Jump to content

Talk:Bishop and knight checkmate

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


How many moves?

[edit]

How many moves are required given best play?84.189.119.170 15:42, 25 February 2007 (UTC)[reply]

It depends on the position, but at the worst, the stronger side can checkmate in 33 moves with best play by both sides. I've added that to the article. Bubba73 (talk), 00:31, 1 May 2007 (UTC)[reply]

Occurrences in actual games?

[edit]

Is there an example of this in an actual game?84.189.119.170 18:10, 25 February 2007 (UTC)[reply]


I'm not sure of this but I think there was once a game between Kasparov and Anand which, from the final position, would have resulted in this endgame, but which didn't get that far because Anand didn't doubt that Kasparov would know how to win it, and hence resigned before reaching that stage. (unsigned)
Now there is a game with Polgar, not using the standard method and another game which does use the standard method. Bubba73 (talk), 15:29, 10 March 2008 (UTC)[reply]

How to?

[edit]

I hate to bring this up, after the AfD fiascos of last week, but this article does read quite like an instructional guide. The discussion at WP:NOT seeems to have quieted down, but it is not clear to me what the consensus was there. Nonetheless, this article could use some reworking to be more encylopedia-ic. Baccyak4H (Yak!) 19:57, 26 July 2007 (UTC)[reply]

I have to agree with you. This reads exactly like what I have in mind for a how-to. I feel there's too much detail given here for an encyclopedia article. I mean, even the reference work quoted (Fundamental Chess Endings) doesn't actually go into most of this - it explains the "W" manoeuvre and gives an example of how to drive the king into the corner that isn't actually explained in detail with words. And if a chess endgame encyclopedia doesn't go into this kind of depth, I don't think we should go overboard either.
Concrete suggestions:
  1. Sections 1 and 2 could simply be removed - an encyclopedia article shouldn't be teaching proper piece coordination. This is almost like saying "develop knights before bishops" and "castle early".
  2. Avoid the terms "active" and "passive defense" in section 3, since it's not established terminology. The lines can still be given but don't need nearly as much verbal explanation.
  3. The rest of the article seems all right.
In general, I think we should aim to inform, not to instruct. If we write articles as objective summaries of the existing body of chess knowledge rather than as our version of Dvoretsky's Endgame Manual and the like, that seems encyclopedic enough to me. youngvalter 21:44, 4 August 2007 (UTC)[reply]
This seems like a good first go through.
I believe it would be appropriate to describe that cooperation of the pieces to block off a contiguous (w.r.t. King moves) set of squares (in order to drive the defending King to the edge) is key to the process, would be acceptable, so long as the terminology used was sourced, and there were no diagrams with pretty virtual fences and how-to annotated lines as is now. That could take the place of the first two sections. Unfortunately, I only have the 1969 ed. BCE to work with, so cannot be of that much help. Baccyak4H (Yak!) 03:18, 5 August 2007 (UTC)[reply]
I don't think there is a problem with it. The article doesn't tell you how to do it - it tells you how it is done (i.e., third person pasive voice). It doesn't say "move your knight here then move your king there...". Bubba73 (talk), 15:50, 10 March 2008 (UTC)[reply]

Two more factoids...

[edit]

...cited from memory: 1) even some titled players are not able to execute this checkmate and must draw the game, and 2) nearly all chess programs do it without a hitch. Can't find a reference for either, though. GregorB (talk) 14:05, 29 April 2008 (UTC)[reply]

(1) is definitely true. GM Larry Evans once related (in a column in Chess Life and Review, I believe) that at the end of a long simul (that was his excuse, anyway) he was unable to mate in this ending. Once GM Dragoljub Janosevic, as I recall (or was it Kovacevic?), got this ending in a tournament game and proved unable to win it. Former World Champion Mikhail Botvinnik said that one had no business being a GM if one can't mate in this ending. I have to agree. I think (2) is right, also -- but yeah, good luck finding sources for these propositions. Krakatoa (talk) 01:23, 30 April 2008 (UTC)[reply]
(1) ChessBase turns up dozens of draws, some between players rated over 2400. Bubba73 (talk), 03:45, 30 April 2008 (UTC)[reply]
(2) Any program with even the four-piece endgame tablebase will have no problem doing it. Programs such as Sargon and early versions of Chessmaster could not do it.
I've never been on either side of this endgame. I doubt that I could beat a computer at it, but I probably could do it against a player weak enough that I'm two pieces ahead. I came close to it in a tournament once. I had B+N+P vs a minor piece. After three or four moves merrily supporting the advance of my pawn, I realized that if he exchanged his piece for my pawn, I'd be in trouble. I changed plans and forced an exchange of pieces. Bubba73 (talk), 04:01, 30 April 2008 (UTC)[reply]
I'm pretty confident I could win the ending against anyone (as a master, it would be a huge embarrassment if I couldn't). Once you learn a few basic positions cold (the one showing how to drive the king out of the wrong corner, and the ways to lock up your opponent's king with the N+B alone), it's pretty straightforward. I once had the wrong side of the ending in an online game with a 15-second-a-move (as I recall) increment. After 30 moves or so my opponent realized he was getting nowhere and offered the draw. Two knights vs. pawn -- now there's an ending where I wouldn't have a clue. Some Soviet master (Lilienthal?) had the winning side of that ending, with the pawn above the Troitzky line thrice in his career and drew every time. And if I'm recalling the name right, Lilienthal (who has beaten assorted World Champions) is a hell of a lot stronger than I am, probably even today at age 97 or whatever it is. Krakatoa (talk) 07:01, 30 April 2008 (UTC)[reply]
One point missing from the article is something that occurred with me. Knight+Bishop was about to occur, but with only a few moves to the adjournment my opponent knew that he couldn't claim the position as drawn so resigned. Maybe something like this could be covered in the adjournment article, as adjournments are quite a factor in many endings. SunCreator (talk) 07:46, 30 April 2008 (UTC)[reply]
I was right above about Lilienthal getting 2 N's v. P thrice, at least according to Botvinnik's recollection, which is what I had remembered. See Botvinnik's book Championship Chess, pp. 154-55 (analyzing Smyslov-Lilienthal, USSR Absolute Championship 1941 -- actually Botvinnik says that in this game the pawn was above the Troitzky line, so Lilienthal should have won, and he thinks L. had the same ending twice before, but doesn't expressly say whether the pawn was above the line in the prior endings). I gather Botvinnik and Lilienthal weren't big fans of one another. Botvinnik says what a fish L. is for misplaying the ending despite having it twice before. Botvinnik evidently pulled some strings to set up the "Absolute Championship" after L. and Bondarevsky had co-won, and Botvinnik had embarrassed himself, at the previous USSR Championship. There were no other "Absolute Championships" before or since. Oh, and I was wrong about L.'s age above. He's only 96, and won't be 97 for another five days. Krakatoa (talk) 08:10, 30 April 2008 (UTC)[reply]
The beauty of the wiki birth date and age template Andor Lilienthal. Lilienthal is a strong player but I don't really remember reading much about him. SunCreator (talk) 08:20, 30 April 2008 (UTC)[reply]
My daughter had the 2N vs P in a tournament when she was seven. (She gets into some strange ones.) I don't know if it was one of the winnable positions, but I was watching and just hoping she would not let the pawn queen. She did exchange a knight for the pawn. Bubba73 (talk), 13:59, 30 April 2008 (UTC)[reply]
That's amazing. I've never seen that ending personally. Had your daughter won it, she probably would have been future world-championship material. And yeah, Lilienthal is probably the strongest living player that hardly anyone has heard of: wins over six undisputed world champions! The record is Keres' nine. (Not sure where Lilienthal's win over Alekhine occurred -- his biography says it was in "a serious game," whatever that means.) Krakatoa (talk) 16:40, 30 April 2008 (UTC)[reply]
I was just hoping that he wouldn't let the pawn promote (and she didn't). In a tournament the same year, she had 2N+R vs. R (no pawns), and exchanged rooks (exchange when ahead), not knowing that 2 Ns can't checkmate. (I'm violating the rules of talk pages, so I need to stop.) Bubba73 (talk), 17:35, 30 April 2008 (UTC)[reply]

Missing "History" and "Impact" sections?

[edit]

In this case, there may be some history of the subject. For instance, who first realized that checkmate was possible? I doubt that is known. It must have been realized not long after the bishop got its current move. Was the technique given by early authors - Lucena, Ruy Lopez, etc? I don't know - does anyone have information?

As far as a section on the impact, it states that it is one of the basic checkmates. I can't think of anything else to say about the impact on the game. Bubba73 (talk), 01:48, 13 July 2008 (UTC)[reply]

The earliest example of B+N in the ChessBase database is a 1911 game, and the weaker side resigns after a few moves. Other than that, they go to the 1940s, I think. I don't think this is worthwhile to put in as "history". Bubba73 (talk), 03:12, 13 July 2008 (UTC)[reply]

A Faulty Example

[edit]

It appears to me that the example labeled "An Alternate Checkmate" could not possibly be attained through actual play, since the Knight must make its attack against the losing king in from a stalemated position (obviously not a legal play.) —Preceding unsigned comment added by 160.109.98.44 (talk) 21:58, 29 October 2009 (UTC)[reply]

abcdefgh
8
h8 black king
g6 white king
h6 white knight
d4 white bishop
8
77
66
55
44
33
22
11
abcdefgh
Checkmate position with bishop and knight
Agreed. SunCreator (talk) 22:09, 29 October 2009 (UTC)[reply]
abcdefgh
8
f8 black king
g6 white king
d4 white bishop
g4 white knight
8
77
66
55
44
33
22
11
abcdefgh

It can't be forced but it can be reached from this position: 1. Bg7+ Kg8?? 2. Nh6#. But your point is well-taken. I'll look into it more. Bubba73 (the argument clinic), 00:39, 30 October 2009 (UTC)[reply]

I replaced it with one that can be forced (with a reference). Bubba73 (the argument clinic), 00:52, 30 October 2009 (UTC)[reply]
Good, but the top diagram that I added here on the talk page has the same issue. SunCreator (talk) 00:55, 30 October 2009 (UTC)[reply]
I'll fix it. Bubba73 (the argument clinic), 01:00, 30 October 2009 (UTC)[reply]
My apology. Can see how it is forceble after all. SunCreator (talk) 01:03, 30 October 2009 (UTC)[reply]
Well, the new position is referenced and it is the actual final position in the "W maneuver" section. An improvement. Bubba73 (the argument clinic), 01:11, 30 October 2009 (UTC)[reply]

More draws

[edit]
With perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot immediately win one of the pieces (Müller & Lamprecht 2001:19), (Speelman, Tisdall & Wade 1993:7), and the position is not in the "stalemate trap" (see below).
Not true. There are many positions where the defending king can simultaneously attack the bishop and knight without the stronger side being able to recover. There are also a few drawn positions where the lone king, to move, would need at least three moves to reach any of the squares occupied by the opposing pieces. (See examples.)
abcdefgh
8
b4 white king
d4 white bishop
e4 white knight
c2 black king
8
77
66
55
44
33
22
11
abcdefgh
1) Black to play draws
abcdefgh
8
g7 white king
g5 white knight
h4 white bishop
f3 black king
8
77
66
55
44
33
22
11
abcdefgh
2) Black to play draws
abcdefgh
8
d4 white king
g4 black king
b2 white bishop
h1 white knight
8
77
66
55
44
33
22
11
abcdefgh
3) Black to play draws
abcdefgh
8
d5 white king
g4 black king
c1 white bishop
h1 white knight
8
77
66
55
44
33
22
11
abcdefgh
4) Black to play draws

Martin Rattigan (talk) 00:35, 1 December 2009 (UTC)[reply]

The statement in the article is with the normal assumption that White is to move. It also states the exceptions where the defender can win one of the pieces. Bubba73 (the argument clinic), 00:39, 1 December 2009 (UTC)[reply]
OK - I take the point about white to move, but I think the word "immediately" is misleading. After ...kf3 in the last two diagrams, it's white to move but black may still need another two moves to take a piece. —Preceding unsigned comment added by 77.58.249.114 (talk) 00:54, 1 December 2009 (UTC)[reply]
Sorry - forgot to add the twiddles to the previous para. I'm new to this. Martin Rattigan (talk) 01:05, 1 December 2009 (UTC)[reply]
OK. I guess it depends on what "immediately" means - it could be two moves if the knight is in the corner and the enemy king is two squares away diagonally. I would probably interpret "immediately" as being on the next move. I'll try to check if any of the references handle that case more specifically. Bubba73 (the argument clinic), 01:11, 1 December 2009 (UTC)[reply]
or three moves if the colours are switched and assuming white to move. Martin Rattigan (talk) 01:21, 1 December 2009 (UTC)[reply]
No, such things are normally given with the stronger side to move first. So far I've checked four references searching for a better wording of what they mean by "immediately" but I haven't found it. But since the knight is trapped, capturing it on the second move must be considered "immediate". I'll continue to check references. Bubba73 (the argument clinic), 18:31, 1 December 2009 (UTC)[reply]
Whatever the references mean, I would say the wording is unclear. If you published a mate in 2 under the caption, "White to play and mate immediately", I think you would confuse a lot of people. Similarly here. —Preceding unsigned comment added by Martin Rattigan (talkcontribs) 16:17, 2 December 2009 (UTC)[reply]
I agree that the use of "immediately" is unclear. I know that and you know that, but so far I don't have a reference clarifying it and just stating it on what we know is original research, which is not allowed. I'm working on it. Bubba73 (the argument clinic), 16:47, 2 December 2009 (UTC)[reply]
Sorry, I see the problem now. As I said I'm new to this. Martin Rattigan (talk) 16:55, 2 December 2009 (UTC)[reply]
On the other hand, if one of the pieces is lost, he no longer has a bishop and knight. I checked some more refereces, no clarification on this point yet, but I'm working on it. Bubba73 (the argument clinic), 17:05, 2 December 2009 (UTC)n[reply]
Just for interest here is another position (5) in which the weaker side draws and may need three moves (discarding convention) to take a piece - but not in a corner. (Sorry to be awkward.)
abcdefgh
8
f6 black king
c4 white king
a2 black knight
b1 black bishop
8
77
66
55
44
33
22
11
abcdefgh
5) White to play draws

Martin Rattigan (talk) 21:47, 2 December 2009 (UTC)[reply]

Yes,

  • 1. Kb3 Nc1
  • 2. Kb2

forks the pieces, so that is sort of covered in the article under the fork. But the lead section handles it OK "thirty-three moves from any starting position where the defender cannot quickly win one of the pieces". Bubba73 (the argument clinic), 21:57, 2 December 2009 (UTC)[reply]

I put it in mainly for interest. The lead section is now fine; I thought the note could benefit from a slight rephrasing. Martin Rattigan (talk) 22:32, 2 December 2009 (UTC)[reply]

By the way I think it would be safe to say, "... where the defender cannot win one of the pieces within three moves, given the move, or two moves with the stroger side to move", but of course that would need a reference. Martin Rattigan (talk) 22:43, 2 December 2009 (UTC)[reply]

Yes, we have no reference for those details but I think it is safe to say "quickly capture". Bubba73 (the argument clinic), 23:49, 2 December 2009 (UTC)[reply]

I have added a "citation needed" flag to the sentence, "With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces,[1] [2], (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves. and the position is not in the "stalemate trap" (see below).", pending replacement by *The longest win in this ending is 33 moves (Müller & Lamprecht 2001) harvcol error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help)."

From WP:OR: "Wikipedia does not publish original thought: all material in Wikipedia must be attributable to a reliable, published source. Articles may not contain any new analysis or synthesis of published material that serves to reach or imply a conclusion not clearly stated by the sources themselves." From WP:V: "Even if you're sure something is true, it must be verifiable before you can add it." "In Wikipedia, verifiability means that anyone using the encyclopedia can check that the information comes from a reliable source." From WP:OR again: "Even with well-sourced material, if you use it out of context, or to reach or imply a conclusion not directly and explicitly supported by the source, you are engaging in original research." That is what is done here.

I realize that there is an "urban myth" to the effect that the bishop and knight ending is won unless the lone king can immediately take a piece or fork both and that the sentence as it stands goes a long way towards dispelling this. It is however still not completely accurate, because it fails to take account of position (5) right. It's also could give the impression that situations with forks or knights trapped in the corner would generally be draws, but they are generally still white wins. In any case, it is not Wikipedia's job to clarify this. Clarity and accuracy must take second place to reliability and verifiability (though I wouldn't personally be replacing the material if it were clear and accurate). Martin Rattigan (talk) 19:40, 6 July 2015 (UTC) [reply]

"It's Déjà vu all over again." (1) In endgames, if one side is to win, it is assumed that that side moves first. That isn't the case in any of your five positions. (2) I think you are right about the stalemate trap being unreferenced. Bubba73 You talkin' to me? 20:09, 6 July 2015 (UTC)[reply]
(1) The positions I showed were all draws; neither side is to win. I am saying that you do not take account only of position (5), this is drawn, so there is no stronger side. If you want the side with the greater notional value of the material (Black) to move first, you can take position (5) after 1. Kb3. The same applies.
(2) You do take the stalemate trap into account explicitly. If this were an exception to a cited reference that failed to take it into account (it appears not to be), I believe that WP:OR would prohibit you from pointing this out.
The reasons for the proposed change are not related to the clarity or accuracy of the removed material. The reasons are as stated in the single paragraph here.
Martin Rattigan (talk) 08:59, 7 July 2015 (UTC)[reply]
abcdefgh
8
g7 black pawn
a6 black king
c6 white pawn
f6 black pawn
h6 black pawn
h5 white king
8
77
66
55
44
33
22
11
abcdefgh
Reti 1928. White to play and draw.
abcdefgh
8
g7 black pawn
a6 black king
c6 white pawn
f6 black pawn
g6 white king
h6 black pawn
8
77
66
55
44
33
22
11
abcdefgh
Reti-Bubba 2015. Black to play and draw.

Regarding your point about Déjà vu, your position regarding side to move has indeed complicated our discussions unnecessarily in many places.
I would invite you to consider how your criterion in (1) applies to drawn positions, if you think it applies at all.
I would also invite you to consider how it applies in positions of mutual Zugzwang. (You would seem to rule out any discussion of the mutual Zugzwang position shown below. Here if White moves first Black wins and vice versa so it cannot be the side to win who moves first, but there is one side to win in either case.)
abcdefgh
8
d5 black pawn
e5 white king
c4 black king
d4 white pawn
8
77
66
55
44
33
22
11
abcdefgh
Mutual Zugzwang. White to play loses. Black to play loses.
I give a famous and pretty endgame by Reti to the right. If you believe that Reti should have shown this with Black to move because his material is of greater notional value than White's, you would presumably show it as the position labeled Reti-Bubba. The point is then completely lost.
Martin Rattigan (talk) 10:56, 7 July 2015 (UTC)[reply]
(1) The one with the bishop and knight are the stronger side, (2) in position 5, black to move wins. That's all there is to it. Bye. Bubba73 You talkin' to me? 14:45, 7 July 2015 (UTC)[reply]
(1) In a drawn position there is no theoretically stronger side. You are consistently failing to see this.
(2) Position (5) is White to move, so this point is irrelevant.
Taken together your points apply equally to Reti's position. Do you think Reti also had it wrong?
Whatever, you offer no rebuttal of the paragraph detailing the reasons for the modification, so, in the absence of objections from any other user, I will go ahead and make the change.
Martin Rattigan (talk) 18:59, 7 July 2015 (UTC)[reply]

Bubba you have reverted my changes without any notification here and without any attempt to rebut the reasons I gave in this paragraph for making them. I will break the points down here. Please answer.
(a) The first sentence (or clause) I replaced is, "With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces,[1] [2],
(b) The references [1] and [2] you give are respectively (Müller & Lamprecht 2001) harvcol error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) and (Speelman, Tisdall & Wade 1993) harvcol error: multiple targets (2×): CITEREFSpeelmanTisdallWade1993 (help)
(c) The first reference gives only "Longest win: 33 moves.", which is much less than what you state.
(d) I don't yet have the second reference, but I know you changed the sentence to include the phrase, "where the defender cannot quickly win one of the pieces", in order to take account of the draws I gave, so this at least is adding material that is not in the second source either.
(e) What exactly is the citation from the second source?
(f) In view of these points I would say that the first sentence(clause) violates WP:V.
(g) I replaced this with, "The longest win in this ending is 33 moves (Müller & Lamprecht 2001).", which is a faithful representation of the source, and "Just under 90% of positions can be won against the lone king (Wilhelm by Raphael Andrist + Nalimov EGTBs).", which is also reliable and verifiable (the software is free).
(h) The second sentence (or clause) I deleted is, "(1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves. and the position is not in the "stalemate trap" (see below)."
(i) This appears to be an attempt to categorise all draws in this ending. (This fails as mentioned above, but that is beside the point.) There is no source for this, not even me, because I didn't say that all draws were of the type I gave, only that the positions I gave were draws. Such a categorisation without any reference is clearly prohibited by WP:OR. I know of no such reference so I just deleted this.
(j) You give, "I think that the original tells the situation better", as the reason for reverting. If the original research were corrected and clarified I might be inclined to agree, but the same could probably be said of well informed OR in many other areas. WP:OR is specifically designed to prevent this happening, as a policy.
(k) In view of the above, will you reinstate my changes, please?
Martin Rattigan (talk) 11:23, 8 July 2015 (UTC)[reply]
The above is too long to read, so I'm ignoring it. The problem is that you replaced what was referenced from reliable sources and replaced it with what appears to be your own research with a program and the endgame database. (See wp:or, wp:v, and wp:rs.) Did you consider only positions where White is to move? Because for 10% of the positions to be draws, they must be where Black can immediately capture one of the pieces, and that was covered in the original text. If no one responds on this talk page, take it up on the chess project talk page. Bubba73 You talkin' to me? 15:43, 8 July 2015 (UTC)[reply]
(A) "The above is too long to read, so I'm ignoring it."
You have a short version here to which you didn't respond, and a longer version here which you say you are ignoring. How long would you suggest?
(B) "The problem is that you replaced what was referenced from reliable sources and replaced it with what appears to be your own research with a program and the endgame database."
(1) My whole point is that what I replaced has either references that don't correspond with the text (segment (a)) or no references at all (segment (b)).
(2) I included the sentence "Just under 90% of positions can be won against the lone king (Wilhelm by Raphael Andrist + Nalimov EGTBs)." in an attempt to retain as much of the content of your original while providing a reliable and verifiable source. I agree this could also be interpreted as original research. I suggest you undo your revert and delete this single sentence. We can then discuss whether anything can be included regarding the draws without violating WP policy.
(C) "Did you consider only positions where White is to move?"
No, I took all positions.
(D) "Because for 10% of the positions to be draws, they must be where Black can immediately capture one of the pieces, and that was covered in the original text."
I agree with all of that. The problem is that none of the original text had any valid references. See suggestion under B.2 just above.
(E) "If no one responds on this talk page, take it up on the chess project talk page."
(1) The article talk page is the place WP policy documents recommend for discussing article content - see "Resolving content disputes" as one of many such recommendations.
(2) The discussion to date has involved historical references to this page. Without transferring the whole page to the project talk page the discussion would become unfathomable if we continue it there.
(3) I would suggest you respond here in the normal way. If you refuse to do so, I will open a new section on the project page, but simply refer you back here for the reasons mentioned in (2). I would prefer it if you do not put me to that trouble.
(F) Summary of suggestions.
(1) You undo your revert and delete the sentence, "Just under 90% of positions can be won against the lone king (Wilhelm by Raphael Andrist + Nalimov EGTBs).", if you feel this is necessary.
(2) You respond to (e) in the section you describe as too long. ((e) is only nine words.) This suggestion is necessary only if you won't agree that B.1 is accurate.
Martin Rattigan (talk) 21
01, 8 July 2015 (UTC)
  • I've been aware for a while of the disagreements on a few pages on endgames, but didn't intend to involve myself. The discussion is very long and many of the details are very technical, and I think others including those already in the discussion are more qualified to judge them. Most importantly, I think the key points are being brought up and I don't have anything new to add. Since other opinions were solicited at WT:CHESS and really only three editors have been hashing this out across several article talk pages, I thought I should add a few thoughts, as briefly as I can.
  • First, Bubba73 and Cobblet are correct in saying that WP:RS requires reliable published sources, and personal research using endgame databases cannot be included in a Wikipedia article. Editors have argued from time to time that analysis from chess engines should be permitted in articles, but they were and are wrong since unpublished material can't be used.
I have already agreed this point above. I have left Bubba to decide whether he wants to include it.
  • Bubba73 and Cobblet have also commented on how WP:OR does not allow editors to draw their own inferences from reliable sources and include them in articles, unless a reliable source supports the claim directly. This doesn't mean that we have to slavishly copy sources word for word (WP:COPYVIO), and editors with expertise in an article subject can and often do take advantage of their experience to write a better article, but they must remain within the bounds of Wikipedia policy. It does mean that we usually can't "fill in the blanks" if none of our sources has. Some people think this is a terrible thing, but really I don't think it's usually that bad. If those blanks really needed filling in, it's likely that a reliable source would already have done so. If our sources leave a thing unsaid, then generally we have to too. These points have already been made recently on chess endgame talk pages, but I just wanted to reiterate that it isn't just two editors who understand Wikipedia policies in this way.
I totally agree. I did insert what I believe is a faithful representation of the cited source, without this being a verbatim copy. The sentence it replaces bears little relation to the cited source and appears to have been mostly of Bubba's own invention. It is also incorrect or incomplete which is my main reason for wanting to replace it.
Pertaining to the subject matter in the sentence, (Müller & Lamprecht 2001 - Kindle copy) says exactly and only, "Longest win: 33 moves".
I used the sentence, "The longest win in this ending is 33 moves (Müller & Lamprecht 2001)."
I will go into some detail on the sentence it replaces, because this is actually the only sentence I want to change in this article. I reproduce the versions to date below.
Version (1) is an "urban myth" that has been circulating in chess circles since long before Wikipedia was founded. Except for beginners, anyone repeating it would normally add, "or the pieces can be forked". This is presumably where Bubba got it from. I don't believe it could appear in any reliable source, and it certainly does not occur in the cited source.
(Speelman, Tisdall & Wade 1993:7) was added in version 3 (without any content change) but from the comment on the edit it appears that this was intended as a citation for the 33 move diagram and is actually misplaced here.
Version (4) appeared after our first discussion. In version (4) the phrase "the defender cannot immediately win one of the pieces" was replaced by "the defender cannot quickly win one of the pieces" and a footnote added detailing the ways this could be possible. The footnote is incomplete, missing positions such as (5). This is my only objection on technical grounds. Had the footnote been omitted I believe the sentence would have been correct, but this is only a personal opinion. I still wouldn't include it without a valid citation.
All the above versions were produced by Bubba.
The citations were shrunk to "ref" format and the footnote appended to the sentence in version (6).
Version (6) was produced by Amortias after a 5.5 year period without change. This change occurred shortly after Bubba first broke off communication in our renewed discussions in the "Two knights endgame" talk page.
Bubba did not revert to a version last changed by himself, but since the content remains unchanged by Amortius' amendments (only the format changes), he did revert to material of his own authorship.
________________________________________________________________________________
(1) Bubba 1 May 2007 (# moves)(new ref format)
Checkmate can be forced in at most thirty-three moves from any starting position (unless the defender can immediately win one of the pieces) (Müller & Lamprecht 2001:19).
________________________________________________________________________________
(2) Bubba 23 Dec 2007 (→‎A stalemate trap: change captions, link)
Checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot immediately win one of the pieces (Müller & Lamprecht 2001:19), and the position is not in the "stalemate trap" (see below).
________________________________________________________________________________
(3) Bubba 15 Apr 2009 (add ref for position requiring 33 moves)([ .. ])(move ref to intended place)
With perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot immediately win one of the pieces (Müller & Lamprecht 2001:19), (Speelman, Tisdall & Wade 1993:7), and the position is not in the "stalemate trap" (see below).
________________________________________________________________________________
(4) Bubba 2 Dec 2009 (reword to clarify and put details in a footnote. I'm still searching for a reference)
With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces (Müller & Lamprecht 2001:19), (Speelman, Tisdall & Wade 1993:7),}},[1] and the position is not in the "stalemate trap" (see below).
Footnote;
(1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves.
________________________________________________________________________________
(5) Amortias 8 Jun 2015 (→‎top: Adding/improving reference(s))
With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces,[2] (Speelman, Tisdall & Wade 1993:7) harvcol error: multiple targets (2×): CITEREFSpeelmanTisdallWade1993 (help),[3] and the position is not in the "stalemate trap" (see below).
Footnote:
(1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves.
________________________________________________________________________________


(6) Amortias 8 Jun 2015 (→‎Notes: Fixing style/layout errors)(→‎top: Adding/improving reference(s))(→‎References: Clean up)
With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces,[1][2] (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves. and the position is not in the "stalemate trap" (see below).
________________________________________________________________________________


In view of the fact that no version of the sentence I attempted to replace has been fully correct over its 8 year history and each has been bogusly attributed to (Müller & Lamprecht 2001) for all of that time and (Speelman, Tisdall & Wade 1993) for the last 6 years, I believe that Wikipedia could be on shaky legal ground. Because of Wikipedia's large readership it is conceivable that the reputations of the authors involved could be damaged and that sales of the books referred to could be adversely affected.


  • Bubba73 is absolutely correct when he says that in chess, "to win" positions are almost universally given with White to move, and in theoretical endgames White is the side with the material advantage. That's the convention, and it's surprising to see you return to this page after 5.5 years and bring up some of the same points again that Bubba73 already explained a half decade ago. If this were the only page under dispute it wouldn't be a big matter, but when you propose major changes to several articles at the same time, it can be a lot to deal with.
abcdefgh
8
d7 white king
d5 white pawn
h4 black king
b1 black knight
8
77
66
55
44
33
22
11
abcdefgh
J. Moravec 1938. White to play and win.
abcdefgh
8
c7 black pawn
h7 white king
b6 black king
c3 white pawn
d3 black pawn
d2 white pawn
8
77
66
55
44
33
22
11
abcdefgh
Bianchetti 1925 White to play and win.
(a) Bubba is absolutely incorrect, as are you. This is not a matter of opinion, but fact. I find it astounding that two editors of the Wikipedia chess pages should be so ignorant of normal conventions. I reproduce two sample positions right contradicting what you say. There are thousands upon thousands of others in the literature.
(b) I have said several times in this talk page and many times in the two knights talk page, Bubba's talk page and the project talk page, that the positions under discussion are draws. They are not "to win" positions. Therefore what you say has no relevance. It is this constant re-presentation of the same arguments to which I have already responded, with apparently no attempt to read or reply to my responses that is inflating these discussions beyond reason. Please read the discussion, at least from, say, here.
(c) I didn't return to this argument. Bubba did here
(d) I could have added in (c), "and here and here and here and here ...", but I will leave you to trawl through the mess it's created on the talk pages to locate these. — Preceding unsigned comment added by Martin Rattigan (talkcontribs) 09:44, 11 July 2015 (UTC)[reply]
  • The sheer volume of your talk posts and the breadth of your changes (it has the appearance of saying "everything here is wrong and I demand a rewrite nearly every paragraph in several established articles") proposed all at once, is almost overwhelming and probably discourages responses. As a pragmatic matter I think you might get more useful dialog if you would focus on a small part of one or at most two pages. Maybe start with a single point that you think could be clarified or corrected, but make sure the correction is supported by one or more reliable sources.
The reason for the volume of the discussion is as stated previously.
If you were to read the discussions you are commenting on you would see I propose.
(i) In the "Bishop and knight" article to replace a single incorrect and unsourced sentence by a correct sourced sentence.
(ii) In the "Two knights" article to
replace a single incorrect and unsourced sentence (the Troitzky line rule itself) by a correct sourced sentence
move a single sourced sentence from the lead to replace a duplicate unsourced version in the body
add a single sentence with a single explanatory diagram to balance the sentence that was moved.
To say that you are exaggerating in your preceding paragraph would be putting it rather mildly.
  • I'll end with my opinion that the chess endgame articles are generally good, many of them even very good, so I don't think they are the best place to begin editing chess articles. This is especially true when proposing major rewrites. There are many poor chess articles begging for attention, some so bad that nearly any change at all would have a good chance of being an improvement. I think a dedicated editor could make a bigger contribution to improving coverage of chess on Wikipedia by giving the weak articles on important chess topics some attention, rather than insisting on major rewrites of articles that are already pretty good. On Wikipedia just as in the chess world, the openings get all the attention, and the endgame gets what little is left. Our articles on the middlegame, strategy and tactics are pretty rough. I'm not saying that the endgame articles can't be improved and that you shouldn't try to improve them, but you might get more done if you started smaller and got other editors to agree to a few little improvements before you begin to rewrite entire articles. Quale (talk) 01:43, 10 July 2015 (UTC)[reply]
I'm not intending to become a dedicated editor. My main intention is to stop Wikipedia from propagating the "urban myth" I mentioned earlier in relation to this article and to reverse a new one that Wikipedia seems to have already established in relation to the Troitzky line.
Martin Rattigan (talk) 18:40, 10 July 2015 (UTC)[reply]
Please read WP:TALK. For starters, don't mix your responses in the middle of other people's posts. (WP:TPO: "It is generally inappropriate to Add replies in the middle of another editor's post."). You can quote the material you are replying to or we can number my paragraphs, but if you hope to collaborate here you're going to have to stop violating talk page norms immediately. Beyond that, I disagree with pretty much everything you wrote, certainly on the matters of fact and also on the matters of opinion. But you've lost me completely when you wrote: "I'm not intending to become a dedicated editor. My main intention is to stop Wikipedia from propagating the "urban myth" I mentioned earlier". Great. It's good to know that you aren't really here to improve the chess articles, instead you're on a crusade to WP:RIGHTGREATWRONGS. I don't think this is going to work out. Quale (talk) 04:29, 12 July 2015 (UTC)[reply]
In interspersing my replies, I was following the lead of far more experienced editors than I, for example Cobblet's reponses to my points here. In this instance my points were numbered.
While this practice is against the WP:TALK guidelines, when you insert a large amount in your post and do not number either the paragraphs or the multiple points therein, as you did here, it is very difficult to respond to the individual points without breaking the guidelines. A full answer would involve reproducing your large post in addition to the responses. ("Editors should attempt to follow guidelines, though they are best treated with common sense, and occasional exceptions may apply. WP:Policies and guidelines.)
If you disagree on matters of opinion then we can agree to disagree if you don't want to discuss them further. But when you say, "I disagree with pretty much everything you wrote, certainly on the matters of fact", while not mentioning any particulars, it seems that you are just hoping that any potential readers of the discussion will be gullible enough to assume that the details can safely be ignored. It also paves the way for continued filibustering.
I invite you to give reasons for disagreeing with any particular facts I introduced above, and also, in the case of facts you asserted, where I disagreed, your reasons for disagreeing with the rebuttal I gave.
As far as your response, "But you've lost me completely ..." goes, I have to say you get very easily lost. My main intention is to correct two particular incorrect sentences in the articles under discussion. In my view this will be an improvement to the site.
Martin Rattigan (talk) 12:30, 12 July 2015 (UTC)[reply]
I don't assume that readers will be gullible. I said my bit, you replied. I'm confident enough in what I wrote to allow anyone who reads this to consider what is on the page and make their own determination. (I don't expect that many people will read this, since I think most people can find better uses of their time.) I made my brief reply to your response only to make it clear that I still don't agree, since sometimes people assume that silence equals asset. I think discussion about points of disagreement can be healthy and often it benefits the encyclopedia, but since I don't enjoy interaction with you I'm going to limit it as much as I can. In particular I don't think getting caught up in a drawn out tit-for-tat discussion of every point on which we disagree would be helpful to us or to anyone else, nor do I think it would improve the encyclopedia. I understand that it can be frustrating to have someone say that you're wrong and then refuse to engage in a dialog to discuss the specifics in detail, but I'm afraid that when you are unpleasant sometimes people aren't going to want to talk to you. Quale (talk) 16:54, 12 July 2015 (UTC)[reply]

The sentences that Bubba73 reinstated now read:

With the stronger side to move and with perfect play, checkmate can be forced in at most thirty-three moves from any starting position where the defender cannot quickly win one of the pieces.[4][5] The exceptions occur when (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves, and the position is not in the "stalemate trap" (see below).[citation needed] These exceptions constitute about 0.5% of the positions.[6]

The clause "The exceptions occur when" and the last sentence were added by Bubba73 after the reversion, at 00:07, 11 July 2015‎.

The first sentence is jointly attributed to Müller & Lamprecht (2001:19) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) and Speelman, Tidsall & Wade (1993:7) and the remainder to Speelman, et. al:7).

The actual quotations in the cited sources are as follows:

Müller, Lamprecht & 2001:1.4
Longest win: 33 moves.
Speelman, Tidsall & Wade (1993:7)
(Table entry)
Table of Computer Database Results
Material Longest Win Target Position Percentage Winning Starting Position Analyst(s) Year
BNvK 33 M 99.5 Ka8,Be8,Nh2 v Kc8 BAV 1981
p8. White to play in all cases except QPvQ, QvRP and RPvP.

Neither source says, "With the stronger side to move and with perfect play, checkmate can be forced ... from any starting position where the defender cannot quickly win one of the pieces", or anything similar.

Similarly the second source says nothing like, "The exceptions occur when (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves, and the position is not in the "stalemate trap"

Again the second source doesn't anywhere say that the list of "exceptions" that Bubba73 gives account for 0.5% of the positions. The list nowhere appears.

Further the last sentence of the paragraph; "Although this is classified as one of the four "basic checkmates" (the others being king and queen; king and rook; or king and two bishops against a lone king), it occurs in practice approximately only once in every 6000 games.[7]", is also bogusly referenced.

M&L include five endings in section 1. Basic Mates and give a frequency of occurrence of one in 5000, not one in 6000.

I will therefore delete most of the material from the lead and add a properly cited set of statistics in a section following the lead. I will first delete the bogusly sourced material, then wait a week or so before inserting the statistics section in case the first change is again reverted to Bubba73's text.

84.73.201.218 (talk) 18:09, 24 July 2015 (UTC)[reply]

It is on PAGE 19 of M&L - that is what the reference means. The references are by page numbers, not sections. Bubba73 You talkin' to me? 23:35, 24 July 2015 (UTC)[reply]
What exactly is on p19 of M&L? So far as I can see the only thing relevant to the deleted text in the whole of the book is the sentence I quoted above. Are you saying what I wrote above is inaccurate? If so where? If not why have you reinstated your text yet again? Martin Rattigan (talk) 02:15, 25 July 2015 (UTC)[reply]
Please state clearly and concisely (one sentence) what you think the problem is. We don't have time to read your long, rambling drivel. Bubba73 You talkin' to me? 02:28, 25 July 2015 (UTC)[reply]
Clearly and concisely you reinstated four sentences, each of which is your own original research and each of which is given bogus citation(s).
PLEASE REINSTATE MY CHANGE and stop wasting both our time. Martin Rattigan (talk) 02:44, 25 July 2015 (UTC)[reply]
Once again, NONE of that is my original research. Bubba73 You talkin' to me? 03:16, 25 July 2015 (UTC)[reply]
It was added by your id according to the history file. Martin Rattigan (talk) 03:25, 25 July 2015 (UTC)[reply]
But it is NOT my original research. Bubba73 You talkin' to me? 03:35, 25 July 2015 (UTC)[reply]
Well it's certainly not ascribable to the sources you cite, and some of it is incorrect, so will not appear in any reliable source. So whose original research is it if not yours? Perhaps you could scan p19. of your paper version of M&L so we can see what the "it" is that you say is there. Martin Rattigan (talk) 03:43, 25 July 2015 (UTC)[reply]
Ask for a third opinion. Bubba73 You talkin' to me? 04:20, 25 July 2015 (UTC)[reply]
OK. In the meantime can you make the scan available, so we can be sure we're not talking at cross purposes , please? Martin Rattigan (talk) 04:24, 25 July 2015 (UTC)[reply]
This - https://www.dropbox.com/s/vfa36du2hceeuwz/MLextractKBNK.png?dl=0 - is the Kindle version of the page. I've highlighted the sentence I quoted above. Martin Rattigan (talk) 10:32, 25 July 2015 (UTC)[reply]
Karsten Muller in his first FritzTrainer Endgame DVD says "With optimal play, you need from the worst position thirty-three moves. So accuracy is required, otherwise your opponent will claim a draw based on the fifty move rule". Jkmaskell (talk) 11:32, 25 July 2015 (UTC)[reply]
And I believe he was perfectly correct. That the longest mate is 33 moves is not in dispute; it's published in the cited source. What is in dispute is the rest of what Bubba73 has inserted and twice reinstated, none of which is in the cited sources. Martin Rattigan (talk) 13:02, 25 July 2015 (UTC)[reply]

"M&L include five endings in section 1. Basic Mates and give a frequency of occurrence of one in 5000, not one in 6000."

They give the figure of 0.02%, which is rounded off to one significant digit. That does give a figure of 1 in 5,000. However, on that page they give the more accurate figure of 283 in 1,700,000 games, which is 1 in 6,007. But rounding that properly gives 1 in 6000. It is a routine calculation. Bubba73 You talkin' to me? 21:01, 25 July 2015 (UTC)[reply]

1. You do not address the first half of the sentence. There it is implied that there is a generally accepted classification of "basic checkmates" numbering four and Müller & Lamprecht (2001:11) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) is cited as a source for this. It is not.
2. Concerning the frequency of occurrence of the kbnk ending, this is a minor point compared with all the discrepencies between text and citation that you reinstated and have not anywhere addressed, but I believe you are also wrong on this point. In detail:
(i) M&L gives the number of games in the database as 1,687,182 so the ratio 283 to this figure is 5962 to the nearest integer. If this is rounded to the nearest hundred or the nearest thousand this would give one in 6000. Müller & Lamprecht (2001) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) give all the frequences to two decimal places of one percent, so it is apparent that the figure given is within 0.005% of the calculated figure. No indication of how the figure was "properly" rounded is given in the article.
(ii) When you originally changed the format from 0.02% to one in 5000, if you were to preserve the information from Müller & Lamprecht (2001) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) intact, you should strictly have said, "somewhere between one in 4444 and one in 6667".
(iii) In attempting to correct (ii) you have taken the ratio betwen the figures 283 shown under "quantity" to 1,700,000. (Müller & Lamprecht (2001) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) say that the ChessBase, Mega Database 2001, "contains nearly 1.7 million games".)
What you have apparently overlooked is that in the same section M&L say, We searched for endings that were on the board for at least two half-moves (a very short time: just one move by each player). This means that the figure of 283 you used is , where is the number of knbk cases found and is the number of these that lasted less than two half-moves.
You have used the formula , where is the number of games in the database, but if you are going to discount from the sample, these at least should be discounted also from the denominator in calulating the frequency, as in , but it is probable that the formula actually used was closer to , where is the total number of cases discounted for all endings. Karsten Müller is a PhD in mathematics as well as a GM and Frank Lamprecht is an IM, so I think whatever formula they used to arrive at the published figure of 0.02% can be assumed to be correct.
The fact is that neither the formula used nor individual values for and nor any value for actually appear in the text, so it is not possible to give an accurate figure using the method you have attempted.
You are allowed to perform routine calculations, but these have to be obvious, correct, and a meaningful reflection of the sources. Otherwise it is original research.
(iv) I can accept that this particular mistake was made in "good faith".
(v) I have uploaded copies of the relevant M&L (Kindle) pages here - https://www.dropbox.com/sh/fd1fvs72hi5jif1/AABqs6k40Et849BTEupRqlV4a?dl=0.
(vi) In my proposed "statistics" section, I will give the figure 0.02%, exactly as published.
3. In focusing on point 2. you are ignoring "the elephant in the room", namely that none of the text I deleted and that you have twice reinstated corresponds with the sources to which it is attributed.
(i) You give as your reason for reverting my change, (Not a bogus source - it is on PAGE 19 of Muller and Lamprect, just as it says).
The text in question is variously attributed to Müller & Lamprecht (2001:19) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help), Müller & Lamprecht (2001:7) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help), Müller & Lamprecht (2001:11) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help), Speelman, Tidsall & Wade (1993:7) and Speelman et. al:7).
What exactly is the "it" on page 19 of M&L that explains all these bogus references? Even were all the text you have composed actually contained on said page 19, most of the references would still be bogus, because they do not refer to this page. In fact, the only relevant statement on that page appears to be exactly and only, Longest win: 33 moves.
(ii) I retained the statement Checkmate can be forced only with the defending king in a corner controlled by the bishop or on a square on the edge next to such a corner. Apart from the first sentence (which is unexceptionable) this was the only part of the first paragraph for which I could find an alternative valid citation (Fine 1941:4). In reinstating your text you have removed this citation leaving it unsourced or bogusly attributed to Müller & Lamprecht (2001:11) harvcoltxt error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help) - it is unclear which.
(iii) You have so far declined to quote or provide scans for your cited sources. Can you, please, at least confirm that the quotes I gave above and the scans contained in the two links I gave above correspond with your paper versions?
4. You suggest that I ask for a third opinion. I can do this, but given that you must be perfectly well aware that the reasons for the deletions were valid and your reason for reinstatement without any foundation, do you really want to waste somebody else's time as well? I suggest it would be better for all concerned if you just undo your reversion.
Martin Rattigan (talk) 17:36, 26 July 2015 (UTC)[reply]
"You do not address the first half of the sentence. There it is implied that there is a generally accepted classification of "basic checkmates" numbering four"
There are quite a few books that have a chapter on "basic checkmates" - I'll provide one as a reference. I'm having computer problems right now, but I should get to it tonight. Bubba73 You talkin' to me? 22:54, 26 July 2015 (UTC)[reply]
No indication of how the figure was "properly" rounded is given in the article
Not necessary. Bubba73 You talkin' to me? 22:54, 26 July 2015 (UTC)[reply]
There are indeed quite a few books that have a chapter on basic endgames, but probably no general agreement. M&L and Fine both give five, but these are not the same five. If you wish to make the point then you should give a selection (with valid references). Is the bishop and knight ending the best place to do this?
An indication of the error range is probably not so important, but the error range is clear in the source and if the same error range is not given in the quote, the quote is not exactly what is cited. The obvious answer is to report it as a frequency of 0.02% as I will do in the proposed statistics section.
You still haven't agreed the quotes I've provided from the sources - this shouldn't really be too hard. What's the problem?
Martin Rattigan (talk) 00:17, 27 July 2015 (UTC)[reply]
Fine gives only four checkmates in the beginning. There is a section about two knights, but it isn't a checkmate.
You write too much nit-picky stuff for me to possibly read. If you want to use a reference for four basic checkmates, you can. If you want to discuss the 1 in 6000 calculation, you can. You don't have to keep pestering me about every trivial thing. Bubba73 You talkin' to me? 01:28, 27 July 2015 (UTC)[reply]
As I already said in 2. above the game frequency is a minor point. I don't know why you are focusing on this. You're still ignoring the elephant in the room.
I want you to use a valid reference for four basic checkmates if you wish to keep it in. It's your text. I see you have just now added Fine and Benko as a citation for the existence of four basic checkmates, but you've put this in the middle of the sentence. Is it also meant to be a citation for the list that follows? Since there is a lack of agreement among authors regarding which endgames are basic or elementary, the sentence is in any case a violation of WP:NPOV.
For the record Fine "Basic Chess Endings" gives five elementary mates. See this link: https://www.dropbox.com/s/vsevazuuk107wqn/Fine_contents.png?dl=0
Could you please refrain from further editing of the first paragraph until the content dispute we have is settled. This only serves to obfuscate the argument to date, because this is related to the references in text you reinstated. If, as seems increasingly likely, it will be necessary to go through formal dispute resolution, changes to this text will make it harder for anybody dealing with the case to follow.
You appear to have overlooked a point in your reply. I say again - You still haven't agreed the quotes I've provided from the sources - this shouldn't really be too hard. What's the problem?
Martin Rattigan (talk) 04:04, 27 July 2015 (UTC)[reply]
That is an OLD version of Fine - read the one I REFERENCED. And quit pestering me with your trivia. Bubba73 You talkin' to me? 04:52, 27 July 2015 (UTC)[reply]
You don't reference Fine anywhere.

In the first paragraph: "Although this is classified as one of the four basic or elementary checkmates (Fine & Benko 2003:1)". And it is listed in the references. You can click on the blue link to take you to the reference.Bubba73 You talkin' to me? 14:58, 27 July 2015 (UTC)[reply]

That's Fine & Benko. We both said Fine. What I uploaded was also Fine (not Fine & Benko). Martin Rattigan (talk) 15:17, 27 July 2015 (UTC)[reply]

References

  1. ^ (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves.
  2. ^ Müller, Karsten (2001). Fundamental Chess Endings. Gambit Publications. p. 19. ISBN 1-901983-53-6.
  3. ^ (1) The defending king may be forking the bishop and knight so that one of them is lost on the next move, or (2) the knight may be trapped in a corner by the defending king and the knight is lost in one or two moves.
  4. ^ Müller, Karsten (2001). Fundamental Chess Endings. Gambit Publications. p. 19. ISBN 1-901983-53-6.
  5. ^ Speelman, Jon; Tisdall, Jon; Wade, Bob (1993). Batsford Chess Endings. London England: B.T. Batsford. p. 7. ISBN 978-0-7134-4420-9.
  6. ^ Speelman, et. al, p. 7
  7. ^ Müller, Karsten; Lamprecht, Frank (2001). Fundamental Chess Endings. Gambit Publications. p. 11. ISBN 1-901983-53-6.

Error in "W" maneuver section

[edit]
abcdefgh
8
b3 black knight
c3 black king
c2 black bishop
c1 white king
8
77
66
55
44
33
22
11
abcdefgh
mate, but it cannot be forced

The claim that checkmate can only be accomplished in the same color corner as the bishop is incorrect. With the White King on c1, mate can be achieved with Bc2, Kc3, and Nb3... —Preceding unsigned comment added by 70.186.196.209 (talk) 05:46, 1 December 2010 (UTC)[reply]

Checkmate can only be forced near the corner that the bishop controls. It can't be forced in the other corners - the defender would have to blunder into it. Bubba73 You talkin' to me? 05:50, 1 December 2010 (UTC)[reply]
For instance, in that position, Black's last move was ...Nb3#. But the white king had to move to c1 from b1, b2, or d1. If he was on d1 he should have moved to e1 instead of c1, etc. Bubba73 You talkin' to me? 07:13, 1 December 2010 (UTC)[reply]

Original 1749 Philidor treatise does NOT cover B+N endgame

[edit]

Here is the current Wiki text:

A method for checkmate when the lone king is in the corner of the opposite colour from the bishop, using the "W" method, was given by Philidor in his famous 1749 treatise, Analyse du jeu des Échecs.[1]

There are at least 3 problems here:

  1. The 1749 treatise has only 1 endgame in it (R+B vs R).
  2. The reference here takes you to an updated edition dated 1803(?).
  3. The actual title to the 1749 treatise is: "L'Analyze des Echecs"[2]

The updated and expanded 1777 English edition (also in the 1777 French edition), "Analysis of the Game of Chess" [3] on p 244 describes the checkmate method with B+N.

TommyWP (talk) 18:09, 2 May 2016 (UTC)[reply]

Could you update with the correct information please?

Also, do you know of an English translation of Deletang's method? I'm planning to replace this in the section "Closed Positions" with some accurate analyses, but I think a section on Deletang's triangles should be retained for historical interest. Unfortunately my French isn't up to the original, but from the number of mots in it I would say it is more comprehensive than appears on the site. The method can't be corrected and remain Deletang's method, but at the moment I don't know if the wrong moves that appear are original or Pandlfini's. — Preceding unsigned comment added by Martin Rattigan (talkcontribs) 19:52, 9 May 2016 (UTC)[reply]

OK - I've made the change. I think it's correct now. Martin Rattigan (talk) 20:44, 14 May 2016 (UTC)[reply]

"Open" and "closed" positions

[edit]

Who invented this terminology? I've never seen these words used in the context of this endgame before. Cobblet (talk) 06:07, 29 May 2016 (UTC)[reply]

The terms are local to the article, not in general use. I should have made that clear.
I have a substantial update in preparation (almost complete) in which the footnote:
The terms "open position" and "closed position" are local to this article and used for ease of exposition. They are not in general use. Outside the context of this article the terms would normally refer to positions from open or closed games.
appears against both the terms when introduced and the subsequent headings.
This should appear in the next few days. — Preceding unsigned comment added by Martin Rattigan (talkcontribs) 09:23, 30 June 2016 (UTC)[reply]

It's taking me longer than planned to complete the update, so I've copied the footnotes to the existing version in the meantime. Martin Rattigan (talk) 10:45, 19 August 2016 (UTC)[reply]

Still taking me longer than planned, but in the meantime I've changed diagram J to be more typical and added the game that the legend referred to. I've also added a section of PGNs because I think some readers may have difficulty visualising the longer sequences. For the moment I've included PGNs only for sections that are not expected to change in the new version (in which the section should be more complete). — Preceding unsigned comment added by Martin Rattigan (talkcontribs) 02:50, 28 September 2016 (UTC)[reply]

Also put in the updated version of the Sierawan section to connect it with the PGN. Martin Rattigan (talk) 03:22, 28 September 2016 (UTC)[reply]

@Martin Rattigan: Anyone wanting PGN's can just use a site like chessgames.com or chess.com to look up these games. Per WP:NOTDIR Wikipedia is not a repository for game PGN's; this is like asking for the best track files of a tropical cyclone to be posted in every storm's article because the reader "has trouble visualizing", and that situation is even more difficult to visualize.--Jasper Deng (talk) 08:23, 12 February 2018 (UTC) @Jasper Deng:Very constructive I'm sure. As I said in the comments to my original reinstatement you've left a heap of defunct pointers to the section you've removed in the body of the text. Please sort it out. I'm not going to. Martin Rattigan (talk) 02:19, 19 April 2018 (UTC)[reply]

POV?

[edit]

In the Grandmasters_failing_to_mate section, the assertion that "The superior side had no idea how to win" isn't supported by the citation. While the game record on chessgames is clear evidence that Epishin failed to win the game, saying he "had no idea how to win" seems a bit harsh, given that at some points he was in the correct direction. — Preceding unsigned comment added by Rcaetano (talkcontribs) 05:15, 13 December 2018 (UTC)[reply]

I agree and have changed it. Clearly he had an idea of how to win, he just didn't know the technique. Bubba73 You talkin' to me? 05:30, 13 December 2018 (UTC)[reply]

references

[edit]

Error in diagram F at the top of the page?

[edit]
F.
abcdefgh
8
a8 white king
c8 black king
e8 white bishop
h2 white knight
8
77
66
55
44
33
22
11
abcdefgh
White checkmates in 33 moves against best defense (Müller & Lamprecht 2001:400) harvcol error: multiple targets (2×): CITEREFMüllerLamprecht2001 (help)

This isn't a checkmate since the king isn't in check. Seems like the knight was most likely intended to be at A7? Unless this was supposed to depict a starting state, in which case it's very out of character with the other diagrams and should probably be labelled more clearly. Hope that makes sense, sorry if I've missed something obvious.

SpectatorRah (talk) 20:08, 28 November 2019 (UTC)[reply]

white checkmates in 33 moves from that position, which is one of the worst starting positions for White. Bubba73 You talkin' to me? 19:24, 29 November 2019 (UTC)[reply]

Diagrams T, U, V are incorrect?

[edit]

I believe the X’s in diagrams T, U, V are incorrect, and the boundaries of the triangles are in error. I’d fix them myself, but I came here trying to learn this endgame, and having never played it correctly, am not really qualified to edit it! — Preceding unsigned comment added by RPellessier (talkcontribs)

I think I got those from one of the references, I'll check. Bubba73 You talkin' to me? 18:20, 16 June 2020 (UTC)[reply]
I checked the source (Pandolfini), and they are correct. Bubba73 You talkin' to me? 23:06, 16 June 2020 (UTC)[reply]

Importance section

[edit]

@Mebden: The biggest issue I took with your edit was that Soltis' position cannot be binarily classified, as he gives reasons both for and against learning it, ultimately concluding that learning it is good overall.--Jasper Deng (talk) 06:28, 15 December 2020 (UTC)[reply]

@Jasper Deng: Thanks for this clarification. In my editing I divided Soltis's arguments (not his position) between the two subsections I'd created. However, lest there be any confusion about his position to someone reading the article quickly, I've reworded the Importance section just now to keep his two arguments in the same sentence. If you see any other issues, I'd be grateful if you didn't please revert the whole section again before we have a chance to discuss it. Mebden (talk) 17:00, 15 December 2020 (UTC)[reply]

Error in diagram E?

[edit]

Diagram E's caption says that checkmate can be forced. However, that checkmate is in the "wrong color" corner. So, presumably the caption is wrong? 79.176.188.163 (talk) 09:11, 1 September 2022 (UTC)[reply]

Method - three phases

[edit]

I want to discuss referencing the three phases from Müller & Lamprecht:

  1. Driving the opposing king to the edge of the board.
  2. Forcing the king out of the "wrong" corner to the "right" corner, if necessary.
  3. Delivering the checkmate.

The three phases are relevant only for reaching the starting position or an intermediate position for the W manoeuvre. From the example lines in the book, it is clear that the three phases only relate to the W manoeuvre, but there is no mention in the text.

For Delétang's triangle method, these phases are irrelevant, as the starting position is different.

As such, I deem the reference from Müller & Lamprecht, not a universal approach. But mentioning here sort of promotes it to the universal approach, which I deem incorrect. So one should bring the W manoeuvre onto the table.

I would also like to discuss the following points. What are your opinions on these points? I consider these facts from a chess-playing perspective but wondering about other opinions:

No systematic pattern exists for driving the king to the edge of the board and then reaching the initial or intermediate position of the W manoeuvre. So this part is often more complicated than the W manoeuvre itself.

Also, proposing the best defence for the opponent to go to the wrong corner must be clarified. Once the king is in the wrong corner, one can easily reach the starting position of the W manoeuvre, whereas it is otherwise, the opponent king not being in the wrong corner, often much more complex. Dlbbld (talk) 08:05, 22 May 2023 (UTC)[reply]

“Two moves deep”

[edit]

The phrase “two moves deep” appears twice in the current version of the article, but is not defined or clarified in any way. One is in the lead section, while the other appears in the description of Diagram E, which shows a checkmating position.

Frankly, I have no idea what the phrase means here and comes off as jargon. It either it needs an explanatory note, or should be removed. Jedzz (talk) 15:00, 20 December 2023 (UTC)[reply]

It means that White can checkmate in two moves from certain positions, but it needs more context and an example of the position. Bubba73 You talkin' to me? 03:04, 21 December 2023 (UTC)[reply]