# No.352,352.1 (IS,IS&AT)

 No.352 Ivan Skoba (Czech Republic) No.352.1 Ivan Skoba (Czech Republic) &Arno Tüngler (Germany) Original Problems, Julia’s Fairies – 2013 (II): May – August  →Previous ; →Next ; →List 2013(II) Please send your original fairy problems to: julia@juliasfairies.com

No.352 by Ivan Skoba – Series autostalemate in 49 moves with conditions Madrasi, Consequent. Through one-side play, the problem seems rather difficult! (JV)

No.352.1 by Ivan Skoba & Arno Tüngler – An improvement to No.352, as few cooks were found by A.Tüngler in No.352: 1.Ke4 2.gxf6ep! (possible because of the consequent condition!) 4.fxg8=R 6.Rxc6 8.Rd4 9.g8=S 10.Se7 11.Sc8 12.Sb6 13.Sc4 14.d4!= ,   6.Kxa5 9.Kxd7 10.Kxc6 13.Kd4 and now 14.dxe6ep! (again possible due to consequent condition) 16.e8=S 20.Sg2 21.d3!=.  Authors write that it took them about a month to find this new position. Enjoy detailed comment to the problem! (JV)

Definitions by author:

Stipulation ser-!= X : Series autostalemate in X moves (here X=49).

Madrasi + Consequent:  According to conditions Madrasi + Consequent white moves in series in order to be stalemated.

Madrasi: Mutually attacking black and white units of the same type paralyse each other, so that they may no longer move or give check. The paralysis may be removed, for example by interference (in the case of line pieces), or by capture of either of the paralysed units. The paralysis needn’t be mutual, e.g. … in the case of en passant capture (in position wpc2, bpb4, after move c2-c4 is white pawn paralysed for 1 move, while black isn’t, it means that pawn, that just double-stepped, can be captured en passant). Paralysed unit looses whole movement and activity, except the ability to paralyse…
Invented by Abdul Jabbar Karwatkar in 1979

Consequent: This is usual play with the only distinction that all intermediary positions are evaluated independently of the earlier moves (similarly to consequent series-helpmate, where the legality of the position is reconsidered after each move). In series-mover this fairy condition is known as Chess, Consistent. Problem legality of the position is determined after every move.
Invented by Michel Caillaud in 1979 [ENCYCLOPEDIA OF CHESS PROBLEMS, 2012]

 No.352 Ivan Skoba Czech Republic original-22.07.2013 Dedicated to Naoko Sakata trio   ser-!=49                              (6+11) C-Madrasi + Consequent     Solution: (click to show/hide)   Main plan:1.d3? 2.Ke4! – black has only got f7-f5 as last move (the white king didn’t move into check,because the pawn is paralysed due to the Madrasi condition). White must remove black Pg6.1st foreplan:1.Ke4? 2.Kf4 3.Kg4? with intention Kxh4 … Kxd7 …Kxg6 etc., but no last legal black move exists (after 3rd move illegal position arises). Other foreplan is needed – white must first of all capture bBd7 and then bSh4. Solution:1.Kd3 2.Kc2 3.Kb2 4.Ka3 5.Ka4 6.Kb5 7.Kb6 8.Kc7 9.Kxd7 10.Kc7 11.Kb6 12.Kb5 13.Ka4 14.Ka3 15.Kb2 16.Kc2 17.Kd3 18.Ke4 19.Kf4 20.Kg4 21.Kxh4 22.Kg4 23.Kf4 24.Ke4 25.Kd3 26.Kc2 27.Kb2 28.Ka3 29.Ka4 30.Kb5 31.Kb6 32.Kc7 33.Kd7 34.Ke7 35.Kf6 36.Kxg6 37.Kf6 38.Ke7 39.Kd7 40.Kc7 41.Kb6 42.Kb5 43.Ka4 44.Ka3 45.Kb2 46.Kc2 47.Kd3 48.Kd4 49.d3 ==White is stalemated.   In final position the move 50.Ke4? isn’t possible. In this case black pawn f5 isn’t paralysed; last black move could be either 49.- f7-f5 or 49.- g6xf5. Finally, move 50.g6? isn’t possible, too. It leads to illegal position (white is checked from Pe5). Problem isn’t computer tested. (Author) No.352.1 Ivan Skoba & Arno Tüngler   Czech Republic / Germany Correction to No.352, 26.08.2013 Dedicated to Naoko Sakata trio   ser-!=54                              (4+11) C-Madrasi + Consequent     Solution: (click to show/hide)   Without bRg4 White could immediately play 1.Kd4 2.d3!=. That would be auto-stalemate as it can be proved that the last move of Black could only be 1…e7-e5 and thus the wK is not in check. However, 3.Kxg4? does not work as Pf5 protects the rook. Therefore, White captures the bPf6 so that the last move after Kxg4 could only be …f7-f5, thus again having this move without a check according Madrasi condition: 1.Ke2 9.Kc7! 12.Kxf6. (Now the bPg6 cannot be captured 13.Kxg6? 15.Kxg4?? as then the last move could also be gxf5!) So, White must first capture the bR: 13.Ke6! 25.Kxg4! (last black move f7-f5!) 27.Ke2. Now it is early to finish as 29.Kd4? as 30.d3 is no auto-stalemate since White can play 31.Ke4! – a negative result of the capture of f6! Therefore, he must first capture the bPg6 so that there is gxf5 as an alternative last move which prohibits moving the wK to e4: 28.Kd1! 39.Kxg6 and finish with switchback. 40.Kf6! (not 40.Kh5? with intention 41.Kg4? 47.Kd4 48.d3 != as Pf5 protects the field g4 – again the last move could also be gxf5!) 53.Kd4 54.d3 !=. In the final position White cannot play 55.g6? as this would place the wK in auto-check since the last move could be fxe5+. Additionally, it is not possible to capture the bPc6 (9.Kxc6? or later) as in the final position White could play 55.dxe6 ep! Solution: 1.Ke2 2.Kd1 3.Kc1 4.Kb2 5.Kb3 6.Ka4 7.Ka5 8.Kb6 9.Kc7! 10.Kd7 11.Ke6 12.Kxf6 13.Ke6! 14.Kd7 15.Kc7 16.Kb6 17.Ka5 18.Ka4 19.Kb3 20.Kb2 21.Kc1 22.Kd1 23.Ke2 24.Kf3 25.Kxg4! 26.Kf3 27.Ke2 28.Kd1! 29.Kc1 30.Kb2 31.Kb3 32.Ka4 33.Ka5 34.Kb6 35.Kc7 36.Kd7 37.Ke6 38.Kf6 39.Kxg6 40.Kf6! 41.Ke6 42.Kd7 43.Kc7 44.Kb6 45.Ka5 46.Ka4 47.Kb3 48.Kb2 49.Kc1 50.Kd1 51.Ke2 52.Kd3 53.Kd4 54.d3 !=. (Authors)

### 26 Responses to No.352,352.1 (IS,IS&AT)

1. Nikola Predrag says:

The 48th move is probably Kd4, not Ke4 as written in the solution. The whole idea seems very nice and interesting and the final 50.g6? nicely ends it.
But, as far as I’m able to understand the ‘Consequent’ condition, if 18.Ke4 is legal then 50.Ke4 or 50.g6 should also be legal.
I sincerely hope that I’m wrong and that a precise interpretation of the condition will prove that solution is correct.

2. Ivan Skoba says:

“Problem legality of the position is determined after every move.”
Difference between 18.Ke4 and 50.Ke4 is in presence (or absence) black Pg6. When Pg6 is present, final black move (in position after 18.Ke4) can only be f7-f5 and that’s why is legal.
If Pg6 is absent, move 50.Ke4 is illegal therefore no final black move exists (last black move could be f7-f5 or g6*f5 and we can’t prove, what it was – therefore only move g6*f5 can be considered). Analogously we can judge move 50.g6.

3. Nikola Predrag says:

Yes, it was clear that without bPg6 there is possibility of gxf5 and without wPg5, fxe5 is possible, thanx anyway.
But I don’t understand what would that change in the interpretation of the rule. Either ‘Consequent’ is not enough clearly and precisely defined or I miss something in the definition of a seriesmover.

First I don’t understand how (from the diagram) White can play Kd4-e4 and not Kd4-c4. The diagram is legal because the last move was e7-e5. This means that bP played to f5 earlier and is not paralised. So square e4 is attacked in the diagram position. Nevertheless, White plays Ke4 saying: Now We’ll forget the previous ‘history’ (and the move e7-e5), the new position is legal with White to move, because now the last move was f7-f5. If this reasoning is correct, then why the first move would not be 1.Kc4? The new position would be legal again with White to move and now the last move would be either Sb7-a5+ or Sb3-a5+?.

What makes a difference in case if possible last move is either f7-f5+ or gxf5+? Ep. capture is illegal if we can not prove that the last move was a double step of a Pawn, but this is not a proof that a double step was not the last move. Or is it a convention?

The only explanation I see is logically dubious, but I might accept it for the sake of an interesting condition and this problem:

Step 1 – we accept that the diagram is legal and e7-e5 was the last move

Step 2 – then we immediately forget step 1 and say that f7-f5 was the last move, bPf5 is temporarily paralysed, the square e4 is not attacked and wK could step on it

Step 3 – we play 1.Kd4-e4 (Normally, after white move, Pf5 would not be paralysed any more and wK would be in check (if Black would now have to move), but in a seriesmover White can move again and must parry the check. With the added condition ‘Consequent’, wK is now not even in check on e4, because the last move was f7-f5 and Pf5 is paralysed.)

Without bPg6, we could also say (in step 2) that f7-f5 was the last move but now bPf5 would not be paralysed and the square e4 would be attacked.

Even with accepting such a strange logic of step 2, what is with the final position after 49th move?
Position is legal and e7-e5 was the last move, bPe5 is paralysed and wK is not in check
We can play 50.g5-g6.
Normally, after white move, Pe5 would not be paralysed any more and wK would be in check (if Black would now have to move), but in a seriesmover White can move again and must parry the check. The position after 50.g5-g6 is legal, the last move was either e7-e5 (not a check) or f6xe5#. If the steps 2 and 3 allow 1.Kd4-e4, then why 50.g5-g6 would be illegal, although it might lead to the ‘auto-mate’? 🙂

4. Ivan Skoba says:

Dear Nikola,
my English isn’t so good to understand all you wrote. Similarly I don’t know how precisely answer your questions in this chat.
Majority of fairy conditions is relatively free to explain them and this entry can stand for only one point of view. Maybe, it can’t be correct standpoint.
I propose to continue our discussion via e- mail.
During Chess Composition Festival Marianka 2013 (1.-4.8.2013) I would like to talk over our opinions with Michel Caillaud, inventor of this condition for series-mover. Then I’ll inform you.

• Nikola Predrag says:

Dear Ivan, thank you very much for your patience. My speculations were confusing and what is important – wrong. I apologize for writing so much.

The trouble was in the definition of a series mover and what is the meaning of forbidden “play into check”, or “a King must not play to an attacked square”. Now I understand what is the intented interpretation of the “Consequent”. Your idea is beautiful and the solution is correct with such interpretation (I hope there are no cooks).

I shall write tomorow about a detail in the rules which is still not precisely defined for my understanding. I need some time to reconsider it.

5. Georgy Evseev says:

I think there is some mistake here.

Let’s start with position
W: Kd4 Pd5 – B: Pe7

Black play 1…e7-e5. This is not a check as the pawn is temporarily paralyzed.

But may white now play 2.d6? I think they may not, as ep paralysis is now gone.

So, in the problem being discussed 1.d3 should be impossible, as ep history is already lost when black are to play in answer to this move.

Or am I wrong?

• Kostas Prentos says:

Georgy, it is all about the Consequent stipulation: The legality of the position is determined after every move. This means, in other words, that e7-e5 is always the last move by Black and the pawn e5 is “constantly” temporarily paralyzed. If White needed another move after 49.d3 (say, 50.Pb2-b3), he would still be allowed to play it, without exposing his King to check. Just because of the Consequent stipulation.

It must be noted that Consequent is a retro stipulation and such problems usually belong in the retro section (whenever there is a separate section).

6. Kostas Prentos says:

This is a nice consequent problem and it seems perfectly clear to me.
The wK cannot move on the square g4 with the bBd7 present. Even if the bPf5 does not attack the King, retracting f7-f5 would be illegal due to the preexisting check by the Bd7. So, the Bishop must be eliminated first, then the Knight that protects g6 and finally the bPg6.

The final position is critical. The wK is not in check, because the last move by the bPe5 must have been e7-e5, temporarily paralyzing this pawn (by the wPd5) – a typical effect in madrasi.

The important point is to understand why 50.g6 is not allowed. The reason is that after g6, Black’s last move could have been f6xXe5, and the wK would be in check; therefore, 50.g6?? is illegal because it exposes the wK to self-check. So, the wPg5 cannot move (let’s say that it is pinned), the wK is not in check by the bPe5, and the move Ke4 is not allowed, because after the elimination of the bPg6, Black’s last move could have been g6xXf5.

7. Arno Tüngler says:

Yes, I think that Nikola and Kostas came to the right conclusion. The main point is that in consequent seriesmovers the active side can play an en-passant move (or have the “right” to play such a move) only if you can PROVE that the last move of the opposite side could only have been a double-step of the pawn to be captured. Thus, in Madrasi you are not allowed to play the king into a position where he is attacked by a pawn that could also have moved to that place by another than a double-step. That is because of the definition of a seriesmover that does not allow moves into auto-check. I recommend all to have a look at the following link http://originals.chessproblems.ca/pdf/T61.pdf with an earlier gem of Ivan with a similar idea that won the first prize in the informal tournament of chessproblems.ca 2011 and even will appear in the FIDE-Album 2010-2012 (hope that I did not reveal a secret…).

8. Georgy Evseev says:

Still I do not understand.

Some position is proved to be legal only if the last move was f7-f5. Why in this case the possibility of last move g6:f5 breaks the legality, while the possibility of move Sb7-a5 does not?

• Arno Tüngler says:

The move Sb7-a5 would not have been a legal move for Black in the position with wKe4 because in that case White would have been in check by bPf5 already before the move.

• Georgy Evseev says:

Maybe, f7-f5 was played a move earlier and so Sb7-a5 is legal))

• Arno Tüngler says:

No, please look at this Madrasi logic (that is quite extraordinary…) in the simple example in PDB P1235849. After 2.b4 Black cannot move as any move of the bQ would bring the bK into check. On the other hand Black is not in check as after the white move the wP is temporary paralyzed (not pinned!) and so does not check the bK. See this effect also in a helpstalemate by Ivan here: http://www.jurajlorinc.com/chess/madrep.htm#uloha6

• Georgy Evseev says:

If we use the logic from P1235849, then there is absolutely no sense in No.352.

For example, after 1.d3 any “normal” proof game leading to the resulting position with black to move allows black to capture white king, so 1.d3 should be considered illegal.

• Arno Tüngler says:

This is exactly the point in No. 352. The goal is auto-stalemate. So, in the final position it is not Black but White to move! (Remember, stalemate can only be the side that is on move!) However, if you now (with White to move!) check whether White is in Check or not, you see that the only possible move by Black in his last move could have been e7-e5 and White therefore is not in check. Also, he has no legal move that does not bring him into auto-check – thus he is stalemate.

9. Nikola Predrag says:

wK may be in check if White is on the move. We intuitively accept the concept of forbidden ‘auto-check’.
But what is the exact moment in which we determine that wK would be in check and White is yet not allowed to move. There should be some virtual moment in reasoning when Black would be able to capture wK, so we could say – That is this illegal moment and the particular supposed move is illegal.

What would be a perfectly precise self-consistant definition of autocheck in seriesmovers?

• Arno Tüngler says:

@Nikola: I fear that I do not have a response to your question – that is better left to people who know better English… However, the point in No. 352 is the paradox combination of seriesmover (where “autochecks” are not allowed), auto-stalemate (somehow additionally indicating that the passiv side does not have any right to move!), Madrasi (with the “strange” en-passant rule) and the consequent condition (requiring that after each series-move the position is viewed as if the passive side just had played a move). For me the logic resulting of this combination is fine.

• Georgy Evseev says:

The move may only be considered self-check, if the checking side (black in case of 352) is at least virtually receives the right to move. And at this moment a pawn which has made a double step is no longer ep-paralyzed!

This is a kind of unfortunate paradox: white king is not under check when it is white to move, but any white move leading to the position is illegal auto-check.

10. Julia says:

Ivan has asked me to correct the solution, so it is done: 47.Kd3 48.Kd4 49.d3.

11. Nikola Predrag says:

There’s a position A and after a single move X there will be a position B.
What decides whether X is legal or illegal?
1.Definition of legal movements of the pieces
2.Definition of a seriesmover
3.Definition of ‘Consequent’

In case of an ordinary seriesmover, X must not be an ‘autocheck’. So we need the exact definition of auto-check and the exact moment when we can determine auto-check.

If we look at the position A and think about possible consequences of the move X, that is a moment 1.
If we play X and look at the position B that is a moment 2. In this case we or some arbiter can conclude that X was illegal because of auto-check and we must withdraw X, turn to the position A and try some other move.

In case of ordinary seriesmover we apply a kind of intuitive concept of “auto-check” and the mentioned moments 1 & 2
seem irrelevant for the legality.

With the condition Consequent, the histories of the positions A and B are different and the moments 1 and 2 become relevant. Consequent rules should also be obeyed when we try to determine whether X is an auto-check or not.

In moment 1, we should apply the rule 3.8a from The Laws of chess (http://www.fide.com/fide/handbook.html?id=124&view=article):
– There are two different ways of moving the king:
by moving to any adjoining square not attacked by one or more of the opponent’s pieces or by castling…. –

If in the position A some square is attacked, a King simply can’t move to that square.
If nevertheless we play X and wait for the arbiters decision about the legality of that move, we have reached the position B and the moment 2. The history of A is forgotten and the arbiter says that the position of B is legal with White on the move.
This means that wK may be in check, it could be even mated. We don’t remember how we actually reached the position B – it is simply a legal position with the last (check)mate move made by Black.

In Ivan’s problem, looking in the moment 2, the illegal kind of auto-check would be determined only after 1.Ke4 2.Kf4 3.Kg4 – in the history of this position, wK must have been in check even before the last black move. But 1.Kc4 would be legal.

Looking in the moment 1, questioning the legality of some move X before we play it, we must apply the history of that position (A). Now, a King can’t move to a square which is attacked according to the history of the position A.

Still, the combonation of Seriesmover+Consequent+Madrasi, offers this amuzing possibility of extended/repeated temporary ep. paralysis and it seems worth searching for a proper definition of legal moves which would allow Ivan’s solution and the logic behind it.
Such definition would have to allow that while we analyse the new history of the position B after the move X, we still remember what was the move X – which happens between the moments 1 and 2.

Perhaps my reasoning is confusing, I’ll try to say the point shortly:
How can we explore the history of a position before we play a move which leads to this position? Once, when we play such move, it must be forgotten, due to the condition Consequent.

• Arno Tüngler says:

In fairy chess there are conditions where you need to take into account the future. In checkless check you may well be allowed to do a check, if the chain after all potential future moves (that never will be played) finishes with a check. The move becomes by these “future moves” a mate and thus legal. Really, what happens in these Madrasi consequent seriesmovers is similar: you do a move that normally would not be allowed. However, taking into account which “consequences” arise in the position after this move, you see that the move is legal. That those “consequences” are then connected with “the history” of the new position is just because of the fairy condition “consequent” – not because of “real” retroplay.
For that reason, in my opinion, it is right to treat those consequent seriesmovers as fairies and not as retros.

• Kostas Prentos says:

Arno, I would agree that in most cases of consequent problems, the retro play is not very deep. Consequent problems usually deal with proving that the en passant capture can be played; sometimes, they build a castling or two and there are some examples in which the retro play is deeper. The fact is that consequent problems can be found in both the fairy and the retro sections in magazines. Moreover, in the FIDE albums, they have appeared in both sections, apparently depending on where the authors submitted them. In the 1998-2000 Album, there are two consequent problems as fairies and four as retros, but in previous years, I found all of them in the fairy section.

Generally, I follow the rule that all problems with fairy conditions and at the same time with retro content, belong in the retro section. In other words, the retro element takes priority. But this is not a universal rule. For example, in “the Problemist” you can find Proca Retractors and Proof games without any fairy conditions in the Retro section of the magazine, but when they have a fairy condition, they go in the fairy section.

Since there are no strict rules, does this mean that the authors of consequent problems can submit the same problem in both the retro and the fairy section of the WCCI and FIDE album? The intention of this idea is obvious: To maximize the chances of the problem to make it in the FIDE album. In my opinion, this concept is wrong (maybe also unethical). Imagine the case when the same problem succeeds in both sections. Will the author win 2 points for one problem, or in retrospect, will he (or somebody else) have to choose the section in which the problem will appear?

As far as I know, there is no rule that forbids this scenario. It is usually between retros and fairies, but it can occur with retros and twomovers, or even helpmates. Sometimes, it can also happen with moremovers and studies. The last case can produce rather ridiculous results: If the problem is accepted in the moremovers section, it will receive 1 point, while in the studies section 1.66 points. Anyway, I do not want to revisit the heated discussion from the Mat Plus forum, about how many points studies should receive, but maybe it is a good idea for the WFCC to make a rule that allows only one section submission per problem. Of course, it should be the composer’s decision, in which section he would submit his problem.

• Kostas Prentos says:

Reading again the last paragraph of my previous reply, I realize that the scenario I was describing cannot happen between moremovers and studies. In this case, the composer has already chosen the stipulation of the problem (either mate in n, or win) when it was first published.

• Arno Tüngler says:

Yes, you are right, it is quite funny that consequent seriesmovers ended in two different sections in the FIDE Album… As long as there is no double publication, I have no problems with that. For me the only reason to send problems to the FIDE Albums is to contribute to a collection of the best chess problems for the current period. The “titles” are in my opinion just a strange side-effect that I take not serious as it is also not taken serious outside our very small community. I know that others have other opinions and leave it to them to fight for righteousness in this area…

• Nikola Predrag says:

Kostas, you may publish the same position both as #n and a study and send each one of such “versions” to a respective FA section.

12. Ivan Skoba says:

Thank you for all your comments that mean very much for me.