

[Rein Halbersma]

On behalf of Ed Gilbert, who still has technical problems posting on this subforum: Woldouby = a win and De Haas-Fabre = a draw.

Here is some play from Kingsrow that can perhaps serve as a starting point
for further discussion. Since I played into a black win for the Woldouby
position, perhaps Gerard can show where he deviates in his analysis.

In the De Haas-Fabre position, Kingsrow never saw very much threat for black
and sees a database draw after 8 moves. If this is published as a probable
win, then we can see where published play deviates from the computer's.

[Event "Woldouby position"]
[FEN "W:W25,27,28,30,32,33,34,35,37,38:B12,13,14,16,18,19,21,23,24,26."]
1. 34-29 23-34 2. 30-39 18-23 3. 39-34 13-18 4. 25-20 14-25 5. 28-22 23-28
6. 32-14 21-41 7. 22-13 41-47 8. 35-30 24-35 9. 14-10 47-41 10. 10-4 41-14
11. 4-15 12-18 12. 13-22 {from this point forward white's position is dire}
14-9 13. 22-17 9-3 14. 17-11 16-7 15. 15-10 3-8 16. 10-32 8-24 17. 34-29
24-8 18. 29-23 35-40 19. 32-27 40-45 20. 27-32 26-31 21. 32-41 31-36 {black
database win}

[Event "De Haas-Fabre position"]
[FEN "W:W25,27,28,30,32,33,34,35,37,38:B13,14,15,16,18,19,21,23,24,26."]
1. 34-29 23-34 2. 30-39 18-23 3. 27-22 23-29 4. 39-34 29-40 5. 35-44 13-18
6. 22-13 19-8 7. 28-22 8-12 8. 32-28 14-19 {Kingsrow sees a database draw
starting here.} 9. 44-39 12-17 10. 22-11 16-7 11. 28-22 7-11 12. 22-18 11-17
13. 25-20 24-30 14. 20-14 19-10 15. 18-13 15-20 draw

-- Ed

[TAILLE]

On behalf of Ed Gilbert, who still has technical problems posting on this subforum: Woldouby = a win and De Haas-Fabre = a draw.

Here is some play from Kingsrow that can perhaps serve as a starting point
for further discussion. Since I played into a black win for the Woldouby
position, perhaps Gerard can show where he deviates in his analysis.

In the De Haas-Fabre position, Kingsrow never saw very much threat for black
and sees a database draw after 8 moves. If this is published as a probable
win, then we can see where published play deviates from the computer's.

[Event "Woldouby position"]
[FEN "W:W25,27,28,30,32,33,34,35,37,38:B12,13,14,16,18,19,21,23,24,26."]
1. 34-29 23-34 2. 30-39 18-23 3. 39-34 13-18 4. 25-20 14-25 5. 28-22 23-28
6. 32-14 21-41 7. 22-13 41-47 8. 35-30 24-35 9. 14-10 47-41 10. 10-4 41-14
11. 4-15 12-18 12. 13-22 {from this point forward white's position is dire}
14-9 13. 22-17 9-3 14. 17-11 16-7 15. 15-10 3-8 16. 10-32 8-24 17. 34-29
24-8 18. 29-23 35-40 19. 32-27 40-45 20. 27-32 26-31 21. 32-41 31-36 {black
database win}

[Event "De Haas-Fabre position"]
[FEN "W:W25,27,28,30,32,33,34,35,37,38:B13,14,15,16,18,19,21,23,24,26."]
1. 34-29 23-34 2. 30-39 18-23 3. 27-22 23-29 4. 39-34 29-40 5. 35-44 13-18
6. 22-13 19-8 7. 28-22 8-12 8. 32-28 14-19 {Kingsrow sees a database draw
starting here.} 9. 44-39 12-17 10. 22-11 16-7 11. 28-22 7-11 12. 22-18 11-17
13. 25-20 24-30 14. 20-14 19-10 15. 18-13 15-20 draw

-- Ed

Concerning De Haas-Fabre position I have exactly the same analysis and my conclusion is then a draw
Why now about Woldouby position ?
My analysis begin with the very famous G. Jansens - H. Meijer game.
Let's begin with the following starting position :

Black to play
The game continued by :
44...24-29 45.39-33 12-17 46.33x24 17-21 47.38-33 23-29 48.28-23 19x39 49.24x44 18-23 50.27-22 13-18 51.22x13 14-19 52.13x24 21-27 53.32x21 16x27 54.24-20 23-28 55.20-14 reaching the following position

Black to play
In the game black played 55...27-31 followed by a convincing win for white.
What about 55...27-32. Can white manage to make 5 kings while avoiding black to make 3 kings ? For the time being I did not find such way in order to win. That is the reason why why I am not sure that Woldouby position is a winning position.
Maybe Ed. can help us.
Gérard

[Rein Halbersma]

Concerning De Haas-Fabre position I have exactly the same analysis and my conclusion is then a draw

After 1.34-29 23x34 2.39x30 18-23 3.27-22 23-29 there is also the variation 4.37-31 instead of 4.39-34 in De Haas-Fabre. This gives also some troubles for black. What do you play there to draw for black?

Why now about Woldouby position ?
My analysis begin with the very famous G. Jansens - H. Meijer game.
Let's begin with the following starting position :

Black to play
The game continued by :
44...24-29 45.39-33 12-17 46.33x24 17-21 47.38-33 23-29 48.28-23 19x39 49.24x44 18-23 50.27-22 13-18 51.22x13 14-19 52.13x24 21-27 53.32x21 16x27 54.24-20 23-28 55.20-14 reaching the following position

Black to play
In the game black played 55...27-31 followed by a convincing win for white.
What about 55...27-32. Can white manage to make 5 kings while avoiding black to make 3 kings ? For the time being I did not find such way in order to win. That is the reason why why I am not sure that Woldouby position is a winning position.
Maybe Ed. can help us.
Gérard

Gérard, your question is the same as Sijbrands asked in his article. I got the following line from Ed on the defense 3.27-22 (instead of 3.39-34 which seems lost):

[Event "Woldouby position"]
[FEN "W:W25,27,28,30,32,33,34,35,37,38:B12,13,14,16,18,19,21,23,24,26."]

1. 34-29 23-34 2. 30-39 18-23 3. 27-22 12-18 4. 39-34 18-27 5. 34-30 13-18 6. 28-22 23-28 7. 32-12 27-7 8. 38-32 {Kingsrow sees a database draw from here forward} 7-12 9. 33-29 24-33 10. 30-24 19-30 11. 35-24 33-39 12. 24-20 14-19 13. 20-15 39-44 14. 15-10 44-50 15. 25-20 ... draw

-- Ed

Note that he plays 8.38-32 instead of the double sacrifice 8. 33-28 24-29 9. 38-33 and 10. 37-32. So perhaps Woldouby is indeed a draw. Just have to wait for the complete proof.