﻿﻿ MatPlus.Net

Website founded by
Milan Velimirović
in 2006

22:42 UTC
 ISC 2024

Remember me

 CHESS SOLVINGTournamentsRating lists1-Jul-2024
 B P C F

MatPlus.Net Forum General Construct the longest ser-x with a Black homebase

Page: [Previous] [Next] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

On another forum, (396) was used for a discussion about testing long problems, and Paul Raican pointed that it was a ser-h= with colors reversed.
I was confused. When chasing records, a composer can forget elementary principles...
Indeed, in a "true" ser-r, there should be at least one try failing because the side to move has reflex obligation to mate or stalemate. Otherwise, the problem is a ser-h with colors reversed in disguise.
(348),(273) and (396) belong to this category.
As a result, (273) can probably be computer-tested as a ser-h= with Popeye and option Intelligent (and some time...).

It happens that a new ser-r# record can be done using reflex tries :

dedicated to Paul Raican
(= 4+16 )
ser-r#13 C+
1.h4! (1.g6? 2.é×f7‡,g×f7‡) 2.h5 3.h6 4.h×g7 5.g×f8=B 6.g6 7.g×h7 8.h×g8=R! (8.h×g8=Q? 9.é×f7‡,Q×f7‡) 9.Rg2 10.B×é7 11.Bh4 12.Kh2 13.Kh3 Q×h4‡

but it is highly improbable, not to say certainly impossible, that the same can be done for a ser-r=...

@ “Indeed, in a "true" ser-r, there should be at least one try failing because the side to move has reflex obligation to mate or stalemate. Otherwise, the problem is a ser-h with colors reversed in disguise.”

Please see my earlier post (165) with a few comments from others thereafter.

Congratulations on the “real” Ser-r# above! Now we have a good start to get the longest…

Arno, I am afraid that with sooo many posts, I didn't read these ones and concentrated on the ones with records and cooks...
Indeed, you anticipated me (and Paul) on the diagnosis. But sorry there will be no change of dedication A Posteriori.
I knew this beforehand, same as for selfmates that are direct stalemates in disguise, etc. etc.. But somehow I got trapped, and felt guilty...
The defences are lower in the field of records.
Quality is not the priority when you accept Kings in check, adding ugly moves to nice sequences, etc... If a ser-r has no reflex try but more moves, it will be considered the record.

Yes, these standards are often ignored - seems this is not a problem as naturally you want always to find also the “real” length record, even if it is not so nice. For series-reflex stipulations the usual view is that records count only if they have at least one reflex-refutation.

One remark to F3745 in The Problemist, July 2022 - this ser-s#16 is by Michael Schreckenbach & Frank Richter.
Unfortunately, the co-author was ignored both when the problem was printed and now in the solution, although I pointed this out several times.

After a long time I would like to update the list, it's the 16th interim balance.

Michel posted one year ago two new records for ser-!= and ser-r=.
Somehow I overlooked that, sorry.

He posted in August a new record for ser-r#. It is a "real" series-reflex-mate,
because the problem is not a ser-h with colors reversed in disguise - a big achievement.

The ser-h=35 (420) as correction of (411) has been cooked, see
https://pdb.dieschwalbe.de/P1401782
The new record is a version without additional pieces to gain moves.

The ser-hs=18 (202) has been cooked, see (417) or
https://pdb.dieschwalbe.de/P1400972

Not in this forum, but in Strategems, another longest ser-s# was published,
see at the bottom.

It's hard to find the current record in section Ser-s= . It might be (113).
Several attempts surpass his problem but the might be cooked as well.

Current records are:

Ser-# ...11 ........... (39) Hilmar Ebert / Alexander Lehmkuhl (feenschach 1976)
Ser-= ...19 .......... (349) Arno Tüngler / Michael Schreckenbach
Ser-s# ..17 .......... (124) Frank Richter
Ser-s# ..17 .......... (128) Michael Schreckenbach
Ser-s# ..17 ................... Frank Richter / Olaf Jenkner
Ser-s= ..22 .......... (113) Michel Caillaud after D. Stojnić and M. Mladenović (C?)
Ser-r# ..13 .......... (421) Michel Caillaud
Ser-r= ..31 .......... (396) Michel Caillaud (C?)
Ser-!# ..14 .......... (325) Michael Schreckenbach / Frank Richter
Ser-!# MMM 16 .. (343) Michael Schreckenbach
Ser-!= ..30 .......... (386) Michel Caillaud (C?)
Ser-h# .KNC 11 .. (42) Michel Caillaud / Arno Tüngler
Ser-h# .. 15 ........ (222) Arno Tüngler / Miodrag Mladenović
Ser-h= .. 30 ........ (407) Michel Caillaud (C?)
Ser-hs# .13 ........ (331) Dragan Stojnić / Miodrag Mladenović (C?)
Ser-hs= .16 ........ (197) Arno Tüngler (C?)

Everybody is invited to surpass one of the entries or to cook one marked C?.
I am expecting some additional moves in (386) and (197).

All entries of the record list and some more problems can be seen and autoplayed in the PDB:
https://pdb.dieschwalbe.de/search.jsp?expression=creationuser=%27jenkner%27and+originals=%27MATPLUS%27and+k=%27hom%27

............. Ser-#11 is https://pdb.dieschwalbe.de/P1400918
The last Ser-s#17 is https://pdb.dieschwalbe.de/P1402343

Thanks for the update, Olaf. Hard to belive this thread exploded over a year ago.

This can not be the final list.
What about +Sa6 +Pa5 +Sc5 +Rc6 ser-h=34 (407)?

Excellent for new review

(420) (428)
Possibly a way to enhance the record by 1 move as well as to prevent the last cook is adding a unit on h4

(= 14+16 )
ser-h=36

1.a×b6 2.R×a6 3.R×a5 4.Ra4 5.R×ç4 6.R×d4 7.R×d5 8.R×ç5 9.R×ç6 10.Ré6 11.ç6 12.d5 13.d4 14.d3 15.d2 16.d1=R 17.Qd2 18.Qh6 19.Rd8 20.Nd7 21.ç5 22.ç4 23.ç3 24.ç2 25.ç1=B 26.Bg5 27.B×h4 28.Bf6 29.g5 30.Qg7 31.h6 32.Rh7 33.Qh8 34.B6g7 35.Rg6 36.f6 é6=

Capturing the wRh4 adds 1 move to the solution and 3 moves to the cook :
1.a×b6 2.R×a6 3.R×a5 4.Ra4 5.R×ç4 6.R×ç5 7.R×ç6 8.Ré6 9.ç5 10.ç4 11.ç3 12.ç2 13.ç1=B 14.Bg5 15.B×h4 16.Bf6 17.g5 18.Bh6 19.Bfg7 20.Nf6 21.0-0 22.d6 23.Qd7 24.Rd8 25.Qé8 26.Nbd7 27.Kf8 28.Ng8 29.Rg6 30.f6 31.Qf7 32.Ké8 33.Qf8 34.Bh8 35.Rg7 36.Rf7 37.B6g7 38.h6 é6=

(a wRd5 instead of a wQd5 is sufficient as after all the Queen did'nt prevent the cook...)

Ah, a new record if it keeps alive. It looks dangerous with pieces on the d-file.
Maybe Arno or somebody else will find a cooking recipe.

Okay, as Olaf asks so kindly…
(430) cooked by 1.Sf6 2.Sxd5 3.bxa6 4.Bb7 5.Sxb6 6.Sa4 7.Sxc5 8.Se6 9.dxc6 10.g5 11.gxh4 12.h3 13.h2 14.h1=B 15.Kd7 16.Kc8 17.c5 18.Bhc6 19.Bxb5 20.Be8 21.Rg8 22.Rxg4 23.Rxd4 24.Rd7 25.c6 26.Rc7 27.Kd7 28.Qc8 29.Bg7 30.Bf6 31.h6 32.Sd8 33.e6+ Kxf6=
Probably one or two moves can be spared. Makes use of the additional unit on h4 but would most likely be possible even without it. This can easily be avoided by removing the one or other piece that Michel added for length but then I will have another look…

Michel's 30-mover seems is likely correct and arguably the best problem in the thread. Still, I would also like to show the shortest ser-h= that I found so far - maybe we come still to a number of moves that can even be verified by a computer?? This one is at least okay when tested as ser-ha=>b 18.

(= 15+16 )

ser-h=18
1.a5 2.axb4 3.Ra5 4.Rb5 5.d5 6.d4 7.dxc3 8.Qxd2 9.Sd7 10.Re5 11.Rxe3 12.Rg3 13.Qxh6 14.Qg6 15.h6 16.Qh7 17.Rg6 18.f6 Bxc7=

A wonderful problem, in case of ser-h= the shortest setting is more interesting than the longest.
Unfortunately the titel of this thread is "Construct the longest ser-x ...".
If A=>B is correct, then a cook has to find another stalemate position. I have no idea how likely it is.
After Arno cooked (430) the list (426) is up to date until Michel adds some moves on his superb (407).

Well done, Arno! As usual.
Your new cook uses the 2 white units on a5,a6 so maybe it is also valid for one of the previous tries with this feature...
Getting rid of 1 of these 2 units gets also rid of the cook (and of 1 "record move"...).

(= 13+16 )
ser-h=35
Waiting for the next cook scheme?
Making a nice position ugly by adding pieces for the sake of a record is disheartening, and one doesn't wish to spend too much time on it...

Your "short length" record is very nice.
To my knowledge, the best tool to test this kind of problem is the intelligent mode of Popeye but probably only 13 or 14 last moves can be tested in the case of this problem...

In current issue of "Die Schwalbe" you can find an article "Serienzug-Selbstmatts mit schwarzer Homebase",
selfmate seriesmovers with black homebase. There are shown 91 problems with mates on each square of the board.
On 48 squares a selfmate with black homebase is possible. Unfortunately on 5 squares we could not avoid some
holes in the homebase.
For 5 squares the shortest and longest problem is the same.
We'd be glad if somebody will find a longer max-position or a shorter min-position or a more economic position
or fill some holes in incomplete homebases.

There are all 91 problems from the article:
http://pdb.dieschwalbe.de/search.jsp?expression=s=%27die%20schw%27and%20sourceno=%27322%27and%20k=%27hom:s%27

There are 48 shortest problems:
http://pdb.dieschwalbe.de/search.jsp?expression=s=%27die%20schw%27and%20sourceno=%27322%27and%20probno=%27min%27

There are 48 longest problems:
http://pdb.dieschwalbe.de/search.jsp?expression=s=%27die%20schw%27and%20sourceno=%27322%27and%20probno=%27max%27

First addition to the article, found before its publication:

b8max
(= 6+14 )

ser-s#10, C+

1.Sc5xb7 2.Qc1xc7 3.Qc7xb8 4.Qb8xa8 5.Kb5-a6 6.Ka6-a7 7.Ka7-b8 8.Bf2-a7 9.Rf1xf7 10.Sb7-d6 Bf8xd6#

I bet Michel or Arno will plug a hole.

Wondering whether the matrix of (222) could be adapted also to a ser-hs= with much less white force (and a new record length...)

(= 7+16 )

ser-hs=19 C?

Solution: 1.cxd6 2.Sxa6 3.Sc5 4.a5 5.a4 6.a3 7.Ra4 8.Qa5 9.Kd8 10.Kc7 11.Kb6 12.Kb5 13.Kb4 14.b5 15.Se4 16.Sxg3 17.Sxe2 18.Sxc1 19.Sb3 & 1.c3+ Kxc3=

EDIT No, this is cooked in just 11 moves… 1.cxd6 2.e5 3.Qg5 4.Qxg3 5.Qc3 6.Qxc2 7.Qxc1 8.bxa6 9.Ke7 10.Kf6 11.Kf5 & 1.e4+ K~=

Wow! What a thread. With so many contributing so many replies it is a treat to peruse. Thanks to James and all the other experts!