MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

23:01 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Oct-2024

B P C F





 
 
MatPlus.Net Forum General Triple Check (for Rewan)
 
You can only view this page!
Page: [Previous] [Next] 1 2 3
(21) Posted by Jacques Rotenberg [Friday, Oct 30, 2020 11:17]

it appears to be a very complex fairy condition.
in wchloe you have 15 problems registered (all from Probleemblad 1993 or 1994), only 7 of them being correct!
 
   
(Read Only)pid=19878
(22) Posted by Andrew Buchanan [Friday, Oct 30, 2020 13:55]

Hi Jacques: 4 of the 8 cooked problems from 93-94 are because the kings make a beeline for one another, and don't let go when adjacent. This is counter-intuitive but effective. Not really surprising that some composers didn't spot it.

I can fix one of the other 4 cooked problems: version Rainer Paslack 151 Probleemblad 1993, 2nd HM. EDIT: It turns out that a naughty bPc6 was present in the original, and not just a transcription error, and so here's the agreed version.

R.Paslack, corrected A.Buchanan 30-Oct-2020
(= 8+12 )
#2 Bosma C+

All this makes me think that the C+ #3 Meredith Aristocrat with James Malcom shown above is maybe better than we thought. We would have offered it to Julia's, but fellow Singapore-dweller James Quah solved it in about 2 seconds. Maybe he was "in the zone"? Has anyone here had a shot? :)
 
   
(Read Only)pid=19879
(23) Posted by Jacques Rotenberg [Friday, Oct 30, 2020 15:16]

you are right, it looks like a misprint, may someone ask the author ?
 
   
(Read Only)pid=19880
(24) Posted by James Malcom [Friday, Oct 30, 2020 16:55]

Better yet, does anyone has scanned copies of the relevant issues of Problemdaad, i.e. the one that announced the tourney and the one that had the results?
 
   
(Read Only)pid=19881
(25) Posted by Andrew Buchanan [Sunday, Nov 8, 2020 04:15]

The original "Bosma" was by Robert Norman: see https://pdb.dieschwalbe.de/P1377131:

Chess Magazine 06/1988
(= 3+4 )
White to play and win

Thanks so much James for your scholarship. Various new thoughts about this fascinating miniature.

First, see the obit of the author's father: https://www.keverelchess.com/exmouth/pioneers/3199/ For the benefit of non-Brits, let me say that the quiz show "Ask The Family" was a cultural fixture in grey 1970s Britain, when only three TV channels existed. The obit also shows again chess acting as a connector between a father and a son. And I am very sorry the composer died so young.

I visited Dar myself in August 1995 as part of a group with Fulbright scholar Dr Joseph Mbele, expert on tricksters and storytellers, who informs to this day my perspective on the anthropology of chess problem composition, as what are we if not trickster storytellers? But perhaps I digress.

Given that Robert Norman originated the trick later termed Bosma, it's a pity that it's not named after him. As far as I know, there is no sound Bosma composition by Bosma himself: he just invented the fairy variant! Is this correct? However, Bosma is a pleasant, distinctive term, to my English ear, and there we are.

Now for the chess. As an orthodox study, these days we know definitively by tablebase that the position is a draw. But in WinChloe, we can run the Bosma condition but realistically only as a directmate. As a #7 WinChloe gave:
1. g6? [2. Qe8#]
1. ... Sd7+ 2. Kf7+ Sf6+ 3. Kf8 [4. Qxf6+ Rg7 5. Qxg7#] Rf7+ 4. Kxf7 [5. Qxf6#] Bxg6+ 5. Kxg6
[6. Qe7 [7. Qg7#] Se8 7. Qh7/Qxe8/Qf8#
6. ... Sh5 7. Qh7/Qe8/Qd8/Qf8#
6. Qxf6+ Kg8 7. Qd8/Qg7#]
So there is no mention of the intended (short) solution after 2. ... Sxe5+ 3. g7+ Kh7 4. g8=Q+ Kh6 5. Qg7#
but 1. ... Kg8!
The point now is 2. Qe8+? Kf7+! 3. K~5 Kxe8! winning for Black. I have had WinChloe analyze both the original position, and that after 1. g6 Kg8 as #8, and there is no other solution or try. I think that as a study it's probably still a win for White, who following 1. ... Kg8 also has e.g. 2. Qd5/Qf6+ Kf7+ 3. K~5+ etc, but how long would it take as a directmate? Any ideas? And can anyone confirm definitively that the study is cook-free?

In other news, I've exchanged amicable emails with Rainer Paslack (thanks Hauke!) - and have edited an earlier post in this thread.
 
   
(Read Only)pid=19919
(26) Posted by Juraj Lörinc [Sunday, Nov 8, 2020 12:20]

Great thread, thanks for a good reading to everyone involved up to now.

I'd like to comment on two earlier positions.

#3 from post No.7

For me the fact that it is easy to solve would not be a reason not to publish. Let's face it, most fairy chess lovers do not solve all the great problems they come across. Rather it is usually the most pleasant to play through the solutions. (Says somebody knowing something about solving fairies - see pages 21-24 here: https://www.chessfed.gr/old_site/wccc2010/files/wccc_2010_bulletin_final.pdf)

Actually, I am rather amused by still frequent comments of authors of the complex fairies about their expectations directed to solvers... believe me, often it is rather difficult just to understand what is going on in solutions. And then, I can smell silicon from distance, being the author of both very silicon and computer-help-less problems. Most silicon problems are not suitable for human solving.

More important for considering the publication in this case is the content of the problem. From my viewpoint this is just a sequence of checks and cross-checks, using Bosma condition. It is more a scheme than a ready-to-publish threemover. My silicon friend with WinChloe says there is some potential with try 1.Qxd8+? Kg5! and some variations, even if dualistic. Would it be possible to turn it into something more complex, better organized, with variation play, using the line 1.D×d8+? Rf7 2.Df8+ Re7+ 3.Rc7‡/Rc6‡ with royal attraction and Bosma specific battery? I did not try myself, but there is a potential.

#2 from post No. 20

In the presented form I find it rather attractive. Let's look at the extract from WinChloe solution:

1.Fh2! blocus
1…Ta4 2.T×b5‡
1…Tb×c4+ 2.Rd4‡

1.c×b5? [2.Rd4‡]
1…Tbe4 2.b6‡
1…T×b5 2.T×b5‡
1…Tbd4!

With two pins, in the solution only Rb4 can move. 1...Ra4 is actually a random move, allowing simple mate 2.Rxb5#. Then 1...Rxc4! is a correction allowing Bosma mate.

But there is also a valuable try showing some more Bosma play. Opening the 5th rank makes 2.Kd4# known from solution a mate. Black can defend by removing one attack from d4. Moving Rb4 "anywhere" does the job, but then 2.b5# mates. Well, anywhere is specific here as vertical movement is limited and it is actually line closing of Rf4 that applies. 1...Rxb5 is a correction in a sense that is removes piece mating after "random" defence, but there is a new mate. Refutation on the threat square.

I admit the explanation of the try is a bit stretched. But impressive introductory problem in my view.
 
   
(Read Only)pid=19921
(27) Posted by Andrew Buchanan [Monday, Nov 9, 2020 01:53]

Hi Juraj,

Thanks for your detailed response. The #3 was not built by silicon, and took some time to assemble by hand. However it was frustrating in that the end result did not seem to give as much complexity to solve as the effort that went into making it. Most of the compositions I make are retro or h#, and any d# or s# content is secondary, and solving difficulty is not high in my objectives, but this one really didn’t seem to be challenging enough. However as aristocrat Meredith with no cook-stoppers it seemed worth giving some airing. I will have a look at adding tries, but first...

I have some questions about what is the actual Bosma condition:

(1) From the 1985 Laws: "Article 5. THE MOVES OF THE PIECES The King (a) Except when castling, the king moves to any adjoining square that is not attacked by an opponent's piece." This is separate from Article 9. CHECK 9.1 "The king is in check when the square it occupies is attacked by one or two of the opponent's pieces; in this case the latter is or are said to be "checking the king.' 9.2 Check must be parried by the move immediately following." This point may have been an assumption in earlier compositions, as e.g. it prevents kings from being adjacent. Hence many of the problems which WinChloe deems as cooked.

(2) More of a nit, but again from Article 5. "(f) Castling is prevented for the time being - (i) if the king's original square or the square which the king must cross over or that which it is to occupy is attacked by an opponent's piece," Does this prevent your 1.0-0?

Is there an exact statement of Bosma for the 1993 tourney? Can someone finally dig up the relevant copy of Probleemblad, please?
 
   
(Read Only)pid=19927
(28) Posted by Jan Hein Verduin [Monday, Nov 9, 2020 20:10]

 QUOTE 
Better yet, does anyone has scanned copies of the relevant issues of Problemdaad, i.e. the one that announced the tourney and the one that had the results?

I got them (not of the entire issues, just the relevant pages). Where do I send them?
 
   
(Read Only)pid=19931
(29) Posted by James Malcom [Monday, Nov 9, 2020 20:23]

Jan, you can send them to me at rewandemontay@gmail.com , if you will please.
 
 
(Read Only)pid=19932
(30) Posted by Juraj Lörinc [Monday, Nov 9, 2020 21:25]

I was not hinting that #3 is a silicon one, obviously rather the opposite. My rant on this topic was slightly misplaced, but only slightly - if complex fairy problems should be accessible to people, it should not be too difficult. If it means it is easy to solve, let it be easy.

Placing two kings next to each other is so counterintuitive for OTB players and even orthodox chess composers that grasping the consequences of some fairy conditions takes time. Köko and Isardam come to my mind as difficult examples, Grid chess and Patrol chess being more intuitive. Bosma seems rather difficult in more complex situations, but somewhere between two.
 
 
(Read Only)pid=19933
(31) Posted by Jan Hein Verduin [Wednesday, Nov 11, 2020 19:46]

 QUOTE 
Is there an exact statement of Bosma for the 1993 tourney? Can someone finally dig up the relevant copy of Probleemblad, please?

Andrew,

I found it, and am willing to send it to you if you're interested. But since I assume your dutch is worse than my english, I've made an attempt to translate the original article by Bosma here.


The rules of chess and their consequences
(Probleemblad may-june 1992)

R. Bosma
(= 3+7 )

This h#3 has multiple solutions. I invite you however to play throught the following one: 1.Rb8,Bc3 2.Se6+,Rg8+ 3.Kh7,Rh8 ‘Wait a minute’, you might say, ‘this move 2..;Rg8+ is impossible’. But then I’d like to direct you to article 9.1 of the rules of chess: “a king is in check if the square he is on is attacked by one or two enemy pieces”. This formulation excludes the possibility that a king is in check if he is attacked three times. The move 2..;Rg8+ is in accordance with rule 9.2: “a check must be parried on the next move”. With 2..;Rg8 the white king is attacked three times, which is, by article 9.1, not a check.
I write this to show that while fairy chess is shown in a number of forms that do not comply with the rules of chess, this problem (however strange it may look) does so. I’m sure never a problem has been published with such a theme. I take the liberty then to call it the Bosma-theme.

R. Bosma
(= 5+4 )

In this problem the idea is shown twice. Helpmate in four: 1.Rb8,Rh4 2.Bd7+,Rf8+ 3.Kh8!,Bh6+! 4.Kh7,Rh8. After white’s third move the black king is in check because white has reduced the number of attacks on him.

R. Bosma
(= 4+12 )

The third scheme shows how innocent a triple attack can be and how fatal a double one. Helpmate in three 1.Ba5,Rxb5 2.Sf3+,Rd1+ 3.Kb1,Ba4
From articles 9.2 and 9.3 it follows that a check must be parried immediately and that kings can never be captured.

(end of article)

note: I don't see how the second problem is mate (5.Kxh8), yet I copied the diagrams and moves (including the exclamation marks and lack of double checks) meticulously. However I couldn't find a correction in subsequent Probleemblad issues. Am I missing something here? I haven't checked any problem for soundness, btw.


A theme tourney with this idea was subsequently announced, of which the english text read "The Dutch Chess Problem Society asks for direct mates in 2, 3 or 4 moves and helpmates in 2 or 3 moves, in which a check is parried by getting the king in a position in which he is attacked three times".

First I note that it probably should have read attacked at least three times. Secondly, this theme did not just require Bosma chess, but specifically asked for a check being parried by getting the king in a position in which it is attacked three times. Imo quite a few problems in the award did not meet this extra requirement. (also the helpmates were limited to three moves, yet a h±4 got a commendation!)
 
   
(Read Only)pid=19964
(32) Posted by Andrew Buchanan [Thursday, Nov 12, 2020 18:09]

Dear Jan,

Thank you so much for translating and posting this information. It's good to know the origin of this fairy condition. WinChloe reveals that all three of Bosma's example problems are all very cooked, as Bosma suggested. Only the first one actually completed in WinChloe, with only 1893 solutions. I think that you are right that the “solution" in the second is not a checkmate.

Bosma observes that any genuine check must be responded to. However, that doesn't stop a king being captured, and we should specify that. I think we should further tidy the condition to say that the rule of three also applies to a king moving regularly or castling, which a literal reading of the 1985 Laws might refute.

I agree with your other comments. It’s not the only time that a tourney in the end embraced fairy entries which did not meet the spec. Look at the Dardilly Kamikaze tourney in France recently! I will put the tourney entries to PDB. I won’t enter Bosma's three examples right now, because they seem to be intended more as unsound schemas than as compositions. If someone repairs them, then I'm happy to log them, but I think that long helpmates like these may be challenging to fix.
 
   
(Read Only)pid=19970
(33) Posted by Joost de Heer [Friday, Nov 13, 2020 09:54]

Bosma himself already wrote that these diagrams were more to show the idea than to present correct compositions:

 QUOTE 
This h#3 has multiple solutions

 QUOTE 
The third scheme shows

 
   
(Read Only)pid=19971
(34) Posted by Ulrich Voigt [Friday, Nov 13, 2020 15:40]

Is this a legitimate #1?

(= 3+7 )


How about this one?

(= 6+5 )


It would be nice to have an exact definition of the Bosma condition.
 
   
(Read Only)pid=19972
(35) Posted by Joost de Heer [Friday, Nov 13, 2020 15:53]

Yes and Yes.
The definition given in Winchloe: A king cannot be captured on a square where it is threatened by at least 3 pieces.
 
   
(Read Only)pid=19973
(36) Posted by Ulrich Voigt [Friday, Nov 13, 2020 16:04]

How does that help with the second example?

This is exactly my point - your quote is not an exact definition. "The king cannot be captured" says nothing about castling rights. The current FIDE Laws specify that castling is prevented temporarily "if the square on which the king stands, or the square which it must cross, or the square which it is to occupy, is attacked by one or more of the opponent's pieces", similar to the definition of a king being in check. However, it is not at all clear that the "one or two" Bosma condition applies to castling as well.
 
   
(Read Only)pid=19974
(37) Posted by Hauke Reddmann [Friday, Nov 13, 2020 17:24]

@Ulrich: The writers of the rules, obviously no
problemist trolls :-), applied logic: a castling
field to be crossed CAN be attacked by more than two
enemy units, thus "one or more", while in a
"normal" chess game, there are no triple checks
and they thus lacked the imagination someone could
voluntarily run into one.
 
   
(Read Only)pid=19975
(38) Posted by Andrew Buchanan [Friday, Nov 13, 2020 19:46]

@Ulrich:
In post #27 of this heroic thread, I described two holes in the Bosma condition - based upon reading the 1985 FIDE Laws. The castling question is the second hole. I wanted to see if Bosma in his original article had provided for these cases, but it turns out he hadn't. Since then, I've had a chance with check James' compositions that under WinChloe's Bosma, the Rule of Three governs *all* royal encounters with enemy force. This is obviously the simplest solution for this tiresome but somehow appealing condition. I have so clarified in PDB too. If anyone objects, please let me know.
 
   
(Read Only)pid=19976
(39) Posted by Andrew Buchanan [Saturday, Nov 14, 2020 13:19]

I have fixed two more of the cooked entries in the Probleemblad tourney.

The first one is simple and clean:

I.Murarasu corr A.Buchanan
(= 14+10 )
#2 Bosma

The second problem is much more complicated thematically, and to keep all the play, the only solution I can find does have a flight-taking key: Tg4-g6. On the other hand in Bosma every piece comes with a downside, and Black can use wTg6 as a defence, always threatening e6-e5 when bK has shifted to d6. Preventing this resource drives solution uniqueness, and forces the place exchange of bK & wS.

The original intended key 1.Dc7-b8 was too strong in a different sense: by unblocking a fourth attacker on c5, it protects wK from almost all checks. Thus effectively the key can be viewed as just addressing unprovided checks, which helps make it nearly sound. I find flight-taking a lighter defect than "unprovided checks".

I chose g4 as the new start square for wT, because if it remains there, it would become a third guard of f4, after the line from wQ opens up. This ensures uniqueness against potential cooks like 1.wTh5xh6.

I have to leave intact the core matrix which delivers the amazing choreography more like a helpmate than a directmate. I am guessing that this problem was only unawarded because the cook was identified during judgement. If it has been sound at the time, it might have seen clear first place. The set of cool features and themes which the WinChloe engine automatically identifies fills three columns.

D.Wissmann corr. A.Buchanan
(= 10+14 )
#3 Bosma

C+ under WinChloe:
1.Tg6! [2.C2ç4+ F×ç4 3.Rd5‡
2…R×d6 3.Cé5‡]
1…F×d2 2.Cf7+ f4 3.Cd6‡
2…ç2 3.Cg5‡
2…Rd6 3.Cé5‡
1…T×d2 2.Cç4+ f4 3.Cb2‡
2…ç2 3.Cd6‡
2…Rd6 3.Cé5‡

If someone can put me in touch with Dolf, I would really appreciate.

The other five cooked tourney entries, including one by Bosma, in WinChloe seem more fundamentally broken. Before looking in detail to fix them, I would like someone to confirm the published diagrams, please. Or just send me scans of the relevant pages, even in Dutch.
 
   
(Read Only)pid=19979
(40) Posted by Andrew Buchanan [Saturday, Nov 14, 2020 16:52]

I have amended the definition of Bosma theme in PDB to the following. Comments welcome.

 QUOTE 
Under a literal reading of Article 9.1 of the FIDE laws of chess from 1985-1993, leaving one's king in a square under attack by three or more or the opponent's pieces was legal.

This inspired Rinke Bosma in 1992 to define a fairy theme, which has recently (2020) been rendered more robust (see Matplus.net discussion). Basically it states:

(1) A king in a square under attack by three or more of the opponent's pieces is not considered to be checked. This applies for *all* relevant purposes: king entering or remaining in a square, pins, castling start, transit and end squares, checkmate, stalemate and anything else.
(2) A king cannot be captured (this needs to be stated despite Article 9.2).
This interpretation matches that implemented in WinChloe.

A few Bosma problems are not fairies but jokes, and do not mention the theme in the stipulation. The first of these was by Dr Robert Norman in 1988. They follow the general Rule of Three Units, but in principle such jokes allows space for to take other related 1985 rules too literally:
- Article 5a says a king cannot *enter* a square if there are any number of attackers
- Article 5f says that a castling king cannot transit or end in a square with any number of attackers >=1.
- Despite what Article 9.2 seems to hint, kings might be legally captured.

 
   
(Read Only)pid=19982

Read more...
Page: [Previous] [Next] 1 2 3

MatPlus.Net Forum General Triple Check (for Rewan)