Website founded by Milan Velimirović in 2006
2:51 UTC
| |
MatPlus.Net Forum General Law 9.2: Draw by Repetition |
|
|
|
You can only view this page!
| Page: [Previous] [Next] 1 2 3 | (1) Posted by Andrew Buchanan [Wednesday, May 1, 2013 06:22] | Law 9.2: Draw by Repetition Happy May Day, folks.
Does anyone here understand the rules of chess? :)
Mostly I guess we all do. But I confess myself a bit baffled by Law 9.2, which says:
"The game is drawn upon a correct claim by the player having the move, when the same position, for at least the third time (not necessarily by a repetition of moves):
a) is about to appear, if he first writes his move on his scoresheet and declares to the arbiter his intention to make this move, or
b) has just appeared, and the player claiming the draw has the move.
Positions as in (a) and (b) are considered the same, if the same player has the move, pieces of the same kind and colour occupy the same squares, and the possible moves of all the pieces of both players are the same.
Positions are not the same if a pawn that could have been captured en passant can no longer be captured in this manner. When a king or a rook is forced to move, it will lose its castling rights, if any, only after it is moved."
Please let's keep any discussion focused on the Laws, not on the Codex.
Concerning en passant, I think the Law is relatively clear. Following a double pawn move, an e.p. may be illegal due to a discovered check or pin. Only if every potential e.p. is illegal might the position be repeatable.
Concerning castling though... brrr! Chaos. I read the clause "...the possible moves of all the pieces of both players are the same", and this seems unclear at several levels, when applied to castling. For the active player, are "possible moves":
(1) just those available in the current position (i.e. depend on inherent castling rights (= prior movement of K or R) together with dynamic situation (intervening or attacking units) or
(2) including also possible *future* moves after any interveners and attackers have gone away. (i.e. depend on inherent castling rights but also some analysis as to whether those rights are capable of being exercised in the future) or
(3) purely depending on inherent castling rights. (Which to me would be the simplest basis for the rule - but is it what's actually intended?)
The situation is complicated by the fact that we need to consider the other player too. What happens if this person is checking their opponent? It is completely impossible for the checker to have the move, and yet somehow we must interpret the phrase "possible moves" in this impossible situation. Or is "possible moves" following all possible moves by the active player?
So... pretty vague so far. OK so I look for clarification later in the Law, and I find: "When a king or a rook is forced to move, it will lose its castling rights, if any, only after it is moved."
Christmas! It's full of stars! I haven't a clue what's going on here - please can someone help.
Thanks,
Andrew. | | (2) Posted by Joost de Heer [Wednesday, May 1, 2013 09:08]; edited by Joost de Heer [13-05-01] | - Does white have castling rights in the begin position?
- Does white have castling rights after 1. e4 f5 2. Qh5+?
- Does white have castling rights in the position W: Ke1 Rh1 B: Ke8 Pe2? (this is to clarify the vague sentence about losing rights after moving)
The intended maening of the article is about rights. Not whether castling itself is technically possible at the moment of determining the 3-fold repetition, but whether the castling (and ep) right is present.
I agree though: awful wording. | | (3) Posted by Sarah Hornecker [Wednesday, May 1, 2013 10:37]; edited by Sarah Hornecker [13-05-01] | Andrew, you bring up an interesting question!
(= 2+2 )
Black plays g7-g5, and the position is repeated two more times. Can a claim be made for threefold repetition?
As for castling rights, it is decided that the permanent loss of those rights creates a different position, even if the temporary rights were the same. But what about en passant?
For the example we assume that Ke8 and Ra8 have not yet castled.
(= 6+5 )
a) It follows 1.e7 g5 2.Kc5 Rc8+ 3.Kb5 Ra8 4.Kc5 Rc8+ 5.Kb5 Ra8. Draw by repetition?
b) It follows 1.e7 g5 2.Kc5 Rh6 3.Kb5 Rh5 4.Kc5 Rh6 5.Kb5 Rh5. Draw by repetition?
c) It follows 1.-g5 2.e7 Rb8+ 3.Ka5 Ra8+ 4.Kb5 Rb8+ 5.Ka5 Ra8+ 6.Kb5. Draw by repetition?
d) It follows 1.-g5 2.e7 Rh6 3.Kc5 Rh5 4.Kb5 Rh6 5.Kc5 Rh5 6.Kb5. Draw by repetition?
In case a, I would say no because the rook that could have castled can not castle anymore.
In case b, I am uncertain: the permanent en passant right has changed, even though the temporary right prevented en passant anyway because of the self-check.
In case c, I would say no because the rook that could have castled can not castle anymore.
In case d, I would say yes. The same position has arisen three times.
(It does not matter for this that White could easily win by playing Rd8+, just let this be a schematic thing!)
-----
Joost:
QUOTE - Does white have castling rights in the begin position?
- Does white have castling rights after 1. e4 f5 2. Qh5+?
- Does white have castling rights in the position W: Ke1 Rh1 B: Ke8 Pe2? (this is to clarify the vague sentence about losing rights after moving)
In the first and second case, yes. In the third case yes, provided that king and rook have not moved yet. It does not matter for the third case that White needs to give up his castling right anyway.
Of course, this is my opinion. But from what I read, the permanent castling right matters as well. | | (4) Posted by Hauke Reddmann [Wednesday, May 1, 2013 13:07] | I'm a chess arbiter (OK, lowest level), my word is LAW :-)
The e.p. case is easy:
The "right" to an illegal e.p. is nonesuch. Thus at the first
time, it's illegal because of selfcheck, at the second time
it's even more illegal because it expired. Since
illegal+illegal=illegal, the position is the same.
The Castling case is much worse than you just outlined!
Assume the following alternate wording of the rule (which
is the intended meaning, since the *actual* right of playing
O-O in the position is not relevant):
"If in Position A later on O-O can follow, whereas in Position B
this is impossible, then A and B are different positions."
Fine, but now think of Ke1 Rh1 - Ke8 Re7! (A) Repeat three times. (B)
By "standard" interpretation A and B are different, because White
lost castling rights. But he never could have castled anyway in
the further game!
Hauke | | (5) Posted by Andrew Buchanan [Wednesday, May 1, 2013 16:12]; edited by Andrew Buchanan [13-05-01] | Thanks Joost, Siegried, Hauke for your replies.
Let me turn to Arbiter Hauke first.
I get what you are saying about e.p. There is no such thing as an e.p. right. If the e.p. capture was illegal (as in Siegfried's diagrams) then the position is the same when the diagram is visited again. The ruling on e.p. is completely clear.
Please be a bit more explicit about castling though.
Note the term "Castling rights" is defined in Law 3.8.b:
(1) The right to castle has been lost:
if the king has already moved, or
with a rook that has already moved.
(2) 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, or
if there is any piece between the king and the rook with which castling is to be effected.
I suggested three different interpretations above, which were essentially:
(1) consider just moves available in the current diagram. (But how to handle moves by the other player in any kind of clean way?)
(2) consider all possible future moves for the rest of the game. (Consistent but it involves open-ended help play analysis to evaluate the current state of the game, and I thought that kind of mischief was restricted to areas such as dead positions and time trouble.)
(3) consider just castling rights, as defined above. (Simplest, cleanest. But is this what 9.2 is saying?)
So bottom line, which option is correct? Or if none, then what is the right interpretation, please?
And once the basic rule is clear, let's address the really challenging question. Let's extend your example by adding WNd1 to (A) to get (A'). Suppose that it never moves in getting from (A') to (B'). According to the current rule, there is still some impact that the knight has on repetition because White was never forced to lose castling rights, unlike in A. Please show how this stationary knight affects the evaluation of the game. It does seem that this sentence of the rules is inconsistent with the earlier definition of castling rights, in that now rights can in some circumstances be lost *before* a move takes place.
Thanks so much!
Andrew. | | (6) Posted by Georgy Evseev [Wednesday, May 1, 2013 17:01] | As far as I understand, all this words about "both players" intend to cover the case when _white_ is to play, but during the intermediate moves _black_ has lost the right to castle (or vice versa, of course). | | (7) Posted by Joost de Heer [Wednesday, May 1, 2013 18:28]; edited by Joost de Heer [13-05-01] | In Siegfried's first position: No, it's not possible to claim the draw. That the move isn't legal doesn't change the fact that white has the ep -right-. So the position after 1...g5 is a different one than the one after 1... g5 2. Ke4 Rh8 3. Ke5 Rh5.
In Siegfried's second position, the answer to all four questions is 'no'.
Bonus question: How many times can you repeat Siegried's second position after 1...g5 without having a legal draw claim?
For those who have the Dutch version of 'Schaakcuriosa': Tim Krabbé talks about this phenomenon in the castle-chapter. | | (8) Posted by Sarah Hornecker [Wednesday, May 1, 2013 18:29]; edited by Sarah Hornecker [13-05-01] |
QUOTE ...the possible moves of all the pieces of both players are the same
Hijacking the thread again. I don't see how a situation would be the same if a potential move is prevented against a situation where that potential move does not exist at all.
QUOTE Positions are not the same if a pawn that could have been captured en passant can no longer be captured in this manner.
Why is this outlined in addition to the above, if not for this case?
Maybe we should ask Geurt Gijssen. | | (9) Posted by Joost de Heer [Wednesday, May 1, 2013 18:58] | Perhaps the following definition would be easiest to understand for a composer:
The game is drawn upon a correct claim by the player having the move, when the same position, for at least the third time (not necessarily by a repetition of moves):
a) is about to appear, if he first writes his move on his scoresheet and declares to the arbiter his intention to make this move, or
b) has just appeared, and the player claiming the draw has the move.
Positions as in (a) and (b) are considered the same, if the FEN-string has the same first four fields.
For the description of the FEN-string, see e.g. https://en.wikipedia.org/wiki/Forsyth%E2%80%93Edwards_Notation
For Siegfried's first example: FEN string after 1...g5 is '8/8/8/4KPpr/8/8/8/8 w - g6 0 1'. Continue with 2. Ke4 Rh8 3. Ke5 Rh5. The FEN string now is '8/8/8/4KPpr/8/8/8/8 w - 4 3'. Compare the first four fields, and you see it's a different string, so no repetition.
For those interested in the Krabbé article I mentioned earlier: I made a scan of the relevant page and placed it at http://sanguis.xs4all.nl/chess/repetition.jpg (text is in Dutch). | | (10) Posted by Andrew Buchanan [Wednesday, May 1, 2013 20:11]; edited by Andrew Buchanan [13-05-01] | @Siegfried: On e.p., Hauke explained that illegal+illegal=illegal. If the move cannot be done, then it doesn't matter whether it can't be done for one reason or for two. Having said that, it doesn't appear that the same kind of "phenotypical" approach is being used for castling. I really look forward to Hauke clarifying on that.
@Joost: thanks for the suggestion of FEN, but:
(1) What is "castling availability"? the term is not defined. i suspect it's castling rights - as that's the easiest thing to track. But we don't really know the right rule anyway. So really can't say whether FEN is appropriate for this field.
(2) The e.p. field would definitely be wrong. it's only tracking if a pawn has just moved twice. not even whether there's an enemy pawn now adjacent too it, let alone whether the capture would be legal.
If there was an online version of the Dutch text, anyone could use google translate, but I am not sure they do character recognition yet.
Thanks,
Andrew. | | (11) Posted by Steven Dowd [Wednesday, May 1, 2013 21:38] | Siegfried said:
Maybe we should ask Geurt Gijssen.
I don't think this is a bad idea, but I have followed his column for several years and find him often both perfunctory and contradictory. And perhaps that is simply the life of an arbiter, that is, someone who has to interpret rules in a practical manner. You are not given a lot of time for reflection. | | (12) Posted by Mario Richter [Wednesday, May 1, 2013 22:55] | This issue was also discussed in the Retros Mailing List: see the thread "Castling temporarily changed and repetition of position", http://www.pairlist.net/pipermail/retros/2007-October/002180.html. Furthermore, it was also discussed in some depth in Jörg Kuhlmann's article "Die neuesten Schachregeln der FIDE" in "Die Schwalbe", Heft 145 (Febr.1994), p.397ff. | | (13) Posted by Nikola Predrag [Thursday, May 2, 2013 02:49] | When the positions are surely not the same? If there is at least one possible line of play(variation) of any length, starting from one position, which can not be exactly repeated starting from the other position, then those two positions are not identical.
An exact definition of the identical positions would not be just the opposite one, there should be added few things, depending on the purpose of the rule. | | (14) Posted by Hauke Reddmann [Thursday, May 2, 2013 13:42] | @ Andrew: Exactly. In principle, the laws could be interpreted
along (1) (moves), (2) later potential moves, (3) rights.
I would always judge that (3) is "meant".
Note that e.p. is easier since (2) falls flat.
Any interpretation along (2) could lead to hair-raising paradoxa,
e.g. Black claiming a threefold and White can't castle because
any continuation leads to mating of Black. (That hair *has* been
raised in "a posteriority" retros.)
In any case, I know an international arbiter of highest rank
(my word may be law, but HE is GOD :-). Maybe he'll have a
concise answer.
Hauke | | (15) Posted by Andrew Buchanan [Thursday, May 2, 2013 17:28] | Thanks for the replies. I think the answer is at hand, hurray!
But first, yes I had already approached Chief Arbiter, Guert Gijssen. See: http://www.chesscafe.com/text/geurt170.pdf. I *think* he accepted my suggestion. I reproduce the text here, because I've noticed that Chess Cafe articles are sometimes substantially re-edited after the fact. I wish I'd spent a little more time on drafting my letter to him - I have made a few minor additions shown in square brackets below, which I don't think affect the sense.
"Paraphrasing Art 9.2 of the Laws, the state of a chess position comprises:
(1) Diagram? Effectively a function from 64 squares -> {WK/WQ/WR/WB/WN/WP/BK/BQ/BR/BB/BN/BP/empty}. Note that different copies of the same unit are not distinguishable.
(2) Whose move? {B/W}
(3) Castling rights, for WKside, WQside, BKside & BQside? Each can be {Yes/No}
(4) En passant, identifying the file if any as {a/b/c/d/e/f/g/h/No}
[At the beginning of the game set all castling rights are "Yes".]
Moving K sets both castling rights = "No" for that player.
Moving R sets one castling right = "No" for that player on that side.
To castle, a player must have castling right = "Yes" for that side [as well as the castling not being temporarily prevented by attacking or blocking unit]
[At the beginning of the game, set the en passant value to "No".]
If:
(1) a pawn moves double, and
(2) there is an adjacent opposing pawn, and
(3) it is not prevented from capturing e.p. [which can happen, if there is a check or pin, including the funny one through the pawn to be captured],
then set the en passant state value to the file of the moving pawn: "a.".. "h" as appropriate. In any other case, set the en passant state value to "No."
The position is now exactly characterized by (1)-(4) above. If the game returns to a situation where (1)-(4) are all the same as some prior state, then that's a repetition."
So essentially Guert endorses my option 3 in the base note. What had thrown me (and why I opened this thread) was reading the bizarre sentence about forcing moves. But I discovered today that we don't need to worry about that ever again! See: http://russiachess.org/upload/iblock/057/0573183c574e528168940e62eb291802.pdf. From 1-July-2013, we have new Laws, and the key rule becomes:
"Positions are considered the same only if the same player has the move, pieces of the same kind and colour occupy the same squares and the possible moves of all the pieces of both players are the same. Thus positions are not the same if:
a. at the start of the sequence a pawn could have been captured en passant.
b. a king or rook had castling rights, but forfeited these after moving. The castling rights are lost only after the king or rook is moved."
Perfect! So the Law for Draw by Repetition is now crystal clear. I don't think the intended interpretation has ever really changed over the years, but the phrasing certainly has. So three cheers for whoever came up with the new phrasing, and persuaded the rules committee of its rightness!
All the best,
Andrew. | | (16) Posted by Joost de Heer [Thursday, May 2, 2013 17:39]; edited by Joost de Heer [13-05-02] | I really disagree with
QUOTE
The e.p. case is easy:
The "right" to an illegal e.p. is nonesuch.
IMO, after a doublestep of the opponent you have the -right- to capture ep. Because there are other rules that prohibit executing this right (don't place your king in check) doesn't change that the right does exist. The same with castling: you have the -right- to castle, but there are other rules (intervening pieces, in check, crossing check) which could prohibit executing this right. | | (17) Posted by Andrew Buchanan [Thursday, May 2, 2013 18:17]; edited by Andrew Buchanan [13-05-02] | Hi Joost,
"Rights" are nice, aren't they? We bristle if we feel our "rights" are infringed. But really this is just emotional attachment, Captain Kirk, to which we Vulcans are impervious. :) The logical point is that there is some castling state which we need to keep track of throughout the game. This four-bit state is termed "castling rights" by the rules. On the other hand, en passant is always a transient concern - the rules do not need or use the term: "en passant rights".
There are three criteria which must be satisfied for en passant to be legal (1) double move just made (2) adjacent pawn (3) no self-check. You may choose to build the excitement by verifying these three criteria sequentially, and say that after you have verified the first two then you have some kind of "right" which may or may not be cruelly dashed when you examine the third criterion. There is no justification in the rules for this kind of drama.
But inventing a concept of "en passant rights" (as Guus Rol also playfully did a few years ago) is confusing, since it can lead someone to think that there is some significance between situations where different combinations of the 3 criteria fail. There is no distinction between failure cases: either the three criteria all succeed (position 1) or at least one fails (position 2). Those are the only two possibilities.
Please forgive my flippant tone tonight, but I am just so happy about the new Laws' handling of this issue.
All the best,
Andrew. | | (18) Posted by Georgy Evseev [Thursday, May 2, 2013 19:39] | Ok, to the practice.
See the following position, black king and rook unmoved.
(= 7+7 )
White play 1.e7 and after
A) 1...Kd7 2. Ka1 Ke8 3. Kb1 Kf7 4. Ka1 Ke8
or
B) 1...Ba7 2. Ka1 Bb8 3. Kb1 Ba7 4. Ka1 Bb8
want to claim a draw.
It seems to me that judge's decision will be different in A) and B), though black will never be able to castle in both cases (or, at least I tried to provide it so). | | (19) Posted by Sarah Hornecker [Thursday, May 2, 2013 20:01] | I'd give the draw based on 10.2 instead 9.2 since Black is obviously unwilling to win. *SCNR*
But back to the talk. Yes, I think that by word of the law the judges might need to make a different decision. However, if they are sane they give the draw in any case. This is certain to go for an appellation, so in the end there is something to cite from. | | (20) Posted by Geir Sune Tallaksen Østmoe [Thursday, May 2, 2013 21:21] | 10.2 may only apply if White has less than two minutes left and there is no increment.
And why would an arbiter give the draw in any case if they are sane? 10.2 is open to personal judgment, but I didn't think 9.2 was. Maybe the game is played with increments, and Black wants to have some additional time to figure out the best continuation, so he repeats the position as many times as he can without risking a 9.2 draw. Admittedly, it is unlikely that Black would need that in the given position... | | Read more... | Page: [Previous] [Next] 1 2 3
MatPlus.Net Forum General Law 9.2: Draw by Repetition |
|
|
|