Talk:Chess strategy and tactics
|
|
Hi,
I'm a newcomer here and I'd like to contribute to the chess theory pages! I'm a chess player and teacher and I was wondering if this project could have sense: building a complete free chess encyclopedia! Perhaps it's a too big deal, but Wikipedia is bigger than one could imagine 2 years ago! What do you think? The starting point is choosing a standard for classifying openings (and symbols for evaluation, like +-, =, etc.), middlegame and endings. Then contributors will start editing a specific page (i.e.: Marshall attack in Ruy Lopez, or identified by a code like C90 for instance) simply inserting the PGN game score (with variations). The perfect thing would be using some browser-based tool, like an applet, to show the moves. I hope this could be done, in a future (perhaps not so far). Teutoburgo
Alex, why add origin squares to some of the moves? That's not normal notation, and it might be confusing because it's only done for a few of them and not in general. --LDC
Well, I thought it would be helpful to know where the pieces came from. I reverted. AxelBoldt
Usually the entire sequence of moves is known, or at least the starting position. When the only diagram is the position after the move, I think think it makes sense to include the square of origin. True, it isn't _the_ standard, but is _one_ standard, and a good idea in some situations. --Karl Juhnke
- This issue is really a function of who the audience is. (There's a Wikipedia rule about that.) If we intend the reader to be the experienced chess player, then by all means omit redundant origin squares. But it's more likely that the reader will be very inexperienced and fumbly when it comes to reading notation. Including all origin squares is just as "normal" as omitting them. If origin squares are shown they act as a reality check to whether a previous move was made correctly, and they avoid the problem of omitting them when they really aren't redundant. Eclecticology
- I just noticed that we have a bigger problem: our diagrams don't show the file letters and row numbers, so that beginners won't be able to read the algebraic notation. AxelBoldt
Karl, your example diagrams are great! For the endgame section, where we say that kings should fight for the squares in front of passed pawns, I'd like to request a diagram of white: Ke5, Pd5 black Kd8. White moves, the only winning move is Kd6. AxelBoldt
Diagram added. I'll leave the spacing, format, caption, etc. to you. My work on this page is convincing me that it is vastly harder to discover/design a good position than it is to make and upload the image. If anyone has any more suggestions, I'll happily make more pictures.
P.S. Kd6 is not the only winning move. Ke6 wins too, e.g. 1.Ke6 Ke8 2.d6 Kd8 3.d7 Kc7 4.Ke7 etc. However, if you back everyone up a rank, i.e. white Ke4, Pd4 black Kd7, then white Kd5 is the only winning move. Should I include that diagram instead?--Karl Juhnke
You are right, the Ke4 position would be much better. Also, I was thinking for the king attack section: a standard bishop sacrifice on h7, maybe in the Colle system. AxelBoldt
- ditto about the diagrams. We may be close to spinning off endgames as a separate article. I agree completely with the point that you are trying to make with your example. An even more fundamental point to illustrate and understand yours would be something to illustrate the concept of the opposition. Eclecticology
- Actually, the diagram that Karl just added is a nice example of opposition; we just need to cover the concept in the text. Personally, I'm not a big fan of splitting the article into pieces: that way, people have to click around too much and get lost. AxelBoldt
The diagram has been modified. (Interesting how the page doesn't show as modified, because I only overwrote the image with one of the same name.) I'm still taking requests for more diagrams if anyone wants to think them up and compose the surrounding text. --Karl Juhnke
| Contents |
Values Of The Pieces
The "values of the pieces" section is incomplete in one important way. To be sure, there is no significant error in rounding-off the relative piece values (expressed in simplified ratios) as 1, 3, 3, 5, 9 for all of the pieces except the king. Unfortunately, I do not know the proper relative piece value of the king.
There is a popular misconception that the relative piece value of the king, despite its low attack value, is infinite since a game of standard chess can be won by capturing a single, royal piece (to use chess variant terminology). This compelling logical truth is deceptive and ultimately, simplistic. For, nonetheless, programs (by example) written to play chess which ascribed an extremely-high, finite value (as a useful approximation of infinity) to the king in this game achieved less-than-impressive ELO ratings. The reasons can be explained deterministically.
Where the king is given an extremely-high, finite value, the AI player or alert human player will demonstrate a tendency to make unnecessary moves of the king or other pieces to enhance king safety trivially. Essentially, such irrationally-defensive play wastes moves. This works to the advantage of the opponent in development. Typically, this leads to positional advantages, then material advantages and ultimately, the capture of the king for the opponent. One could choke on the irony.
Relative piece values are extremely important to any chess variant (including standard chess). In the case of standard chess, this fact is not so obvious due to opening book theory being so richly developed and catalogued. Strategy and tactics can become chaos, nonetheless, in the absence of accurate and complete information if playing against a competent opponent in possession of accurate and complete information.
There are surely some computer chess experts or people deep into researching this game out there who know a useful, moderate finite value for the king. Not being one myself, your input is much needed.
- There is no logical finite value for the king in chess because it cannot be exchanged for any other piece or any other combination of pieces. The whole point of the game is to checkmate the enemy king. Alex.tan 15:32, 1 Dec 2003 (UTC)
"This value will be completely dependent upon the particular program which is being coded. But consider the following proposition: The King is worth the value of all the other pieces combined.
Thus with the 1/3/3/5/9 value system, the King would be worth 39. A relatively low value but significantly higher than all the other pieces. This will effect the program's move evaluation, givng the piece an aggressive profile. And the look-ahead of the program will take care of potential mating."
L. Lynn Smith
computer programmer & chess variant expert
- I think this is an issue for the computer chess article rather than here. In pure chess terms the king really is of infinite value, but when programming a computer, I suppose there may possibly be reasons why you wouldn't want to give it an infinite (or stratospherically high) value (though I admit, I don't see what those reasons would be myself). Either way, this is really a computer programming issue, not a chess issue. --Camembert
Well, it goes against the Turing Test principle to make an arbitrary distinction between rational criteria used by computers and humans. Nonetheless, the difficulties many have with accepting the aforementioned ideas are understandable. I only want Wikipedia to be more incisive and informative than a typical introductory chess text on this important point, to describe the vast theoretical complexities of this game correctly (within realistic limits). Unfortunately, I am at a loss to prescribe exactly how- to give you an exact, authoritative value for the king to publish.
The secondary source of the 200 point valuation of the king in chess:
This information was migrated from that long-standing Wikipedia article to this one. The primary source of this value would be known to the Wikipedia editor Siroxo who wrote that article. In my opinion, he/she is probably rational. Please ask him/her about the details, any verifiable citations in the literature or effective chess programs which use this appr. value. You may discover that you need to replace what you have deleted without first investigating. --BadSanta
- Thanks for pointing out that article; I didn't know of its existence. I'll raise some points on its talk page. --Camembert
WikiProject, possible move to wikibooks
Hi, A discussion on what should go in a wikipedia article on games is underway at WikiProject Games. The current concensus (by only two users, including myself) is that strategy guides like this page should be moved over to wikibooks. So far, I don't think there are enough particpants to make such wide reaching decisions, so I'm asking for more voices to be heard. The discussion is at the wikiproject's talk page.
Please note that I do not advocate deleting this type of article completely, but feel a chess strategy guide (and strategy guides for all games) fits better with the mission of wikibooks, so this page would be moved over there, and be linked to from the wikipedia article about chess. Premeptive thanks for your participation. Gentgeen 08:03, 14 Jan 2004 (UTC)
- Indeed, I think this article should be merged into Wikibooks:Chess. --ZeroOne 16:12, 26 Mar 2005 (UTC)
Zugzwang
Should the Zugzwang page be deleted and entered in here under tactics? I just noticed that Zugzwang had its own page but no mention on this one. kmac 14:58, Mar 6, 2004 (UTC)
- It certainly ought to have a mention on this page (thanks for putting something in about it), but I don't see the harm in it having its own article as well for a more detailed look at the subject. It's a pretty important concept, and quite a lot could be said about it - the article as it stands is just a start, really. --Camembert
Piece values (again)
The article sayeth:
- There is another system for evaluating chess pieces, as follows:
- pawn = 3,1st captured Bishop = 11,2nd captured Bishop = 10,Knight = 10,Rook = 15,Queen = 29.
- Using this system, in a B&N trade for a R&pawn (for example, on square f7, after castling), the B&N are worth more (20 or 21 vs. 18 for R&pawn), hinting that this kind of a trade is not equal, and therefore unadvisable. The traditional system indicates that these pairs are equal.
This system seems needlessly complicated to me, and I'm not sure it has very wide currency. Can somebody give an example of a writer who advocates it? If so, lets cite them in the article. If not, I tend to think it should be removed. The fact is that the value of pieces changes with the position, and no simple points system can take account of that. These systems are just a rough guide, and the 9/5/3/3/1 system, which is almost universal, does the job about as well as anything else. --Camembert
- I also think this secondary evaluation of piece values should be removed. Although its author considers it refined and more accurate, such systems are argumentative by nature. Furthermore, the writing style is probably confusing to the general reader. Talk where rules and exceptions to rules and exceptions to the exceptions are mentioned (even if they have a basis in fact) should be avoided if at all possible in general encyclopediac articles. --BadSanta