Immun Chess: Фигура (включая короля) иммунна (не может быть взята), если ее Circe-поле возрождения (поле того же цвета, занимаемое в расстановке фигур перед началом партии для фигур, для пешки – на линии взятия, для сказочных фигур – поля превращений на линии взятия) занято другой фигурой.

No.1316Gani Ganapathi &
Nicolas Dupont India / France original – 08.08.2018

The remark for the judge as of 18-Aug-2018:
Note that there are solutions in 11.0 moves, and there is no general consensus on whether or not it cooks the problem in 11.5 moves. We think that a shorter solution by 0.5 doesn't cook the problem, because the side on move has changed. If you disagree, please modify the entry to a 12.0 PG with the additional move 12...Kf6, so that the problem becomes a sound SPG. (Authors)

Georgy, as you know, a shorter solution is a cook. To avoid a stipulation type “PG in exactly 11.5 moves”, the authors must introduce a last checking move. Fortunately, this move is possible: 12. …Sc2+ (confirmed by Jacobi).

PG 11.5 cannot be cooked in 11 moves because the side to move will be different. Shorter solution (cook) can only be 10.5 or 9.5 etc. I am surprised that Paul Raican, a retro expert makes such a claim.

A similar claim of cook was made for our WCCT entry and it was rightly rejected by the judging countries.

Although the main idea of this problem is to lose a tempo for a PG 11.5, any shorter solution, whether it is in 11 or less moves is valid, unless the authors stipulate that the problem is an exact proof game in 11.5 moves. All proof games are assumed to be shortest proof games (SPG), unless otherwise specified. The stipulation of a PG as an exact proof game is rather unpleasant and composers tend to avoid it. If possible, a better solution is to add an additional half move (like Paul Raican’s suggestion of 12…Sc2+, or preferably a non-checking move) to make the problem sound as a shortest PG and preserve the tempo play, as well.

Thierry Le Gleuher told me that as the retros editor of Phénix he would consider a PG 11.5 cooked if there is a solution in 11 moves or less, unless the stipulation says “exact”. This is in agreement with Paul and Kostas. As for the current problem, adding the half-move 12… Kf6 works (C+ Jacobi, separate tests in 12.0, 11.5, 11.0, …). The extra move doesn’t have to be check — the goal is only to force 12 white moves.

According to standard chess approach, the same placement of pieces with WTM and BTM are different chess positions. So, strictly speaking, the position reached in 11 moves is not the position asked to be reached by authors.

BTW, i know some composers who consider setplays in helpmates as cooks, as they think that solver should be allowed to select the side to play first.

For threefold repetition of position to be claimed by players the side to move is important. Without WTM or BTM being identical now draw by repetition can be claimed. Every player knows this. To ignore this in Proof games seems illogical & strange.

Joost, the stipulation “Shortest Proof Game” does not need to specify the side to move, or the number of moves that are needed to reach the diagram. In the case of 1316, an SPG would be in 11 moves, not in 11.5. If you specify a problem as an SPG and further add which side’s turn it is to move as in P1004019, then you are looking only for the SPG ending with a white/black move, not for the (absolute) shortest PG.

In order to avoid all controversy, and especially the stipulation: “exact PG 11.5” or “SPG with Black to move”, composers prefer to add a “tail” move (like 12…Sc2+ or 12…Kf6, as proposed in this case). I remember a PG by Reto Aschwanden (I cannot find it now) that had a shorter solution (cook) and this remedy of adding an extra “tail” move was not possible, so the composer had to specify that his problem was an exact PG, instead of an SPG. If my memory serves me well, Reto’s problem was selected in the award of that informal tourney, even as an exact PG.

If we try to compare a PG with a helpmate, then the set play in a helpmate would be equivalent to a PG in 11 moves (going backwards) for a solution in 11.5 moves. Even in helpmates, where the stipulation specifies the side to move, we still stipulate the set play. This would make sense in a PG if there was a unique solution in 11 moves (not the case here) and this solution was in contrast with the solution in 11.5 moves. I can think of my own P1288899 as such an example.

In any case, No.1316 is not an SPG (it would be cooked as an SPG), so the stipulation has to change (see the first paragraph of this comment). I would definitely prefer the added 12th move for Black, if this was my own problem, but it is not. Let’s wait and see what the authors will choose to do.

Thank you, Joost. This is the Reto PG. It is cooked in 21.5 moves, but not in 22 moves. I don’t have any old Probleemblad issues with me, but according to WinChloe, the award was published in a special issue (December 2009).

Is is true that, in the modern approach, PG is often used to denote in fact SPG, and exact PG to denote in fact non-shortest PG.

Nevertheless another approach seems to be used both in the PDB and in WinChloe. For example P1288899 is labelled a) BP in 20.0 b) BP in 20.5, and a) PJ in 20.0 b) PJ in 20.5 respectively, and not b) exact BP/PJ in 20.5. So it seems here that PG in X moves means PG in exactly X moves.

I think it is the best option (rendez-vous at the restaurant in 3 hours means in exactly 3 hours, not in 1, 2 or 3 hours), even if I can understand the logic of the others. So to avoid any confusion, I agree to add 12… Kf6 in 1316.

PDB uses the word “genau” extensively. The search stip=”genau” and k=”unique proof game” returns 27 problems, including twins that differ by 0.5 (P1091679, P1339295).

In both examples (P1091679, P1339295) the word “genau” could have been omitted, imho. It is clear that in both problems the twin b is not an SPG.

Proof games with tempo play often have the same difficulties as 1316. See for example the discussion on whether the added “tail” move 16…Ke7 was necessary or not, in this fine proof game. It is easy to forget this detail, especially since the PG solving engines don’t search for solutions in less than the stipulated moves.

Good choice, Nicolas. Although there are exceptions to this rule, it is a well established convention that a PG is a shortest PG. Yet, I don’t agree that P1288899 is a good example to support the opposite opinion. The twin a is an SPG, while the twin b obviously is not (as both twins have the same final position). I believe it is all quite clear, without the need to stipulate that twin b is an exact PG.

I have always operated under the assumption that a PG in n moves is in fact an SPG. To this end, stipulating the number of moves is a courtesy to the reader/solver by the composer and could be omitted without changing anything in the solution. Like with classical retros, the least explaining that has to be done by the composer in the form of the stipulation, the better.

Finally, stipulating that a PG in n moves is not an SPG, but an exact PG is not always bad. See for example the popular P0000811 that owes its reputation to the fact that it is a PG in exactly 4 moves and not in 3 or 3.5.

I don’t understand the logic, if there is one at all.
11.5 is AS EXACT a number as it could ever be.
For a PG it means 11 pairs of halfmoves + 0.5 of a pair, so White made the last move.

So far, everything is clear and there’s no room for any arbitrarily imposed bureaucratic conventions.

PG 11.0 is not a solution, since the stipulated position is not matched.
PG 10.5 hypothetically might reach the required position and if such PG is real, that could require an agreement and some convention as a default.

(Even this wouldn’t be an issue for me. As a default, PG 11.5 asks for any play reaching the diagram in 11.5 moves, where 11.5 is actually 11.5, no less, no more. The stipulation tells about what has ALREADY HAPPENED – the exact position and number of moves. “Forward stipulations” tell what could happen, and the default logic of stipulation could be different.)

Nikola, if you do not specify that the solution is in exactly 11.5 moves, then any shorter solution is a cook. The magic word here is “exactly”. A PG is supposed to be a shortest PG, the fastest way to reach the given position. To avoid confusion, you can specify that the problem is in exactly 11.5 moves, but this is really not the best available choice for the PG composer, only the last resort.

Ok, François and Kostas – the PDB is using “exact PG” for a non-shortest game (in the case there are no twins). But it is not true for WinChloe! For example the famous Orban game is labelled “PJ in 4.0” and not “exact PJ in 4.0”.

So it seems that at least WinChloe is sharing Nikola’s argument – 4.0 means 4.0, no different length could be considered a cook as it doesn’t fill the stipulation.

As I said before, this option also sounds the best to my mind, even if I accept (and Gani too) to add a tail move in 1316 in order to avoid any conflict, but without being convinced by its full necessity.

Kostas, you may claim that 11.5 means 173.666 or whatever you wish, and you may bureaucratically impose it to the community. But, is that what you really want?
Direct simple hierarchy of logic is the best default.
11.5 is the EXACT number and the default meaning is exact.
WHY on earth would anyone distort such a clear logic?

You might argue that 10.5 moves leading to THE SAME POSITION makes a short solution. We could accept that as a default for possibly practical reasons, but first you present them convincingly!

However, PG 11.0 can NEVER reach THE EXACT position as the stipulated one. So, what are we talking about?

PG means ANY PG, and if the number of moves is given, both the position and number are known.
For ‘shortest PG’, just stipulate SPG, WITHOUT a number.

But no, upside-down logic claims that indicating ‘shortest’ is not necessary (as if that was obvious???) and that 11.5 is not exactly 11.5 unless explicitly indicated!??

Anyway, you say:
-“A PG is supposed to be a shortest PG, the fastest way to reach the given position”-

Do we really have to read again the rules of the game to comprehend what is a position?

Kostas, it’s by no means about you as an individual.
And it’s not about me as an individual.
Logic suits nobody, there is THE logic or no logic.
It’s terrifying how easily we surrender to ‘out-of-mind’ conventions!
Position after 11.5 moves becomes a position after 11.0 moves.
I am very sorry that Nicolas has surrendered. And one by one, the others will surrender without a single convincing argument.

WHO has invented that ‘convention’ and why do you (or anyone) accept it?

The “convention” is somewhat natural: it should be “easy” to reach any position if there is no limitation to the number of moves to reach the given position, while the shortest way may be expected to be difficult.

Still, there is a number of examples when longer way is less obvious than the fastest one. I think that it is a good move in right direction when the stipulation “SPG” is replaced with “PG n”.

I am 100% sure that the solution that the solution with the other side to move in the required position is not a cook. I think it is a matter for discussion if _any_ shorter solution should be considered a cook if the solution in required number of moves is unique.

Nikola, I was not going to engage in an exercise in futility, but your last question intrigued me a bit. I don’t know the answer, but this link may be of some help. I don’t own the book “Shortest Proof Games” by Gerd Wilts and Andrey Frolkin. My guess is the authors may offer a historic account.

As to the question “why I accept it”, the answer is “because it makes perfect sense to me”. There is no way in hell I would knowingly allow a cook such as this in one of my own compositions. Others can do whatever pleases them.

I don’t believe there is anything else I can add without repeating my previous comments.

The example no 8 from the aforementioned book “Shortest Proof Games” by Michel Caillaud & Jacques Rotenberg, 2HM Europe Echecs 1991 (PDB – P0001681) has the same stipulation mentioned by Joost “Shortest Proof Game?” with two twins: a) White to move and b) Black to move.
That means between 1991 and 2001 there was probably no general objection against using the stipulation “Shortest Proof Games” without specifying the number of moves in such tempo motivated cases.

For the second question: in any stipulation, not specifically for proof games, any way to reach to the conclusion in less than the specified number of moves is considered to be a shorter solution. Therefore, the accepted convention for such cases is to add the word ‘exact’ in the stipulation.
This convention was not invented by the retro composers – it was just imposed in the retro genre like other existing conventions from chess composition. The logic for this convention is perhaps to force the solver find the author’s intention and not claim he found a shorter solution to reach the aim.

Dear all,
I request that the following points be considered:-
1.PG in n.5 cannot be cooked by a PG in X.0 as side to play is important.
2.A stip PG.n does not mean a stip.SPG in n moves.
3.While a PG 11.5 can be cooked by pg 10.5 etc it cannot be cooked by a PG 11 as being claimed.

Dear Julia,

Pl. correct the prob. stipulation as PG 11.5

Is it not same as Strict Circe?

Sorry strict five is different

Jacobi found an alternative in 11 moves: 1.b3 Sc6 2.Ba3 Sd4 3.Bd6 exd6 4.f4 d5 5.f5 Bd6 6.f6 Se7 7.fxe7 f5 8.Sc3 Kf7 9.e8=S Rxe8 10.Sb1 Re6 11.c3 Rg6.

This is not a cook, as a loss of tempo is an idea of the problem.

Georgy, as you know, a shorter solution is a cook. To avoid a stipulation type “PG in exactly 11.5 moves”, the authors must introduce a last checking move. Fortunately, this move is possible: 12. …Sc2+ (confirmed by Jacobi).

PG 11.5 cannot be cooked in 11 moves because the side to move will be different. Shorter solution (cook) can only be 10.5 or 9.5 etc. I am surprised that Paul Raican, a retro expert makes such a claim.

A similar claim of cook was made for our WCCT entry and it was rightly rejected by the judging countries.

A nice idea. Promotion must be to queen, because a knight cannot lose a tempo.

Although the main idea of this problem is to lose a tempo for a PG 11.5, any shorter solution, whether it is in 11 or less moves is valid, unless the authors stipulate that the problem is an exact proof game in 11.5 moves. All proof games are assumed to be shortest proof games (SPG), unless otherwise specified. The stipulation of a PG as an exact proof game is rather unpleasant and composers tend to avoid it. If possible, a better solution is to add an additional half move (like Paul Raican’s suggestion of 12…Sc2+, or preferably a non-checking move) to make the problem sound as a shortest PG and preserve the tempo play, as well.

Thierry Le Gleuher told me that as the retros editor of Phénix he would consider a PG 11.5 cooked if there is a solution in 11 moves or less, unless the stipulation says “exact”. This is in agreement with Paul and Kostas. As for the current problem, adding the half-move 12… Kf6 works (C+ Jacobi, separate tests in 12.0, 11.5, 11.0, …). The extra move doesn’t have to be check — the goal is only to force 12 white moves.

According to standard chess approach, the same placement of pieces with WTM and BTM are different chess positions. So, strictly speaking, the position reached in 11 moves is not the position asked to be reached by authors.

BTW, i know some composers who consider setplays in helpmates as cooks, as they think that solver should be allowed to select the side to play first.

That is the point Georgy.

For threefold repetition of position to be claimed by players the side to move is important. Without WTM or BTM being identical now draw by repetition can be claimed. Every player knows this. To ignore this in Proof games seems illogical & strange.

IMO, SPG 11.5 asks for the shortest proofgame with black to move, and since there’s no proofgame in less moves this is indeed the shortest profgame.

See e.g. P1004019 for a composition with stipulation ‘Shortest proof game? (a) white to move (b) black to move’.

Joost, the stipulation “Shortest Proof Game” does not need to specify the side to move, or the number of moves that are needed to reach the diagram. In the case of 1316, an SPG would be in 11 moves, not in 11.5. If you specify a problem as an SPG and further add which side’s turn it is to move as in P1004019, then you are looking only for the SPG ending with a white/black move, not for the (absolute) shortest PG.

In order to avoid all controversy, and especially the stipulation: “exact PG 11.5” or “SPG with Black to move”, composers prefer to add a “tail” move (like 12…Sc2+ or 12…Kf6, as proposed in this case). I remember a PG by Reto Aschwanden (I cannot find it now) that had a shorter solution (cook) and this remedy of adding an extra “tail” move was not possible, so the composer had to specify that his problem was an exact PG, instead of an SPG. If my memory serves me well, Reto’s problem was selected in the award of that informal tourney, even as an exact PG.

If we try to compare a PG with a helpmate, then the set play in a helpmate would be equivalent to a PG in 11 moves (going backwards) for a solution in 11.5 moves. Even in helpmates, where the stipulation specifies the side to move, we still stipulate the set play. This would make sense in a PG if there was a unique solution in 11 moves (not the case here) and this solution was in contrast with the solution in 11.5 moves. I can think of my own P1288899 as such an example.

In any case, No.1316 is not an SPG (it would be cooked as an SPG), so the stipulation has to change (see the first paragraph of this comment). I would definitely prefer the added 12th move for Black, if this was my own problem, but it is not. Let’s wait and see what the authors will choose to do.

Kostas, I think the Reto composition is P1080421 , but I don’t have my old Probleemblad issues anymore, so I can’t verify.

Thank you, Joost. This is the Reto PG. It is cooked in 21.5 moves, but not in 22 moves. I don’t have any old Probleemblad issues with me, but according to WinChloe, the award was published in a special issue (December 2009).

Is is true that, in the modern approach, PG is often used to denote in fact SPG, and exact PG to denote in fact non-shortest PG.

Nevertheless another approach seems to be used both in the PDB and in WinChloe. For example P1288899 is labelled a) BP in 20.0 b) BP in 20.5, and a) PJ in 20.0 b) PJ in 20.5 respectively, and not b) exact BP/PJ in 20.5. So it seems here that PG in X moves means PG in exactly X moves.

I think it is the best option (rendez-vous at the restaurant in 3 hours means in exactly 3 hours, not in 1, 2 or 3 hours), even if I can understand the logic of the others. So to avoid any confusion, I agree to add 12… Kf6 in 1316.

PDB uses the word “genau” extensively. The search stip=”genau” and k=”unique proof game” returns 27 problems, including twins that differ by 0.5 (P1091679, P1339295).

In both examples (P1091679, P1339295) the word “genau” could have been omitted, imho. It is clear that in both problems the twin b is not an SPG.

Proof games with tempo play often have the same difficulties as 1316. See for example the discussion on whether the added “tail” move 16…Ke7 was necessary or not,

in this fine proof game. It is easy to forget this detail, especially since the PG solving engines don’t search for solutions in less than the stipulated moves.

Good choice, Nicolas. Although there are exceptions to this rule, it is a well established convention that a PG is a shortest PG. Yet, I don’t agree that P1288899 is a good example to support the opposite opinion. The twin a is an SPG, while the twin b obviously is not (as both twins have the same final position). I believe it is all quite clear, without the need to stipulate that twin b is an exact PG.

I have always operated under the assumption that a PG in n moves is in fact an SPG. To this end, stipulating the number of moves is a courtesy to the reader/solver by the composer and could be omitted without changing anything in the solution. Like with classical retros, the least explaining that has to be done by the composer in the form of the stipulation, the better.

Finally, stipulating that a PG in n moves is not an SPG, but an exact PG is not always bad. See for example the popular P0000811 that owes its reputation to the fact that it is a PG in exactly 4 moves and not in 3 or 3.5.

I don’t understand the logic, if there is one at all.

11.5 is AS EXACT a number as it could ever be.

For a PG it means 11 pairs of halfmoves + 0.5 of a pair, so White made the last move.

So far, everything is clear and there’s no room for any arbitrarily imposed bureaucratic conventions.

PG 11.0 is not a solution, since the stipulated position is not matched.

PG 10.5 hypothetically might reach the required position and if such PG is real, that could require an agreement and some convention as a default.

(Even this wouldn’t be an issue for me. As a default, PG 11.5 asks for any play reaching the diagram in 11.5 moves, where 11.5 is actually 11.5, no less, no more. The stipulation tells about what has ALREADY HAPPENED – the exact position and number of moves. “Forward stipulations” tell what could happen, and the default logic of stipulation could be different.)

Nikola, if you do not specify that the solution is in exactly 11.5 moves, then any shorter solution is a cook. The magic word here is “exactly”. A PG is supposed to be a shortest PG, the fastest way to reach the given position. To avoid confusion, you can specify that the problem is in exactly 11.5 moves, but this is really not the best available choice for the PG composer, only the last resort.

Ok, François and Kostas – the PDB is using “exact PG” for a non-shortest game (in the case there are no twins). But it is not true for WinChloe! For example the famous Orban game is labelled “PJ in 4.0” and not “exact PJ in 4.0”.

So it seems that at least WinChloe is sharing Nikola’s argument – 4.0 means 4.0, no different length could be considered a cook as it doesn’t fill the stipulation.

As I said before, this option also sounds the best to my mind, even if I accept (and Gani too) to add a tail move in 1316 in order to avoid any conflict, but without being convinced by its full necessity.

Kostas, you may claim that 11.5 means 173.666 or whatever you wish, and you may bureaucratically impose it to the community. But, is that what you really want?

Direct simple hierarchy of logic is the best default.

11.5 is the EXACT number and the default meaning is exact.

WHY on earth would anyone distort such a clear logic?

You might argue that 10.5 moves leading to THE SAME POSITION makes a short solution. We could accept that as a default for possibly practical reasons, but first you present them convincingly!

However, PG 11.0 can NEVER reach THE EXACT position as the stipulated one. So, what are we talking about?

PG means ANY PG, and if the number of moves is given, both the position and number are known.

For ‘shortest PG’, just stipulate SPG, WITHOUT a number.

But no, upside-down logic claims that indicating ‘shortest’ is not necessary (as if that was obvious???) and that 11.5 is not exactly 11.5 unless explicitly indicated!??

Anyway, you say:

-“A PG is supposed to be a shortest PG, the fastest way to reach the given position”-

Do we really have to read again the rules of the game to comprehend what is a position?

Nikola, this is not a convention I invented, nor am I trying to impose it to anyone. You are free to follow whatever logic suits you best.

Kostas, it’s by no means about you as an individual.

And it’s not about me as an individual.

Logic suits nobody, there is THE logic or no logic.

It’s terrifying how easily we surrender to ‘out-of-mind’ conventions!

Position after 11.5 moves becomes a position after 11.0 moves.

I am very sorry that Nicolas has surrendered. And one by one, the others will surrender without a single convincing argument.

WHO has invented that ‘convention’ and why do you (or anyone) accept it?

The “convention” is somewhat natural: it should be “easy” to reach any position if there is no limitation to the number of moves to reach the given position, while the shortest way may be expected to be difficult.

Still, there is a number of examples when longer way is less obvious than the fastest one. I think that it is a good move in right direction when the stipulation “SPG” is replaced with “PG n”.

I am 100% sure that the solution that the solution with the other side to move in the required position is not a cook. I think it is a matter for discussion if _any_ shorter solution should be considered a cook if the solution in required number of moves is unique.

Nikola, I was not going to engage in an exercise in futility, but your last question intrigued me a bit. I don’t know the answer, but this link may be of some help. I don’t own the book “Shortest Proof Games” by Gerd Wilts and Andrey Frolkin. My guess is the authors may offer a historic account.

As to the question “why I accept it”, the answer is “because it makes perfect sense to me”. There is no way in hell I would knowingly allow a cook such as this in one of my own compositions. Others can do whatever pleases them.

I don’t believe there is anything else I can add without repeating my previous comments.

The example no 8 from the aforementioned book “Shortest Proof Games” by Michel Caillaud & Jacques Rotenberg, 2HM Europe Echecs 1991 (PDB – P0001681) has the same stipulation mentioned by Joost “Shortest Proof Game?” with two twins: a) White to move and b) Black to move.

That means between 1991 and 2001 there was probably no general objection against using the stipulation “Shortest Proof Games” without specifying the number of moves in such tempo motivated cases.

For the second question: in any stipulation, not specifically for proof games, any way to reach to the conclusion in less than the specified number of moves is considered to be a shorter solution. Therefore, the accepted convention for such cases is to add the word ‘exact’ in the stipulation.

This convention was not invented by the retro composers – it was just imposed in the retro genre like other existing conventions from chess composition. The logic for this convention is perhaps to force the solver find the author’s intention and not claim he found a shorter solution to reach the aim.

Different genres have different principles and meaning of the features. And the outcome of a logic procedure may (expectedly) be different.

There’s one logic but the input is essentially different for different genres.

There’s obviously no curiosity about convincing arguments.

This subject was discussed on the MatPlus forum back in 2012:

http://www.matplus.net/start.php?px=1534048897&app=forum&act=posts&fid=gen&tid=1139&pid=8975&mark=yes&#n8975

True. It had been discussed then also. What was the conclusion? 🙂

Dear all,

I request that the following points be considered:-

1.PG in n.5 cannot be cooked by a PG in X.0 as side to play is important.

2.A stip PG.n does not mean a stip.SPG in n moves.

3.While a PG 11.5 can be cooked by pg 10.5 etc it cannot be cooked by a PG 11 as being claimed.