Jump to content

Talk:First-move advantage in chess/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3

Moved from Draw

This article was a section of Draw (chess) and moved here. Bubba73 (talk), 14:29, 18 March 2008 (UTC)

Watson Review of "The System"?

International Master John L. Watson for some years now has been reviewing virtually every chess book of any significance for The Week in Chess and others. From what I have seen elsewhere, he wrote a scathing review of Hans Berliner's book "The System." Somehow I can't find it on the Internet. Can anyone else? Krakatoa (talk) 19:57, 10 April 2008 (UTC)

Could you be thinking of this one by Silman: [1] Bubba73 (talk), 21:04, 10 April 2008 (UTC)
Aren't we in danger here of giving undue weight to Berliner's ludicrous claims? He seems to be the only writer in the history of chess literature to say that 1.d4 leads to a decisive advantage, so why mention it?Pawnkingthree (talk) 08:29, 11 April 2008 (UTC)
I don't think so. The article says that it has been sharply critized, and gives two references. Bubba73 (talk), 14:38, 11 April 2008 (UTC)
Bubba73, I already cite Silman's review (also Randy Bauer's, and Taylor Kingston's). As for undue weight, the article is "First move advantage in chess," so I think it's fair to talk about the two prominent people (Adams and Berliner) who have claimed that White has a decisive or near-decisive advantage. Berliner is/was an extremely strong player (IM over-the-board, and the most successful correspondence player in history, most notably by winning the 1965-68 correspondence world championship by the unheard-of score of 14-2, undefeated, three points ahead of second). I think it's significant when a player of that strength makes such a claim -- although unfortunately I think he's tarnished his image greatly by doing so. And as Bubba73 says, the article says that Berliner's view (like the similar view of his hero Adams) has been sharply criticized, and gives three (not two) references. If you can find Watson's review, to make it four references, great. The article also says that the prevailing view, from at least Steinitz on, has been that White's advantage is not decisive. Krakatoa (talk) 16:00, 11 April 2008 (UTC)
I think everything in there is very well balanced. A sentence or two about why it was criticized would help. Bubba73 (talk), 18:18, 11 April 2008 (UTC)
Thanks for your additions in this area, Krakatoa, that's addressed all my concerns. Superb article.Pawnkingthree (talk) 11:45, 12 April 2008 (UTC)
Do you think Watson's criticism could be in his book Mastering the Chess Openings? It might be in the intro articles in volume 1 or in vol 2 (the 1. d4 openings). They only have an index of players and openings, so I can't look up "Berliner" in the index. Bubba73 (talk), 18:23, 11 April 2008 (UTC)
I am embarrassed to say that I don't think I have "Mastering the Chess Openings, Volume 2." I thumbed through Volume 1 and don't see anything about Berliner, nor is his book mentioned in the bibliography at the front (as I assume it would be if Watson said something about it). Could you or someone else who has Volume 2 please thumb through it and see if he talks about Berliner in there? If he does, that would be a decent thing to cite even if he doesn't include the whole review. Krakatoa (talk) 06:24, 12 April 2008 (UTC)
I have vol 2, I'll try to see. It might be there, but probably not. Just a shot. Bubba73 (talk), 14:07, 12 April 2008 (UTC)
There is a Watson review from 1999. Here[2] you can see Silman refer to someone else disagreeing with Watson's criticsm: "In the December 1999 issue of The Chess Journalist, Mr. Savage took exception to John Watson’s highly negative review of THE SYSTEM." Unfortunately although I looked for Watson's review I couldn't find it. 1999 is just early enough so that it's possible that it isn't online, although some Watson reviews dating to 1998 can be found at http://www.chesscenter.com/twic/twic.html. The Silman article mentions that Tim Harding, another good chess writer, defends Berliner's work. Quale (talk) 20:52, 11 April 2008 (UTC)
Right, I saw the same Silman piece you did, mentioning Watson's criticism and correspondence players' attempts to defend Berliner (rather amazing in my view, and I am a senior master at correspondence chess -- I guess I shouldn't be surprised that saying "the greatest correspondence chess player of all time is a snake-oil salesman" gets their hackles up). I also looked at Watson's TWIC reviews, and couldn't find his review there or anyplace else. Maybe as you say it's not online because it's 10 years old. Too bad; I'd love to see it. I daresay Watson "tears Berliner a new one," as they say. Krakatoa (talk) 06:15, 12 April 2008 (UTC)

Sections

Thanks Krakatoa, your complete rewrite of this article is outstanding. It exceeds the sum of its parts, and is now better than any of the individual references used to compile it. If possible we should try to divide it up into two or three sections so that it isn't just a large chunk of text. I can't think of any images or diagrams that could be used to make it more visually appealing, so I think a few sections could really help. Any ideas? Quale (talk) 02:37, 11 April 2008 (UTC)

Thanks, Quale. Well, obviously we could have a Weaver Adams section and a Hans Berliner section, though those wouldn't encompass everything. I guess we could have something like "The classical view" section, then Adams, then Berliner, then "The modern view" or something like that. Krakatoa (talk) 16:03, 11 April 2008 (UTC)
Sure. Another possibility might be to use three groupings arranged by color rather than by the individual person: "Arguments that White wins" (or "White has a decisive advantage"), "Chess is a draw" (or "Dynamic balance"), "Black has some advantages" (or "Disadvantages of the first move"), and then Adams and Berliner would go in the first section, Kasparov et. al. would go in the second, and the (small) third section could include Suba and possibly Breyer ("White's game is in its last throes!", although this is about a specific white opening move not white's chances in general). I'm not really happy with those section titles, but breaking it up by the three main competing views might be logical. My answer to the question asked above whether the article gives undue weight to Berliner is no, as Berliner's views have attracted a fair amount of attention and commentary (generally critical), and that's noteworthy. Putting his name in a section title might be undue weight. Putting Adams and Weaver together in a section would emphasize the similarity between their views. Quale (talk) 16:57, 11 April 2008 (UTC)

Reversed openings

A brief discussion of why reversed openings (KIA, English Sicilian, etc.) are not always favorable for white might make a good addition to the last part of the article that talks about why white might not have a full move advantage. I guess it would expand on the final remarks by Watson a little. Quale (talk) 02:40, 12 April 2008 (UTC) Done

Article rating?

Anyone think the article is better than start class now? Krakatoa (talk) 08:57, 12 April 2008 (UTC)

I think it's very good, and your sections and section titles are much better than the ones I was thinking about. I'm bumping the rating to B. I would say it's an A-class article, but many WP:CHESS members think that the project A-rating requires a (semi-)formal review. It's certainly within the purview of a project to determine the rating process it wants to use for its articles, and this review requirement doesn't seem to be unreasonable even if I'm not sure I fully agree with it. Quale (talk) 09:12, 12 April 2008 (UTC)
Thanks. Inspired by your comments, I've nominated the article for A-class. In my nomination, I quoted your first three sentences above. Krakatoa (talk) 16:09, 12 April 2008 (UTC)
I'm going to check through it. Personally I find the topic fascinating. SunCreator (talk) 16:11, 12 April 2008 (UTC)

Moved my post to Wikipedia:WikiProject_Chess/Review#First_move_advantage_in_chess SunCreator (talk) 15:39, 13 April 2008 (UTC)

Topic title

Isn't the title POV. Some might claim it's not an advantage, or at least not an advantage for them. SunCreator (talk) 16:32, 12 April 2008 (UTC)

There's a typo at the end of your last sentence, so I don't know exactly what you meant. I didn't create the title, but it seems reasonable. Although the article references a few contrary views at the end, the overwhelming consensus is that White starts the game with at least some advantage, and White has consistently scored 53% or better for the last 150+ years at least. What else would one call the article -- "Does White or Black have the advantage in chess?" The current name seems a little cumbersome, but any alternative I can think of seems worse. Krakatoa (talk) 17:12, 12 April 2008 (UTC)
Corrected my thinking mistake. I'm not talking about the article just the topic name with the use of the word 'advantage'. SunCreator (talk) 17:18, 12 April 2008 (UTC)
Until March 18 there was a section in Draw (chess) called "first move advantage". I didn't think it was appropriate over there, so I moved it to this new article and added "in chess" so people would know what it was about. Of course, the article has changed radically since then. I can't think of a better name, but I'm not against renaming it if a better name is found. Bubba73 (talk), 21:47, 12 April 2008 (UTC)
I thought that you were right to split this out of draw (chess), and Krakatoa's complete rewrite has certainly proven that to have been a very wise decision. At first I also wondered about the page name, but actually it's grown on me and I rather like it now. Quale (talk) 06:44, 13 April 2008 (UTC)
Yes. It's not easy to find another name. Been looking for something similar in other sports. In tennis you have the Serve (tennis) but as they alternate server per game an advantage is soon gone, this is typical in most sports but not all. More comparable is Cricket where you has the Toss (cricket) and motor sports with Pole position. SunCreator (talk) 11:59, 13 April 2008 (UTC)
Right. As long as the page name isn't POV, the main rule for the name is to use what it is usually called in an English language encyclopedia, or what would a reader expect it to be called. This will help make the page findable by an interested reader. Unfortunately the current title is an unlikely search term, but I can't think of anything better. To make the article discoverable we need to link it in some other appropriate articles (chess, chess opening, and draw (chess) would be good places if we can find appropriate spots) and put it in any relevant chess categories. Krakatoa has made this a fine article, so it would be a shame if no one can find it. Quale (talk) 06:33, 14 April 2008 (UTC)

Someone said

Someone said that white can afford two mistakes but black can afford only one, or something like that. That would be good to include if the details are known. Bubba73 (talk), 21:47, 12 April 2008 (UTC)

I am almost sure Evans said that. I looked for it, but couldn't find it. I'll keep looking. Krakatoa (talk) 00:22, 13 April 2008 (UTC)
Anyone have a copy of the RHM Press book How to Open a Chess Game (1974) by Evans, Gligoric, Hort, Keres, Larsen, Petrosian, Portisch, and Burt Hochberg? I think the quote might be in Evans' chapter in that book, and maybe some of the other contributors said things relevant to this article. I should have a copy of the book someplace, but can't find the damned thing. Krakatoa (talk) 04:44, 15 April 2008 (UTC)
I found a statement by Suetin that I think is along the lines of what we were looking for, and put it in the text -- White can make a small mistake and generally only lose the initiative, whereas a similar mistake by Black may have more serious consequences. Krakatoa (talk) 08:42, 15 April 2008 (UTC)

Chess Speak

Some of these words I'm not sure are in any dictionary.

'hypertheoretical' and 'drawish'

While 'hypermodern' seems to have a different meaning in chess then it does in the dictionary where it means up-to-the-minute.

I think "hypertheoretical" is used in a quote from Rowson or someone; I wouldn't use it in Wikipedia other than in a quote. "Drawish" means "tending toward a draw" -- a commonly used chess word. You're quite right about "hypermodern": in chess, hypermodernism is now about 90 years old. Krakatoa (talk) 01:47, 16 April 2008 (UTC)
All  Done Hypertheoretical and drawish now with wikified link. Hypermodern is actually not used in the text but is a name of a chess publisher. SunCreator (talk) 14:14, 16 April 2008 (UTC)
What reference did you use for the definition of hypertheoretical? It isn't a commonly used term and in my view it does not belong in List of chess terms. Quale (talk) 14:38, 16 April 2008 (UTC)
I don't have one. I did quite a lot of checking around the web but didn't find any place with a definition. I do think there has to be some definition however as this article uses it(the only one on wikipedia to use the word 'hypertheoretical' by the way). SunCreator (talk) 14:44, 16 April 2008 (UTC)
That is original research and not permitted. The fact that this article is the only one on wikipedia to use the term "hypertheoretical" actually makes it a lot less important to include a definition. Quale (talk) 15:09, 16 April 2008 (UTC)
I do appreciate that it shouldn't be there as it is original research however the word in this article really wants to be a wikified link because the word is so rare. I will see if I can find a referenced definition of the word. SunCreator (talk) 15:17, 16 April 2008 (UTC)
I think this concern is overstated, but that is just my personal opinion. When Rowson used "hypertheoretical", I don't think it was as particular chess jargon more than it was just a way to try to amplify the well-understood term "theoretical". He could just as easily written "megatheoretical", "ubertheoretical", "ultratheoretical", etc., although hypertheoretical sounds a bit better and probably has the more precise connotation he was aiming for. In fact according to the spell checker in Firefox, none of these are words including "hypertheoretical" which suggest that "hyper-theoretical" might be a better spelling, although naturally we don't change spelling in quotes. I think that all that needs be done (and probably the best that can be done) is to get good definitions for "theory" and "theoretical" in the chessic sense and link to that. Throwing "hyper" on the front doesn't change the essential meaning—it's basically just attempting a nuance. Quale (talk) 16:04, 16 April 2008 (UTC)
I forgot to add that putting terms in List of chess terms that are used only or are extremely rare in the chess literature is a bad idea, unless the term is for some reason extremely important. "Hypertheoretical" doesn't qualify in my opinion. If it requires special explanation (and I'm not convinced it does), that explanation belongs in the only place that the term is used (which is this article). Quale (talk) 16:08, 16 April 2008 (UTC)
Your right to remove it from List of chess terms. To my mind that still leaves work to be done to clear it up in this article. I don't think Rowson meant "hyper" and "theoretical" as the meaning of those words together do not seem to fit into his quoted statement. The word hypertheoretical does exist in other literature, see, Find sources: Google (books · news · scholar · free images · WP refs· FENS · JSTOR · TWL. SunCreator (talk) 16:25, 16 April 2008 (UTC)

I think that what Rowson meant by "hypertheoretical" is as follows: An ordinary chess theoretician would probably say of the opening postion, based on his/her experience and analysis, that in his/her opinion White has a small advantage, but with best play the game probably should end in a draw. The "hypertheoretical" view, which could only be stated authoritatively by Caissa or some other omniscient creature, very likely is that "this position is a draw." No mere mortal could authoritatively make such a pronouncement, since it is impossible to analyze the enormous number of possible continuations and definitively conclude that the game should end in a draw. It's like the difference between how chess punctuation (=, plus over equal, etc.) in used in books on the opening and the ending. In opening books, situations where White's winning chances are a little better than Black's (maybe translating to an overall percentage of 55-60% or thereabouts) are rated plus over equals. Endgame books, on the other hand, will affix an equal sign to, say, a rook and bishop versus rook ending where it's known that the side with the extra bishop can't force a win, even though in actual practice against an opponent who's not an expert in this ending, the superior side's winning chances are probably higher (particularly given that there are essentially no losing chances) than in the plus-over-equals opening position. Yet the hypertheoretician calls "equal" an ending where the superior side will score better, in actual practice, than in the "plus over equals" opening position. Krakatoa (talk) 01:54, 17 April 2008 (UTC)

The difference between the "theoretical" and "hypertheoretical" levels is esoteric, but failing to understand it sometimes has consequences. Rowson writes that Andras Adorjan's Black is OK! series is important, but complains (again, if I'm understanding "hypertheoretical" correctly) that Adorjan uses a sort of linguistic sleight of hand in conflating the two levels. Rowson says Adorjan misleadingly says things like, "Chess is a draw, so it's wrong to say that White stands better in the initial position." Rowson notes that the statement "Chess is a draw" is only true at the hypertheoretical level, whereas in actual practice no one knows how to force a draw as Black, and White scores better than Black from the initial position; thus, at the "theoretical" level, and in actual practice, White is better. Another example: some players in slightly inferior endgames (say, R+B+P v. R+minor piece) have unnecessarily sacrificed to reach the inferior side of R+B v. R, having heard that "Rook and Bishop versus Rook is a draw." See, e.g., Mednis-Weinberger, cited in Edmar Mednis, Practical Endgame Lessons, pp. 195-97. But that statement too is only true in the hypertheoretical sense; in actual practice, if the inferior side doesn't know the ending very well, he/she is likely to lose, as Weinberger did in short order.

As to what to do with the text's quote of Rowson's use of "hypertheoretical," probably the thing to do is write the article on Chess theory, as I keep threatening to do, then (if we can find a good explanation by Rowson or someone else) mention that some chess writers distinguish between "theory" and "hypertheory," explain that difference, then wiki-link "hypertheoretical" in the present article to the Chess theory article or the relevant section thereof. Krakatoa (talk) 09:30, 17 April 2008 (UTC)

OK, partly there. Created the article on Chess theory; it needs major work, but it's a start. I wiki-linked "theorist" in the present article to that article, which also defines "theorist" and "theoretician." Nothing about "hypertheoretical" in there yet; I have to look to see whether Rowson actually defines that term, or whether my thoughts above are simply my own surmise as to what Rowson is talking about. Krakatoa (talk) 11:38, 17 April 2008 (UTC)

Now the reference at the end of the "hypertheoretical" quote cites Rowson explaining what he means by "hypertheory." It's basically what I said above (which I'm sure I dimly remembered from Rowson, rather than making it up on my own). Krakatoa (talk) 22:29, 21 April 2008 (UTC)
I think the explanation looks good, as does the rest of the article. Every time I think the page is nearly as good as it could be, you find a way to make it even better. Quale (talk) 01:54, 22 April 2008 (UTC)
Thanks. "My cheeks are crimson," as Larry Evans once said (of himself, not me). :-) By the way, you haven't actually voted on the proposal to make the article A-class -- so if you have an opinion, you might want to do so. Krakatoa (talk) 04:43, 22 April 2008 (UTC)
Sorry, although I'm a fairly experienced editor (about 3 years here), I must confess that I am completely at a loss to understand how to contribute at Wikipedia:WikiProject Chess/Review. The collapsible divs leave me baffled. I also might not have a rigid enough approach to the review process to fit in. Although I have a basic check list of requirements for what I consider to be an A-class article (some do's and a lot of don'ts), I don't have it reduced to as mechanical a process as seems desired. I also may not be critical enough, because my only complaints fall to the trivial. (Such as a preference that in the references, "Id." not be used in favor of a brief Harvard-style reference. As an example, cites 69 and 70 are in separate paragraphs, but since 70 is an "Id.", if anyone were to move that para or insert extra material, the ref would be messed up. "Id." is great for papers completely under your control, but seems to me to be too fragile a device for a document with the potential of drastic revision from many other, possible less careful or experienced editors.) Quale (talk) 02:42, 23 April 2008 (UTC)
Have undived the review. Was having similar issues reading/editing. The point about the "Id." is very insightful and something that I doubt would of occurred to me. SunCreator (talk) 03:03, 23 April 2008 (UTC)
Yeah, I was having trouble with the collapsible divisions too. You're also right about the "id." thing. In fact, the "id." in 70 was wrong, since 69 was a Yermolinsky cite, and 70 a Watson cite. I had inserted a bunch of stuff between the prior Watson cite and 70, and hadn't noticed the need to change the "id." Now I've killed all the "id."s except those within a single reference number. Krakatoa (talk) 05:28, 23 April 2008 (UTC)
Well, sorry about the collapsible divisions. As someone had complained that the review was not structured enough, I had figured this as a way to clearly separate each review. Of course I did not thought it would confuse editors so much. SyG (talk) 19:41, 23 April 2008 (UTC)
Nothing ventured and all that. There's probably no perfect solution. Krakatoa (talk) 20:31, 23 April 2008 (UTC)
No need to apologize for the divs. Just because I was confused doesn't mean everyone was, and anyway trying things out is good. You and SunCreator have done some nice work to improve the functionality and appearance of WP:CHESS and related project pages, and I think that's good. Quale (talk) 01:07, 24 April 2008 (UTC)

Anyone have "New in Chess" books after 2000?

I stopped collecting "New in Chess" books after 2000. The 2000 Yearbook is the most recent I have, so that's what I quoted in the "Winning percentages" section. Does anyone have a more recent volume we can cite? I doubt the percentages have changed significantly, but it would be nice to cite a 2007 or 2008 book. Alternatively, if we could cite someone writing about the statistics from the ChessBase Mega 2008 database or something, that would be a good substitute, I think. Krakatoa (talk) 01:47, 16 April 2008 (UTC)

I don't have what you need, but I want to commend you on avoiding orginal research on this. It would be easy to compile the numbers ourselves, but that would be borderline OR already, and we would not be able to make any qualitative analysis of them without definitely going afoul of the prohibition on OR. A lot of editors don't seem to understand this, but the Statistics (oops, now "Winning percentages") section you wrote is a great example of the right way to do it. Quale (talk) 05:26, 16 April 2008 (UTC)
Personally I am not greatly enamored of this rule when it comes to this sort of thing, but I do understand it and don't want to condemn this well-received article to the purgatory of perpetual Class-B status by running afoul of this hallowed rule. Krakatoa (talk) 07:55, 16 April 2008 (UTC)
I just ordered the New in Chess Yearbook 85 (2007), so I'll see what its statistics look like. Krakatoa (talk) 05:28, 29 April 2008 (UTC)

Can it be solved

A section on whether it's solvable 'given best play by both sides, the game should conclude in a win for White or a draw.' could be added, there is some material in the topics Computer_chess,Minimax_algorithm and Game_tree. Obviously the answer is that it's currently impractical, but it might be a useful addition to the article if suitable references can be found. SunCreator (talk) 01:51, 16 April 2008 (UTC)

This is a very good idea. All we need is a short mention with a link to one of those articles. Quale (talk) 05:22, 16 April 2008 (UTC)
Also Shannon number. SunCreator (talk) 19:44, 16 April 2008 (UTC)
I now explain what Rowson means by "hypertheoretical," and cite (part of) what he says about chess theoretically being solvable. He says that in some number of decades or centuries, we'll have 32-piece tablebases. Krakatoa (talk) 22:26, 21 April 2008 (UTC)
I added a section on this to the article. See what you think. I'm rather out of my depth here, so any additions/editing would be welcome. Krakatoa (talk) 05:37, 27 April 2008 (UTC)
I think the section is great, but it lacks an introductory paragraph making the link with the subject of the article. It should be made clear for the reader that, in absolute/theoretical terms, White having an advantage means the solved game is a white win. SyG (talk) 14:20, 2 May 2008 (UTC)
I (1) added a sentence to the end of the lead paragraph about "solving" chess; (2) added a sentence to the beginning of the "Solving Chess" section tying it in with the rest of the article; and (3) moved the "Solving Chess" section to the end. It seems most logical there -- first we talk about all the different theoretical perspectives (White has some advantage, but it's drawn with best play; White wins; White has no advantage, and it's drawn with best play, etc.), then we talk about the possibility that computers will eventually resolve the debate once and for all. See what you think. Krakatoa (talk) 02:34, 3 May 2008 (UTC)

Philidor

I have seen vague references to Philidor thinking that the game was a forced win for White, but can't find anything definitive. Most mentions of him in the chess literature are just "Pawns are the soul of chess yada yada yada." Harding Simpole reprinted an English version of his book a few years ago, so I'm going to get hold of it and see what he said. Krakatoa (talk) 10:52, 16 April 2008 (UTC)

I got and read the book, and saw nothing addressing this. He did say that the King's Gambit and other gambits should be drawn with best play by both sides!? Krakatoa (talk) 05:56, 19 April 2008 (UTC)

On expanding the statistics section.

This is a superb article! Here are some suggestion on expanding the statistics section. I would be very interested in:

  1. Are the percentage dependent on the type of game, eg. any statistics available on correspondence chess, computer chess, blitz vs standard? I expect differences can be found.
  2. Is there an evolution over time? A graph would be nice, but will be hard to find.
  3. Is it dependent on the rating? eg more decided games between low rated players?
  4. Would be nice to give the statistics on the chessbase megabase (the largest commercial database as far as I know)
  5. How about the % for opponents with different rating. This chessbase article [3] has the graph. It also gives the number 0.541767 + 0.001164 (i think for games 1994-2001)
  6. How about the influence of the opening. Eg, is there also such an advantage in comp games played without opening book. Or is there such an advantage in chess960? The article has a little bit on it (eg sicilian does better), but could be expaned. Statistics on chess960 can be found here [4] (51.865%, gm games, but only around 3000 games).

I think realistically, we can hope to find information on nr 1, 4 and 5; the rest would be a bonus. The challenge will be to stay away from OR. I am sorry I was only able to give one new reference, but I will keep on looking. Voorlandt (talk) 20:35, 18 April 2008 (UTC)

Answering nr 3, there seems to be a correlation; I just checked the statistics with ChessDB and the huge database it comes with

All games        3663979 1416648 1109964 1137334    53.8%
Both rated 2700+    3711    1010    2114     587    55.6%
Both rated 2600+   28509    8111   15539    4858    55.7%
Both rated 2500+  134658   37669   73447   23541    55.2%
Both rated 2400+  403513  121111  202842   79555    55.1%
Both rated 2300+  753392  242713  341650  169020    54.8%
Both rated 2200+ 1162932  397798  475002  290113    54.6%
Both rated 2100+ 1464819  516963  561844  385993    54.4%
Both rated 2000+ 1684823  605717  620120  458960    54.3%
Both rated 1900+ 1769646  639415  642916  487285    54.2%
Both rated 1800+ 1839420  667586  660985  510819    54.2%
Both rated 1700+ 1891033  688631  673758  528614    54.2%
Both rated 1600+ 1934476  706414  684187  543845    54.2%
Both rated 1500+ 1964069  718682  690842  554515    54.1%
Both rated 1400+ 1985761  727875  695244  562612    54.1%
Both rated 1300+ 2001871  734855  698349  568637    54.1%
Both rated 1200+ 2014606  740395  700634  573547    54.1%
Both rated 1100+ 2023285  744164  702116  576975    54.1%
Both rated 1000+ 2029943  747145  703053  579715    54.1%

If you want this quotable, then I can download the latest version, and run this again. Since this is 100% verifiable, I think it doesn't qualify as OR. Voorlandt (talk) 20:49, 18 April 2008 (UTC)

Can you explain the columns above. There seems some minor error. 3711=(1010+2114+587) which is correct but 28509 does not equal (8111+15539+4858) but 28508 etc. SunCreator (talk) 22:48, 18 April 2008 (UTC)

Rating	Total#Games  W-Wins   Drawn  B-Wins  T-Calc  White%
2700+		  3711	  1010	  2114	   587	  3711	55.70%
2600-2699	 24798	  7101	 13425	  4271	 24797	55.71%
2500-2599	106149	 29558	 57908	 18683	106149	55.12%
2400-2499	268855	 83442	129395	 56014	268851	55.10%
2300-2399	349879	121602	138808	 89465	349875	54.59%
2200-2299	409540	155085	133352	121093	409530	54.15%
2100-2199	301887	119165	 86842	 95880	301887	53.86%
2000-2099	220004	 88754	 58276	 72967	219997	53.59%
1900-1999	 84823	 33698	 22796	 28325	 84819	53.17%
1800-1899	 69774	 28171	 18069	 23534	 69774	53.32%
1700-1799	 51613	 21045	 12773	 17795	 51613	53.15%
1600-1699	 43443	 17783	 10429	 15231	 43443	52.94%
1500-1599	 29593	 12268	  6655	 10670	 29593	52.70%
1400-1499	 21692	  9193	  4402	  8097	 21692	52.53%
1300-1399	 16110	  6980	  3105	  6025	 16110	52.96%
1200-1299	 12735	  5540	  2285	  4910	 12735	52.47%
1100-1199	  8679	  3769	  1482	  3428	  8679	51.96%
1000-1099	  6658	  2981	   937	  2740	  6658	51.81%

My understanding is that if you're crunching the numbers in a database yourself that is OR, even if those numbers could be verified by anyone else. I'm sure Quale has an opinion on this. That chart is very interesting, and makes a lot of sense: as the ratings go down, draws go down (everyone knows that -- beginners don't draw games) and White's winning percentage goes down (the games become more random, and thus victory is determined mostly by who makes the last blunder). For purposes of the article, I think the numbers at the highest level are most significant -- we want to know the "truth" of how big White's advantage is, and that's determined by the games of the strongest players, since random factors are least significant in their games. My inclination is to think that the "Winning percentages" section of this article shouldn't get much bigger than it is, or it will overwhelm the rest of the article. Krakatoa (talk) 06:09, 19 April 2008 (UTC)
Yeah you are probably right, about the OR. Without a source this thing cannot be quoted. Voorlandt (talk) 06:56, 19 April 2008 (UTC)
I agree with Krakatoa's perception of OR. Although you are still allowed to do some extremely basic computations (say, additions), anything a bit more complex easily becomes OR. Probably the best criteria would be something like "is the pure Average Joe able to understand and reproduce the operation ?". If the answer is yes, it is just "common sense computations", if not it is OR. In the present case I would classify it as OR because the average Joe is not able to query a database. SyG (talk) 08:22, 19 April 2008 (UTC)
Any thoughts or sources on the other questions? Number 5 has a nice reference, perhaps a line or two can be added in the article? Voorlandt (talk) 09:04, 19 April 2008 (UTC)
I added a line about Number 5 and changed the range of recent nos. for W advantage to 54-56% (rather than 55-56%) in light of it. Krakatoa (talk) 09:11, 19 April 2008 (UTC)
Was no 'crunching the numbers in a database' in the post of number above, it's simple maths. However it may exceed OR especially as Voorlandt's numbers don't entirely seems to add up. SunCreator (talk) 13:07, 19 April 2008 (UTC)
Krakatoa and SyG correctly characterize my view on original research in this context. I am actually even a little uneasy about the work we've done on a different page scouring the FIDE ratings lists to try to get title years, as that seems to me to be borderline. (I excuse it because we aren't advancing any novel theses or making any qualitative statements—we just need the years.) I think it would be great if we could source the well-known observation that Krakatoa makes—color makes less difference for weak players than strong ones, and White's advantage grows as the players' strength increases. I also wonder if we could make note of some players who are well known for being especially effective with only one color. Kramnik's strong preference for White is a recent example. I also agree with Krakatoa that the stats section shouldn't grow too large or it will unbalance the article, which I think is one of the best we have. Quale (talk) 22:18, 19 April 2008 (UTC)

Year?

"White scored 53% in both 1851-78 (W46 D14 L40) and 1881-194 " What is the last year supposed to be? 1914? Bubba73 (talk), 05:12, 21 April 2008 (UTC)

Woops! Yes, fixed it. Good catch. Krakatoa (talk) 05:17, 21 April 2008 (UTC)

Picture of Steinitz

Could this perhaps be moved further down the article? I appreciate it's included because he was one of the first to discuss the topic but the way it's positioned in the lead section, and where an image would be placed for a biography, makes it look as though the whole article's going to be about him!Pawnkingthree (talk) 08:50, 21 April 2008 (UTC)

I think some image should be top right. My suggestion already was to add a caption below, if the caption refers to the subject it would refocus the image back onto topic rather then onto Steinitz himself. That seem a solution to me. SunCreator (talk) 10:58, 21 April 2008 (UTC)
Added caption and reduced image size. Would be nice to have date with caption, Wilhelm Steinitz, 18xx. SunCreator (talk)
I added "(1889)" Krakatoa (talk) 22:22, 21 April 2008 (UTC)

Balance the structure

Due to the numerous and great additions that were done recently, now I feel the article is unbalanced, with the section "Modern perspectives" being fives times more important than any other section. Would you have suggestions for a more harmonious structure ? Maybe the subsections can stand on as sections on their own ? SyG (talk) 15:48, 26 April 2008 (UTC)

Pictures

I love those pictures you added, SunCreator! Thanks! Krakatoa (talk) 18:57, 29 April 2008 (UTC)

Thank you. As you add so much text - after a time some image is helpful to lighten it up. Added some images to Chess theory also. SunCreator (talk) 20:10, 29 April 2008 (UTC)

reflist 2

Just to explain that the use of {{reflist|2}} causes issues for some readers that see an effect. See Template_talk:Reflist#Multiple_columns_deemed_bad.. In addition it does nothing for IE readers who still are presented with one column. SunCreator (talk) 12:56, 2 May 2008 (UTC)

I would like to know more on that, as this point may impact other articles as well. Is there a consensus for/against multiple columns ? SyG (talk) 13:09, 2 May 2008 (UTC)
I don't know if it's been discussed elsewhere but my reading of Template_talk:Reflist suggest consensus is very much against the use of multi columns until it's improved. Problems of
small text is harder to read
on smaller screens, typically flows into each other
output of long items such as URLs overrun
browser compatibility issues
increase in page loading time
There is also a related issue of font size effecting IE only that is in progress here. SunCreator (talk) 13:44, 2 May 2008 (UTC)
Reflist|2 is used in many featured articles (see todays article Satyajit Ray) and everyone seems happy enough about it. Talk pages of templates are mostly used for reporting problems, and hence you don't see anyone defending it there. IMO consensus is quite something else. Reflist|2 saves space (it is getting pretty lengthy; do we want 1/2 text and 1/2 references?) and gives the article the professional look of published papers. Btw, the issues are mostly due to outdated browsers. Please don't revert honest edits without discussion. Voorlandt (talk) 13:57, 2 May 2008 (UTC)
The talk pages of templates may well be unduly biased. Perhaps an idea would be to raise the question at the village pump. BTW few IE users will understand the situation because for them it always shows one column no matter the setting given in reflist. SunCreator (talk) 14:43, 2 May 2008 (UTC)
Voorlandt appears to be correct that the consensus is in favor of reflist 2 rather than reflist. I looked at the Featured Articles queue for May. Of the 14 articles slated so far (for May 1-14), nine use reflist 2, three use reflist 3, and only two use reflist. This article originally used reflist 2; it was changed, the reason given being that reflist 2 was disfavored and that the article would be unlikely to become a Featured Article while using it. Since that evidently is not the case, and the consensus appears to be in favor of reflist 2, I've changed it back. Krakatoa (talk) 17:07, 2 May 2008 (UTC)
Okay. I haven't seen a reflist formatting error for about three weeks now. I'm not sure if the reason is that somehow reflist(or part it uses) has been fixed, or an update to the latest version of Firefox has fixed it. It's certainly the case that the majority of the daily featured articles now use it so I'm happy to go along with that. SunCreator (talk) 17:47, 2 May 2008 (UTC)

A-class review successfully passed

This article went through a quality review and was awarded the A-class by the community of the WikiProject Chess. You can find the details of the review at Wikipedia:WikiProject Chess/Review or in the table below: SyG (talk) 08:13, 3 May 2008 (UTC)

This review is done in the scope of the WikiProject Chess and is transcluded from Wikipedia:WikiProject Chess/Review/First-move advantage in chess. The edit link for this section can be used to add comments to the review.

Exponents in the Shannon quote

The exponents have merged into the numbers in the quote. Near the end two references back-to-back are identical; one could be removed or a named ref could be used. I've also added a ToC to this page since there's a NoToC in the A-class review. Gimmetrow 10:06, 3 May 2008 (UTC)

Further work needed

Nice article, I learned a few things and it also helped me understand some of my own experiences as a player - so irrespective of B / A / G A / FA grading rules (which bore me), it's a useful article!

Some things I think need work:

  • Too much detail about the stats. The main points to be made from the stats are: White's apparent advantage has been persistent from about 1850 to the present; there's some evidence that it's increasing; there's some evidence that it's more significant at the higher levels of play. If necessary some of the stats can be summarised in footnotes.
I have to disagree here. I think the stats section has just the right length. It certainly not dominating, and for the encyclopedic value of this article, it is absolutely crucial to have this section in full length (as opposed to being tucked away in footnotes). Voorlandt (talk) 13:33, 3 May 2008 (UTC)
  • Too many quotes from world-class players in support of "best play leads to a draw". If necessary some can be summarised in footnotes. The opinions of world champions and contenders are not good evidence, because by definition they are playing much weaker opponents most of the time (in one article that I saw, Kasparov used the phrase "the average weak grandmaster"). Fischer and Kasparov are the least suitable sources, as both were notorious for the way in which their unconcealed will to win intimidated some world-class opponents.
I found a source you might consider more authoritative than Kasparov and Fischer. Bruce Pandolfini, asked the correct outcome of a perfectly played game, wrote, "I have it on good authority (this guy I used to know from Washington Square Park), that the position is drawn . . . ." [5] Krakatoa (talk) 20:47, 7 May 2008 (UTC)
  • I'm in at least 2 minds about the section on hyper-sharp openings as a drawing device for Black:
    • I think the details should in principle be moved to "appendices", either as footnotes or as an "examples" section at the bottom.
    • There are "issues behind the issues" that deserve more attention, if suitable citations can be found. Players who use these as drawing variations clearly believe in the first move advantage, despite the opinions of world champions from Steinitz to Kasparov. And there's a connection to the impact of computers on chess - how else could GMs Kiril Georgiev and Atanas Kolev be confident that Vallejo Pons-Kasparov RoW vs Armenia 2004 "will probably remain the last word of theory" on even one branch of the Najdorf PP?
I am not completely sure this section should be moved to footnotes, but I myself have concerns with the examples used. At first I thought they were very good, but now I realise the Marshall gambit is actually used as a drawing weapon because players believe in black's compensation, not because it has been analysed to death. On the second example, it is a bit ironic to cite as a drawing example the Poisoned Pawn just now that the e5-variation is experiencing a huge revival with mostly decisive games (e.g. Anand won with both black and white last year). SyG (talk) 17:21, 9 May 2008 (UTC)
I haven't been following the vicissitudes of the Poisoned Pawn, but if you're right about that then it probably is a bad example. Krakatoa (talk) 01:43, 10 May 2008 (UTC)
OK, I looked at chessgames.com and you are right about the recent renascence of 10.e5. I have tried to address this in the text by qualifying Georgiev and Kolev's remarks rather than presenting them as the definitive word on 6.Bg5 -- not sure if I completely succeeded. I think the citation of the Marshall is correct: many top players as White avoid it like the plague because if they play imprecisely Black's tactics crash through (e.g. Kramnik-Leko) and if they follow the trodden paths Black typically ends up in some pawn-down ending where he has the bishop pair and can hold a draw. Krakatoa (talk) 08:15, 10 May 2008 (UTC)
  • "White to play and win" contains too many irrelevant asides, e.g. Evans comment (1962) that "Mr. Adams and his cronies may be linked to the radical right wing of chess .." and the later sentence "...thus turned on its head leading hypermodern Gyula Breyer's famous statement that after 1.e4, 'White's game is in the last throes!'" Adam's own playing experiences are not a refutation as the relative strength of his opponents is not stated. Silman's scorn for Berliner's 1. d4 theory is also of dubious relevance.
I do not understand immediately why you consider these sentences are irrelevant. Evans made a statement, and the article reflects how this statement was received by the community. Once an article cite an expert it seems important to me to hint if the judgment of this expert represents the overall concensus or not. Could you please explain a bit further why you think the reception of the statements is irrevelant ? SyG (talk) 17:27, 9 May 2008 (UTC)
  • "Modern perspectives" should perhaps be split into 2 sections, "Moderate expressions of support for first-move advanatage" and "Counter-claims that Black is OK".
I agree, my feeling is that this section is too big and not perfectly organised, there does not seem to be a logical road between paragraphs, and I am not sure the title of the section represents fairly the content. I think we should choose if we want to present the article in a chronological order (and then it's OK to have a section called "Modern perspectives") or in another order still to be found (and we have some work to do). SyG (talk) 17:31, 9 May 2008 (UTC)
  • More detail would be useful (if available) on the idea that Black's OKness depends on choosing the right openings. If true, it would imply that most opening variations are the equivalent of blunders by Black.
  • On further reading "Counter-claims that Black is OK" should have 2 sub-sections, about Portisch's "it depends on the opening" and Suba's "White has to show his hand first". The biggest problem with this strucure is that there's little material about Suba's idea, and it'a mainly asides in the following sections. Philcha (talk) 18:22, 3 May 2008 (UTC)
  • "Reversed openings" and "Symmetrical openings" should be in their own section called e.g. "Effects of special types of openings". There's too much detail of individual variations.
  • Rowson's comparison between the 1st move and the tennis serve will attract the scorn of those who, like me have played both games - in tennis a Grand slam tournament winner will have lost his serve a handful of times in 2 weeks, at club level serving is a very mixed blessing (try playing social mixed doubles!) and it's usually a real disadvantage for public park players.
  • Most of "White's advantages" and "Black's advantages" add little value (and a large chunk of "White's advantages" consists of the ill-informed comparison with the serve in tennis). The most significant part, IMO, is the discussion of the Hedgehog, but that could be moved to accompany Portisch's "Black's OK if he / she uses the right opening".
  • Several cases of ref spamming, e.g. 4 in a row for first mention of Adorjan's "Black is OK", 3 for criticisms of Berliner's "1. d4 should win", 4 for "Steinitz, widely considered the father of modern chess".
  • Should mention that in 19th cent "pawn and move" odds meant the receiver played all games of a match as White, and had an extra pawn. Philcha (talk) 14:18, 9 May 2008 (UTC)
Would you please have a reference for that one ? SyG (talk) 17:34, 9 May 2008 (UTC)
I'll try, but can't promise when - I'm working on Howard Staunton at present. A quick google got me Cochrane's A Treatise on the Game of Chess: Containing the Games on Odds andThe Art of Odds Giving Philcha (talk) 21:07, 9 May 2008 (UTC)
The definition of "pawn and move" is not controversial, e.g. Hooper and Whyld, The Oxford Companion to Chess (1984), pp. 242-43 ("a handicap that consists of playing Black and removing the pawn at f7 before the start of play"). What does that have to do with this article? Krakatoa (talk) 21:45, 9 May 2008 (UTC)
Cochrane (see above) says that a giver of material odds only (P, minor piece, etc.) always plays White. P +1 (odds-giver plays Black) is an additional advantage for the receiver, P + 2 even greater. Philcha (talk) 22:33, 9 May 2008 (UTC)
I'm not quite sure how to weave discussion of pawn and move in so that it fits into the article logically. The subject of the article is "first-move advantage in chess," so it would fit in with the subject matter a lot better if Cochrane had opined on how large White's advantage was -- slight advantage, large advantage, winning advantage? I agree that the practice of giving "pawn and move" does imply that having the first move confers an advantage of some size. But as far as I can tell, that seems to be a pretty uncontroversial proposition for just any notable player in history except Adorjan. Krakatoa (talk) 01:39, 10 May 2008 (UTC)

Solving chess?

Jonathan Rowson apparently believes it's possible, but there are arguments based on Bremermann's limit that say otherwise:

"[...] no computer, however constructed, will ever be able to examine the entire tree of possible move sequences of the game of chess."[6]

However, it may well be that solving chess is possible without actually examining the entire tree, so this argument is not absolutely prohibitive.

Thoroughly fascinating article, by the way. GregorB (talk) 11:23, 7 May 2008 (UTC)

Glad you liked the article. Thanks for the Bremermann stuff. It's way over my head, but he does say in the article you quote, "In order to have a computer play a perfect or nearly perfect game [of chess] it will be necessary either to analyze the game completely ... or to analyze the game in an approximate way and combine this with a limited amount of tree searching. ... A theoretical understanding of such heuristic programming, however, is still very much wanting." That is consistent with your statement that it may be possible to solve chess without examining the entire tree. However, if Bremermann's right about the impossibility of examining the entire tree of possible move sequences, I assume that would refute Rowson's conjecture that there will someday be 32-piece tablebases? Krakatoa (talk) 18:56, 7 May 2008 (UTC)
Not necessarily - even today's tablebases are built without actually traversing the entire game tree, by using all sorts of tricks. However, just for illustration:
"For example, if we turned all the matter in the universe into the best computer possible, it would take at least 100,000 times the current age of the universe to compute all possible games of chess that can be played."[7]
...which, at least to me, makes Rowson's conjecture extremely improbable, as it doesn't appear that any tricks would help here. GregorB (talk) 20:34, 7 May 2008 (UTC)
Time is one limiting factor, but space is another. The six-piece tablebase has been completed, and it is over 1 terrabyte in size. Each additional piece increases the size by a factor of 60 or so. Before long that will exceed what can possibly be stored. Bubba73 (talk), 20:43, 7 May 2008 (UTC)
Sounds reasonable to me (not that I know jack about the subject). I did include a paragraph on Bremermann, so thanks again for the tip, GregorB. Krakatoa (talk) 20:50, 7 May 2008 (UTC)
I do not remember the reference, but I had read somewhere that there were more chess games possible than the number of atoms in the whole universe. As each chess games will always need at least one atom to be coded, it follows that it is not possible to code all chess games. Of course, this assumes we cannot code the chess games on quarks ;-) SyG (talk) 21:08, 7 May 2008 (UTC)
I have read the claim you refer to several times, but elsewhere I've read that it's not true. Don't ask me where I read all this, or who's right. Of course, when talking about the number of possible games of chess, that concept is only meaningful if you add another rule to chess, requiring one of the players, or the arbiter, to invoke the fifty-move rule. Under current rules, no one is required to claim the draw, and the arbiter can't do so, so there are an infinite number of possible games. Of course, eventually one player would die, leading to a time-forfeit -- but again, someone has to claim it. :-) Krakatoa (talk) 22:39, 7 May 2008 (UTC)
Bubba73 and SyG, you have a point: storage density is limited by the Bekenstein bound. From Hans Moravec:
"The "Bekenstein bound" leaves room for a million bits in a hydrogen atom, 10^16 in a virus, 10^45 in a human being, 10^75 for the earth, 10^86 in the solar system, 10^106 for the galaxy, and 10^122 in the visible universe."[8]
This may prove to be not enough... GregorB (talk) 21:50, 7 May 2008 (UTC)

(unindent) It is not necessary to store all legal positions at the same time to solve the game. For starters, room for only those with 32 and 31 pieces would be sufficient (or 31 and 30, etc., if that is higher). Guido den Broeder (talk) 22:07, 7 May 2008 (UTC)

If I did the arithmetic correctly, I get something on the order of 10^42 positions with 32 pieces. Bubba73 (talk), 22:40, 7 May 2008 (UTC)
In fact, game complexity says 10^50 - so yes, that is a good estimate. So, at least the storage won't be a problem... :-) GregorB (talk) 06:59, 8 May 2008 (UTC)
Storage can be reduced further by also dropping /k-1/-piece positions as soon as all /k/-piece positions leading to them have been given a value, which puts us somewhere between 1048 and 1049. Guido den Broeder (talk) 07:44, 8 May 2008 (UTC)
That is still a lot more than we will ever be able to do. Bubba73 (talk), 14:32, 8 May 2008 (UTC)
We can cut this down some more by first searching along the branches where either side is slow to lose their pawns. Once only [k]-positions are left with e.g. 4 white pawns, we can ditch all [k-1]-positions with more white pawns. Guido den Broeder (talk) 10:43, 11 May 2008 (UTC)

Reorganization

I have reorganized the article, especially the "Modern Perspectives" section, in an attempt to group the different ideas (White has an enduring initiative, Black is OK!, Black has more information than White, etc.) better. See what you think. Krakatoa (talk) 21:05, 10 May 2008 (UTC)

One thing that bugs me...

...is this:

Electrical engineer László Lovass found that of the 8577 games in the SSDF's database played between chess engines in 2000-03, White won 3248 (37.9%), drew 2927 (34.1%), and lost 2402 (28.0%), for a total White winning percentage of 54.9%, similar to human winning percentages. However, for chess engines running on the fastest PCs (1200 MHz), White scored much better, winning 910 (41.4%), drawing 974 (44.3%), and losing 316 (14.4%) of 2200 games, for a total White winning percentage of 63.5%. [emphasis is mine]

Now, 63.5% is way too high. This can't be due to stronger hardware, because three very comprehensive computer rating lists, CEGT, CCRL and Frank Quisinsky have total White winning percentages of 54.6, 54.4 and 54.4% respectively, and they all used much stronger hardware than SSDF. So, 63.5% can't be true - is there a way to verify this? (One way would be to download all SSDF games in question and check them, but the download link provided by the SSDF is apparently dead.) GregorB (talk) 22:09, 16 May 2008 (UTC)

Because the fastest 1200 MHz (strong chesswise) computers are playing games many of whcih are against weaker opponents. If you took GM results against IM's you'd have a similar % because the GM's are going to win a lot against IM's. It's only when you are comparing like for like you'll be back to the 'normal' 53-55% result for white wins. SunCreator (talk) 22:37, 16 May 2008 (UTC)
I thought of this, but then I'd expect stronger player to be stronger with white and black pieces alike. GregorB (talk) 10:07, 17 May 2008 (UTC)
You would and that appears to me to be the case from the statistics you gave above. SunCreator (talk) 12:18, 17 May 2008 (UTC)
I assume that Lovass, in writing about results "In the fastest PC category (1200 MHz)," was referring to games with 1200 MHz computers on each side. SSDF plays chess engines against each other to determine their relative strengths. To do so, one would surely run both engines on the same speed of computer -- not, say, run one engine on a 1200 MHz computer and its opponent on a 450 MHz computer. I agree that the 63.5% figure is an outlier. If someone can substitute more recent figures on computer chess results from a reliable source, that would be welcome. Krakatoa (talk) 12:07, 18 May 2008 (UTC)
OK, I have axed the Lovass SSDF statistics and substituted recent CEGT statistics instead. They're very comparable to human stats. At the slowest time control, White and Black both win about 2% less, and there are thus about 4% more games drawn, than in the chessgames.com database. White's overall winning percentage is very similar. Since these statistics will presumably change over time, can someone please change my references so they refer both to the archive of the present (May 18 2008) version and to the version available to a reader reading this article at a later date? Thanks. Krakatoa (talk) 14:19, 18 May 2008 (UTC)
Once a new faster computer comes out in order to work out an accurate Elo rating you having to play it against existing programs/computers with Elo ratings. Thus for some time the (1200 Mhz) for (a simplified example) will play against the (450 Mhz). SunCreator (talk) 11:39, 25 May 2008 (UTC)

Winning percentages table

Just wondered why the New in Chess section is blank for the White wins, drawn and Black wins percentages. Was the information not available? Not applicable? There should maybe be some explanation.Pawnkingthree (talk) 19:50, 6 June 2008 (UTC)

Probably the information was not available. I have changed the table accordingly. SyG (talk) 20:25, 6 June 2008 (UTC)
SyG's surmise is correct. Unfortunately, New in Chess consistently just gives the raw White overall winning percentage, without breaking it into White wins/draws/Black wins. As mentioned in footnote 7, you can see a similar graph (again, not broken down into White wins/draws/Black wins) at http://www.newinchess.com/Yearbook/Default.aspx?PageID=226 Krakatoa (talk) 05:08, 8 June 2008 (UTC)

Passed GA Review

The article has passed GA Review! See here for the details. Thanks to Noble Story for the suggestions and review, and to everyone who worked on the article!

The only significant problem Noble Story found is that "if you're thinking of taking this on to FAC, then I would suggest, per MOS:IMAGES, that you 'avoid sandwiching text between two images facing each other.'" No one has found a great solution to this as yet: (1) at present, we have text sandwiched between two images facing each other; (2) I tried fixing this, but ended up with several white spaces between images or diagrams; (3) SyG solved that problem, but only by eliminating a few diagrams. Since the diagrams in question seemed worthwhile to me, and eliminating them wasn't necessary to pass GA Review, I restored them, leaving the article in its current state. Any ideas? Krakatoa (talk) 01:59, 16 June 2008 (UTC)

What about my version where the diagrams were both aligned to the left? Artichoker[talk] 02:12, 16 June 2008 (UTC)
That version omits at least three diagrams that are in the current version: the final position in Vallejo Pons-Kasparov; the position after 1.e4, where Adams claims a win; and a position from near the end of Polugaevsky-Ftacnik. Krakatoa (talk) 03:34, 16 June 2008 (UTC)
The problem is that this article needs to analyse some opening / early-mid-game positions in some detail, and this requires more than 1 diagram per case. I can think of 2 possible layout solutions, although I don't know how to implement the first one:
  • Multicolumn layout, as in printed books and articles, so the continuation / analysis is entirely below the diagram. Of course it's more difficult to do this in a web page, as you don't control the page width. So a template for such a layout would need to generate some moderately clever (X)HTML and CSS. I can handle (X)HTML and CSS well enough, but I know zilch about Wikipedia templates. Some tech Web pages that might help with (X)HTML and CSS (found by Googling for "web page multi-column"):
  • This might be one of the few cases where sub-pages might be usful in Wikipedia articles. E.g. for Lev Polugaevsky-Ľubomír Ftáčnik, Lucerne Olympiad 1982:
    • on the main page show position after 19.Nf3 without the preceding moves. Then just give the main principles, e.g. "White's position looks ideal. That's the naked truth about it, but the 'ideal' has by definition one drawback - it cannot be improved."
    • on the sub-page present the full score and notes, plus as many diagrams as are helpful to those who are not blindfold experts. Philcha (talk) 08:00, 16 June 2008 (UTC)

← A couple of untried ideas:

  • put small diagrams instead of normal size,
  • or expand the text.

On more general grounds, I am not fully convinced we need to dwelve in such details like giving the full game and a latter position, so I would suggest to delete both the full notation of the game and the diagrams that come second. The casual reader knows very little about chess, so he is able to see that a position is symetric (therefore the first diagram is good) but does not have a clue about the positions not symetric (therefore the second diagram is not useful). Meanwhile, the expert reader (the one that knows a lot about chess) can still follow the full game on www.chessgames.com if we provide a link. Opinions ? SyG (talk) 18:40, 16 June 2008 (UTC)

I agree, this will help the layout of this article, as well as eliminating some information that the casual reader does not care about, but that the expert reader can always find elsewhere (i.e., Chessgames.com.) Artichoker[talk] 18:49, 16 June 2008 (UTC)
In Lajos Portisch-Mikhail Tal, Candidates Match 1965 it's easy to see how the 2nd position evolved from the 1st; the previous moves could easily be omitted; on balance I'd prefer to keep the moves that link the diagrams, as an example of how Black can gain by breaking symmetry.
Hodgson-Arkell, Newcastle 2001 makes the point that White is in "Zugzwang Lite"; I'd skip the first few moves, wikilink "Zugzwang" and summarise the rest of the game in a couple of sentences - and omit the 2nd diagram.
For Lev Polugaevsky-Ľubomír Ftáčnik, Lucerne Olympiad 1982 I'd skip the first few moves and stop after "White's position looks ideal. That's the naked truth about it, but the 'ideal' has by definition one drawback - it cannot be improved," and skip the remaining moves as White's 17th is flagged as an error. That means I'd also skip the 2nd diagram of this game.
For Francisco Vallejo Pons-Kasparov, Armenia vs. Rest of the World Match 2004 I'd be inclined to keep both diagrams but skip the moves after 9. ... Qa3 and just say that after 10. f5 best play by both sides apparently leads to the perpetual check shown in the 2nd diagram ; this makes it possible to contrast more clearly the point that 10. e5 (in the first diagram) recently favours White. Philcha (talk) 23:18, 16 June 2008 (UTC)
All this discussion of how we should dumb down the article for the supposed benefit of non-chessplayers seems seriously misplaced. Let's face it, non-chessplayers won't find this article, and probably wouldn't be interested in it if they did find it. If it becomes a Featured Article, some non-chessplayers will see it on the day it's featured, but most probably still won't find it of much interest. I have no problem with making the article more comprehensible to non-chessplayers, but making it a worse article (from the serious chessplayer's perspective) seems a really bad idea. This article is written for chessplayers, and shouldn't be compromised for the sake of non-chessplayers who probably won't be interested in it anyway. Krakatoa (talk) 02:39, 17 June 2008 (UTC)
OK, I agree the article will mostly be read by chessplayers. At first I thought there was no use of putting all the moves in the article, but now I think it makes sense when the analyses illustrate the fact that a symetrical position becomes advantageous for Black. I am still dubious on the Vallejo-Kasparov game, because:
  1. most chessplayers are lower than 2000, so in the Vallejo-Kasparov game they are not able to see it is perpetual check (BTW, maybe we should state it in the article at the end of the game, even if it is already mentioned a bit earlier).
  2. we do not provide any analysis for this game, just the blunt moves, so I do not see a purpose compared to a mere link to www.chessgames.com
SyG (talk) 09:52, 17 June 2008 (UTC)
I have now tried to use small diagrams on the Hodgson-Arkell game. Please have a look, and revert if you do not like it. SyG (talk) 10:09, 17 June 2008 (UTC)
Hi, I changed it back to the old display as your revision created an excess amount of white space and was not consistent with the layout of the rest of the article. Artichoker[talk] 11:33, 17 June 2008 (UTC)
I have butchered the article by removing four diagrams, making the Weaver Adams picture smaller, and moving the picture, all in order to avoid the dreaded "text between two images" problem. (Why is text between two images so terrible? Beats me.) If you think the cure is worse than the disease (it may well be), please revert to this version. Krakatoa (talk) 05:59, 19 June 2008 (UTC)
It looks pretty good, I just made a few minor tweaks to remove some more white space, and now I think that the article's formatting issues are mostly solved. Artichoker[talk] 14:23, 19 June 2008 (UTC)

Winning percentages

The Adorjan and GEGT results would be clearer if they were presented as tables rather than text; for CEGT, remove the 40/120 entry from the 1st table and make it part of table with the other 2 time controls in CEGT's sample. This would make it possible to introduce the CEGT paragraph with e.g. "Statistics from games between computer programs point to a similar conclusion" (that White does better at higher levels). It would also be helpful to explain the meaning of 40/120, etc. Philcha (talk) 23:14, 16 June 2008 (UTC)

Three items

(from Wfaxon (talk) 06:03, 20 June 2008 (UTC)):

  • The point should be made that, while the advantage of the first move can only be estimated using games between master-level players and above, there are many situations where one side has an undeniable advantage yet the statistics also favor a draw, e.g., K+R+P v. K+R, or as an edge case, the rare K+N+N v. K. The fact that the player with a disadvantage can often steer for such positions is certainly one reason for the high number of draws, meaning also why the advantage of the first move is so low.
  • The main statistics in the article are drawn from over-the-board play, but my understanding is that the percentage of draws is lower in top-level correspondence play. The psychological reasons for this are obvious, mostly involving required effort and fatigue. If this is a fact, it affects the estimate of White's true advantage. (If the statistics are drawn from both types of play, they should be distinguished.)
That sounds interesting and could improve the article. Would you have some references from reliable sources to back these assessments ? SyG (talk) 10:57, 20 June 2008 (UTC)

Great article

Impressive article.

Lead seemed a little long, more a mini article than a summary. But, on reflection, I think it suits the topic. Chess players will love the stats, which are introduced clearly and logically. Non-chess players need only the lead, which covers a great deal of relevant information in a short space. In this case, I'd counsel that footnotes in the lead would actually be a good thing. As a chess player, my intuition runs in sympathy with the lead, but I'm not sure I'm right (or that the lead is). For example, I play the Sicilian, which does have reactive advantages, but I'm not sure how the computers play that out. A book with good commentary on this would be a useful cite. Anyway, I'll shut up, 'cause I was asked to copyedit not review. Alastair Haines (talk) 19:48, 22 June 2008 (UTC)

The lead was less than half its present size before the article was nominated for FA, as you can see if you look at this version. Then an FA commenter protested that the lead was too short, so it was made longer. Hard to make everyone happy. Krakatoa (talk) 20:50, 22 June 2008 (UTC)
Thanks for that. I think these things are sometimes not well understood at Wiki.
  • There is no single, correct, best final version, just many excellent ones
  • FA review comments aren't always correct, just always worth discussing
(they can be challenged and need to win consensus like all other comments)
  • Making people happy isn't the aim of the process, but addressing the issues raised by the people is.
Anyway, I like both leads. I don't think "long" or "short" is much of an issue, content is. The new lead has excellent content. As they say near Krakatoa, hati hati. Alastair Haines (talk) 02:04, 23 June 2008 (UTC)

Removed comment about Fischer and Kasparov from lead

I've removed the bolded section in this passage from the lead:

The prevalent style of play for Black today is to seek dynamic, unbalanced positions with active counterplay. World Champions Fischer and Garry Kasparov were the foremost exponents of this style of play.

Krakatoa (talk · contribs) has correctly noticed that this claim is referenced later in the section on dynamism:

Watson observes that "energetic opening play by Black may ... lead to a position so complex and unclear that to speak of equality is meaningless. Sometimes we say 'dynamically balanced' instead of 'equal' to express the view that either player is as likely as the other to emerge from complications with an advantage. This style of opening play has become prevalent in modern chess, with World Champions Fischer and Kasparov as its most visible practitioners."[79]

However, I believe that this remark should be removed from the lead section of the article. To give context to my rationale, recall from the content guideline on lead sections:

The lead section should briefly summarize the most important points covered in an article in such a way that it can stand on its own as a concise version of the article...in a well-constructed article, the relative emphasis given to information in the lead will be reflected in the rest of the text.
  1. This remark is given undue weight in the lead, because it is not further explained in the body of the article; we spend the same amount of time on the claim in both places, except we don't cite it in the lead. Lead sections should give an overview of the subject, and the emphasis placed on this claim in the article is not well reflected by featuring it prominently in the lead section. Discussing who were the major proponents of this dynamic style of play should fit to expand the claim made in the lead section: The prevalent style of play for Black today is to seek dynamic, unbalanced positions with active counterplay.
  2. I also have to take issue with the way it is cited, or rather not cited, in the lead. Of course, Watson's reputation as a commentator on chess theory is highly regarded. His opinions on this kind of thing merit encyclopedia inclusion if anyone's do. But I have strong reservations about stating his opinions about which players are most prominent as bald fact; not because I personally disagree with him, but because such claims are in principle subject to contention.
  3. Finally, bringing this all together. Given that claims like this are in principle subject to dispute due to their prevalence in Wikipedia occurring primarily due to hero worship, restricting the claim to citing Watson's opinion makes the reader's experience more informative. Consider the difference in experience between a reader who sees a claim made on Wikipedia that "Fischer was a leading proponent of the dynamic style of chess" versus "Chess analyst John Watson (credentials) said that Fischer and Kasparov were the most visible practitioners of this style of play." To the reader, citing the claim makes it more forceful, because now we have expert opinion to back it up; stating it the other way leaves the reader with the impression that "some guy" just added it to the article, which isn't convincing. "I read it on Wikipedia" is much less helpful than "International Master John Watson, author of the seminal Secrets of Modern Chess Strategy, says so." Granted, we can't cite everything in this way, but given the rationale above, I think this claim merits strict scrutiny.

If there are any other objections, I'd be pleased to discuss them here so we can reach consensus. Overall I think this is a very good article on a difficult topic and I'm happy to see more chess articles reaching featured status! Cheers,

--causa sui talk 12:36, 28 June 2008 (UTC)

Is your problem with (a) saying that F and K were exponents of this style of play, (b) that they were the foremost exponents, or both? Proposition a seems uncontroversial - for example, both played the Sicilian Najdorf (including the Poisoned Pawn) and King's Indian throughout their careers, and those are paradigmatic openings for this style of play. I could see that (b) is somewhat less clear. One could say that Topalov and Anand play that way, too, and that both achieved higher ratings than Fischer ever did. (Probably the result of ratings inflation, but that's another story.) I have softened "the foremost exponents" to "prominent exponents"; hope that addresses your concern. Of course, given that F and K both no longer play, neither is now an exponent of this style of play, so probably "were prominent exponents" is the way to go for that reason, as well. Krakatoa (talk) 13:12, 28 June 2008 (UTC)
My two cents:
  1. For an article having reached FA-status, it is good practice to discuss important changes before actually doing them. ;-)
  2. Maybe giving two full lines in the Lead to one single statement is a bit overweight, especially when it seems we have few other references that concur with this statement.
  3. Watson is clearly a highly regarded chess expert. I have a hard time believing, however, that dynamic play has become that prevalent. It seems to me there are still a lot of players who do not play that dynamically with Black, possibly Kramnick or Leko. On that statement I would like to hear the opinion of Krakatoa or Sjakalle, as they are both strong enough to have an authoritative opinion on the leading style of play nowadays. SyG (talk) 13:32, 28 June 2008 (UTC)
Botvinnik was also noted for his success in playing aggressively as Black, although his style was different from that of Fischer and Kasparov. No doubt other editors will find other good examples. That makes me think the lead should simply say e.g. "Some of the greatest world champions were exponents of this style of play," while the content names names. Philcha (talk) 09:35, 29 June 2008 (UTC)
The lead is of course a condensed version of the rest of the article. Watson named F and K as "the most visible practitioners" of this style of play; the lead refers to them as "prominent exponents" of it. This seems certainly true, although of course one could name others. Chess-literate readers are likely to be familiar with F and K's style, and less likely to be familiar with Botvinnik's. It accordingly seems to me to be helpful to the chess-literate reader to give these two examples. Krakatoa (talk) 09:47, 29 June 2008 (UTC)

I want to be clear by restating exactly what I am and am not objecting to. For starters I am not objecting to mentioning this tidbit at all; of course it fits well later in the article. I am not objecting to the factual nature of the claim; Watson's opinion is as notable as anybody's. I am objecting to mentioning it in the lead, especially in its current form. To avoid repetition, as I explained above the short reasons I am objecting to that are that it states Watson's editorial opinion as bald fact, and that it is out of balance given that the lead should be a condensed summary of the article. Therefore I unfortunately have to say that it does not address my objection to reword the statement as Krakatoa so nobly attempted here [9]. --causa sui talk 23:25, 29 June 2008 (UTC)

Are you claiming that the proposition that Fischer and Kasparov are prominent exponents of dynamic play as Black is remotely controversial? Everyone knows that (or so I thought); it is not remotely controversial. As for your other point, I've already explained above why I think it's reasonable for it to be in the lead. Krakatoa (talk) 02:37, 30 June 2008 (UTC)
No, once again, I'm not disputing whether it's true or not. I am disputing why it belongs in the lead. I understand you don't agree; that's why I'm participating in this talk page discussion. :-) I think we are talking past each other here somewhat, and I'm not sure what the cause of the miscommunication is. --causa sui talk 02:43, 30 June 2008 (UTC)
I have eliminated reference to F and K, and changed the previous sentence to "The prevalent style of play for Black today is to seek dynamic, unbalanced positions with active counterplay, rather than merely trying to equalize." (Boldface = new material added) Krakatoa (talk) 05:57, 30 June 2008 (UTC)
I think that's very good. Thank you. --causa sui talk 08:41, 30 June 2008 (UTC)

How important?

This article, recently elevated to Featured Article, is rated as being of "Mid-importance". That is the third-highest category, below "High" and "Top" importance and above only "Low" importance. I'm not sure if "Mid-importance" is the right classification. The article is about the size and nature of White's advantage. It discusses Adams, Rauzer, and Berliner, respected masters who claim(ed) that chess is or may be a forced win for White. It seems to me that whether the game is a forced win for White or not is a pretty fundamental (and thus important) question for chessplayers. I suggest that "High" importance is a more appropriate classification. What does everyone else think? Krakatoa (talk) 19:50, 3 July 2008 (UTC)

"High" importance is OK by me. Bubba73 (talk), 20:15, 3 July 2008 (UTC)
I have no problem with a move to high importance, the article has clearly demonstrated it.Caissa's DeathAngel (talk) 20:52, 3 July 2008 (UTC)
Really, there are many much more important articles than this. Interwikis tell us something, too. I oppose. Lab-oratory (talk) 06:08, 4 July 2008 (UTC)
I... don't know! Probably Krakatoa is right when he says this article is important to chessplayers, but then most of the things that are in "Mid" are important for chessplayers as well, so they should all go into "High" ? Also, the assessment must take into consideration everyone, not only chessplayers. That means the "Top" category should be for everything that a non-chessplayers must know, and maybe the "High" category would be for everything that a non-chessplayer might be interested in. All in all, I would slightly agree to an upgrade to "High" for this article. SyG (talk) 17:22, 4 July 2008 (UTC)
SyG raises another interesting question. I would think that the rating should be somewhat project-specific. For example, Anatoly Karpov is rated Top-Importance by us, but only High-Importance by both WikiProject Russia and WikiProject Soviet Union. Also, it seems to me that something can be "High" or even "Top" importance to a particular project even though it may be of little interest to the general public. Recent Main-Page article Oxidative phosphorylation is about "a subject of High-Importance within molecular and cellular biology", but I daresay most lay people know nothing about it and care less.
Anyway, it seems to me that other aspects of First-move advantage in chess are of even greater practical importance than the one I referred to above. Consider the nature of White's advantage, and how Black should respond to it. In response to 1.e4, for instance, for most of chess history it's been commonly thought that Black should respond with 1...e5 and try to equalize. As late as the Seventh Edition (1946) of Modern Chess Openings, all other responses were lumped into the category "irregular defenses". (MCO-7, p. 331: "the Sicilian, which many claim to be the best of the irregular replies to 1 P-K4".) In the Eighth Edition (1952), the Sicilian, French, Caro-Kann and the rest finally reached the status of "asymmetrical replies to P-K4" (p. 261). Now the Sicilian is the most common and most successful response to 1.e4, and the most important variations of it are the Najdorf and Sveshnikov variations, which would have been considered bizarre for most of chess history. (Even in the mid-1970s, I remember being shocked by the Sveshnikov -- wasn't that just weak? That was the received wisdom from Fine, Horowitz, Reinfeld, etc.) The Poisoned Pawn variation (7...Qb6 intending 8...Qxb2), played by Fischer, Kasparov, Anand and other greats, would have been considered a beginner's move for most of chess history, yet today Watson calls it the main deterrent to 6.Bg5. Similarly, the Hedgehog Variation of the English would have been derided as bizarre, passive, weak, etc. but now it's a very important line. The article addresses theoretical and practical questions that are very important to chess. Krakatoa (talk) 18:52, 4 July 2008 (UTC)
The importance rating is ofcourse highly relative to the project. From Wikipedia:Version_1.0_Editorial_Team/Release_Version_Criteria#Importance_of_topic:
Importance or Priority must be regarded as a relative term. If importance values are applied within this project, these only reflect the perceived importance to this project. An article judged to be "Top-Class" in one context may be only "Mid-Class" in another.Voorlandt (talk) 21:13, 4 July 2008 (UTC)
On the specific question: I think high is ok. When assessing articles, it is often useful to look at the number of interwiki's. Typically, high importance articles will have around 5-15 interwiki's, while mid 1-7 and low 1-2. Ofcourse this is just a rough guideline (and not fixed in times-obviously) and there are tons of exceptions. In this specific case, the subject is obviously very important for the game of chess, and I think high is ok. Voorlandt (talk) 10:14, 5 July 2008 (UTC)

OK, this is enough of a consensus to me, and I think the argument of interwiki, while generally valid, does not apply here because the article is too young. I changed the importance to "High". SyG (talk) 10:23, 5 July 2008 (UTC)

One comment: I think the topic of this article is profoundly important in opening theory. That makes me lean towards "high" importance. Bubba73 (talk), 00:39, 6 July 2008 (UTC)

Congratulations!

This is a great article! I missed the FAC, but it really is accessible and does a very good job of covering the theory on this. Not sure if it covers everything (is there more on the early stuff - what about views before Steintz and more details on the 19th century and early 20th century views?), but it certainly covers a lot. Well done. Carcharoth (talk) 13:18, 8 July 2008 (UTC)

Thanks! Glad you like the article. I have looked for pre-Steinitz views, but found nothing. I bought the English translation of Philidor's book just so I could see what he said, but found nothing. (Weirdly, he did claim that all gambits were drawn with best play. Does that apply to, say, the Jerome Gambit?) I also have Staunton's The Chess-Player's Handbook (it's available online, too), but found nothing relevant there, either. The views of Lasker and Capablanca are discussed a little more in the footnotes, but neither addressed the subject at any length. I have looked high and low for everything relevant I could find (I hope it shows!). If you find anything else that should be in the article, let me know. Krakatoa (talk) 22:17, 8 July 2008 (UTC)

Diagrams side by side

A previous disussion concluded that one can't have chess diagrams side by side on the right, and the article now has several pairs of diagrams on the left. This is better than sandwiching the text, but IMO is bad for readability because it changes the positions of the starts of lines. There's no official technique for grouping image objects on the right, but I found one that works for simple images, see Small shelly fauna.

Here's how it would look for the first example in First-move advantage in chess:

abcdefgh
8
a8 black rook
b8 black knight
c8 black bishop
e8 black king
f8 black bishop
h8 black rook
b7 black pawn
f7 black pawn
g7 black pawn
h7 black pawn
a6 black pawn
d6 black pawn
e6 black pawn
f6 black knight
g5 white bishop
d4 white knight
e4 white pawn
f4 white pawn
a3 black queen
c3 white knight
a2 white pawn
c2 white pawn
d2 white queen
g2 white pawn
h2 white pawn
b1 white rook
e1 white king
f1 white bishop
h1 white rook
8
77
66
55
44
33
22
11
abcdefgh
The Poisoned Pawn Variation, position after 9...Qa3
abcdefgh
8
c8 black bishop
h8 black king
f7 white queen
h7 black pawn
a6 black pawn
c6 black pawn
e6 black pawn
f6 black pawn
e5 black pawn
c3 white king
g3 white rook
f2 black queen
g2 white pawn
h2 white pawn
8
77
66
55
44
33
22
11
abcdefgh
Final position of Vallejo Pons-Kasparov: the perfect game?

In 2007, GMs Kiril Georgiev and Atanas Kolev asserted that much the same was true of the so-called Poisoned Pawn Variation of the Najdorf Sicilian, which arises after 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Qb6!? This has long been considered one of the sharpest and most problematic (or even foolhardy)[1][2][3] opening lines.[4][5] The game usually continues 8.Qd2 Qxb2 9.Rb1 Qa3.[6] Georgiev and Kolev stated that 6.Bg5 is seldom seen at the highest level because the main line of this variation leads, with best play, to a draw by perpetual check.[7] They wrote that the following game "will probably remain the last word of theory":[8]

Francisco Vallejo Pons-Kasparov, Armenia vs. Rest of the World Match 2004[9][10]: 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Qb6 8.Qd2 Qxb2 9.Rb1 Qa3 10.f5 Nc6 11.fxe6 fxe6 12.Nxc6 bxc6 13.e5 dxe5 14.Bxf6 gxf6 15.Ne4 Qxa2 16.Rd1 Be7 17.Be2 O-O 18.O-O Ra7 19.Rf3 Kh8 20.Rg3 Rd7 21.Qh6 Rf7 22.Qh5 Rxd1+ 23.Bxd1 Qa5 24.Kf1 Qd8 25.Qxf7 Qxd1+ 26.Kf2 Qxc2+ 27.Kf3 Qd1+ 28.Kf2 Qc2+ 29.Ke3 Bc5+ 30.Nxc5 Qxc5+ 31.Kd2 Qf2+ 32.Kc3 Qd4+ 33.Kc2 Qf2+ 34.Kc3 1/2–1/2 (After 34...Qd4+, White cannot escape the checks.)

However, Georgiev and Kolev's pessimistic assessment of 6.Bg5 has since been called into question, as White has succeeded with 10.e5 (another critical line) in several recent high-level games.[11]

(end of example)

It's simply a table:

{| align="right" border="0" cellpadding="0" cellspacing="0"
|-valign="top"
|{{Chess diagram
|tright
...
}}
|{{Chess diagram
|tright
...
}}
|}

-- Philcha (talk) 10:04, 21 July 2008 (UTC)

I'm not sure whether it improves readability, but it's certainly an extremely useful thing to be able to do. (As you'll recall, in the previous discussion, no one knew how to do it, so I had to put all pairs of diagrams on the left.) I have indeed changed the example you cited. Thanks! Krakatoa (talk) 15:36, 21 July 2008 (UTC)

feasible that computers will ever solve chess

Just reversed some edits that said it would NOT be feasible that any computer will ever be able solve chess. It is original research(or at least unreferenced) and it also contradicts some of the existing references in this article. It also contradicts my personal view (although granted my personal view on this is irrelevant to wikipedia) SunCreator (talk) 16:42, 4 March 2009 (UTC)

My understanding is that it is theoretically possible but will never be feasible. I'll try to check some of the references. Bubba73 (talk), 16:45, 4 March 2009 (UTC)
Rowson doesn't seem to understand the computation aspect the way Bremermann does. Bubba73 (talk), 16:54, 4 March 2009 (UTC)

There are three relevant references in this section: Shannon, Rowson and Bremermann.

  • Shannon shows that it would take a computer, calculating one complete game per microsecond, 1090 years to "solve" chess.
  • Rowson gives the opinion that it will happen (taking "decades or even centuries")
  • Bremermann mentions several physical limitations (light barrier, quantum barrier, and thermodynamical barrier) that will make it impossible.

Given that 1090 years is immensely much more than the age of the universe, this basically means that two of the three references are saying it is effectively impossible. Note also that Shannon gives a number of 10120 for the number of possible games, and that the number of atoms in the visible universe is about 1080, which means that to actually store the solution to chess would require storing 1040 games per atom in the visible universe. Those are the kinds of limitations Bremermann is talking about. HermanHiddema (talk) 17:05, 4 March 2009 (UTC)

IMO, the current references therefore support the view that it is "infeasible" (not "impossible") that a computer will ever solve chess. HermanHiddema (talk) 17:05, 4 March 2009 (UTC)

Yes, and Rowson talks about constructing 32-piece databases. In the endgame databases, adding one more piece takes about a factor of 40 more computation. It took on the order of 10 years to do the six-piece endgames. So we are talking about a factor of 1041 more computation. Even with a billion times as many computers, each a billion times faster, that is still 1024 years. And the age of the Earth is less than 1010 years. And even if you could do the computation, there would be no way to store that much data.
And there are physical limits to computation. As mentioned, there aren't enough particles in the Earth to store all of that information. And for a computer large, it would have to be physically so big that it would take too long too access the memory, so a computer system that is large enough and fast enough is not feasible. Bubba73 (talk), 17:15, 4 March 2009 (UTC)
The above quotes are out of context. There is a very big difference in solving every possible chess position versus solving the game of chess. Solving the game of chess requires nothing like solving every position. SunCreator (talk) 17:17, 4 March 2009 (UTC)
Direct quote: According to Rowson, "in principle it should be possible for a machine to ... develop 32-piece tablebases." That does mean every possible position. Bubba73 (talk), 17:44, 4 March 2009 (UTC)
Yes. A 32-piece tablebases would solve every possible chess position. Yet, that is NOT required to solve the game of chess. Thus the confusion. SunCreator (talk) 17:56, 4 March 2009 (UTC)

In game theoretical terms, there are three possible ways to solve a game:

  • Ultra-weak: prove that the game is a win/draw/loss without providing perfect play, but by eg strategy-stealing.
  • Weak: provide perfect play from the starting position.
  • Strong: provide perfect play from any possible position.

When referring the number 10120, Shannon is referring to perfect play from the starting position, ie: weak solving. Since the article is clearly not referring to ultra-weak solving, the argument I make above is perfectly valid. HermanHiddema (talk) 17:58, 4 March 2009 (UTC)

My apology. I was misunderstanding what Shannon wrote. He is talking about Weak play when referring the number 10120. SunCreator (talk) 18:25, 4 March 2009 (UTC)
Does anyone have the original text from Rowson? The current link is dead. SunCreator (talk) 18:25, 4 March 2009 (UTC)
No, I'm sorry, I don't have the text for Rowson. As you pointed out above, when talking about 32-piece tablebases, Rowson is effectively talking about strong solving chess. An infinitely more difficult feat than weak solving it (for an idea: there are about 1050 legal positions, from each of which perfect play must be given...)
Personally, I think Rowson is simply talking in general terms (note that he talks about "decades, or centuries") and is extending current computer chess to a point where heuristics based engines approach reasonably perfect play. In such a case, chess would perhaps by functionally solved, but not solved in any strict game theoretical sense, where even for the most ridiculous counter move we must consider all options until we know perfect play. HermanHiddema (talk) 18:57, 4 March 2009 (UTC)
I've undo one of my reverts, putting back most of what you previously inserted. I'm not sure if it's correctly referenced or not as I don't know the Rowson text. SunCreator (talk) 19:11, 4 March 2009 (UTC)
For the purposes of this article (whether chess is a draw or if White/Black has a forced win), we are talking about ultra-weak solving. Bubba73 (talk), 20:21, 4 March 2009 (UTC)
An ultra-weak solution would suffice, but it is not necessarily the case that there is an ultra-weak solution that is easier than the weak one. There is no known ultra-weak solution, so for purposes of this article it is better to stick with the weak one (because it is certain that that one exists). HermanHiddema (talk) 22:37, 4 March 2009 (UTC)

(unindent) From Rowson, pp. 205-06 of Chess for Zebras:

Before proceeding I should just clarify that 'hypertheory' does not mean 'computer theory'. 'Hypertheory' is the omniscient view of chess, the view of chess 'under the aspect of eternity'. The reason I mention computers in this context is that chess is theoretically finite. No human could ever grasp chess from a hypertheoretical perspective, but in principle it should be possible for a machine to reach this fundamental perspective and develop 32-piece tablebases. This may take decades or even centuries, [end of page 205] but unless runaway global warming or nuclear war gets in the way, I think it will theoretically eventually happen.


Rowson is indeed talking about strong solving. There's no indication whether he has any idea of the practical problems involved in this, or what if any expertise he has on the subject of computers. Krakatoa (talk) 05:36, 5 March 2009 (UTC)

I don't understand what he means by "I think it will theoretically happen." In theory/principle, solving chess is possible, given enough computing power and time. In reality/practice, probably not. Bubba73 (talk), 05:44, 5 March 2009 (UTC)
Sorry, my mistake. He wrote "eventually" - not "theoretically". Krakatoa (talk) 06:37, 5 March 2009 (UTC)

According to the article about him, Jonathan Rowson has no expertise in either mathematics or computers. Shannon and Bremermann on the other hand clearly do. On top of that, Shannon has actually researched the issue scientifically, and is quite well-known for it (eg: Shannon number). I will try to do a bit of a rewrite to incorporate this in the text. HermanHiddema (talk) 08:49, 5 March 2009 (UTC)

About 30 years ago I read an article by Bremermann on the physical limits of computation. Bubba73 (talk), 15:27, 5 March 2009 (UTC)
This bit about checkers that someone added to the end of the article is interesting:

"The game of checkers was solved in 2007,[162] but it has roughly the square root of the number of positions in chess. Jonathan Schaeffer, the scientist who led the effort, said a breakthrough such as quantum computing would be needed before solving chess could even be attempted, but he does not rule out the possibility, saying that the one thing he learned from his 16-year effort of solving checkers "is to never underestimate the advances in technology".

I would think that Schaeffer must have considerably more understanding of the issue than Shannon, given that (a) he spent 16 years solving checkers, whereas Shannon was just theorizing about the possibility of solving chess and (b) Schaeffer has far more understanding of what's possible with computers today (and may be possible in the future) than Shannon could have had in 1950. Krakatoa (talk) 16:39, 21 March 2009 (UTC)
  1. ^ Watson observed that 7...Qb6 "is an astonishing move that those raised with classical chess principles would simply reject as a typical beginner's mistake. Black goes running after a pawn when he is undeveloped and already under attack." Watson 2006, p. 199.
  2. ^ "Referring to the Poisoned Pawn Variation ... the brilliant, classically-oriented grandmaster Salo Flohr commented, even as late as 1972: 'In chess, there is an old rule: in the opening, one must make haste to develop the pieces, and must not move the same piece several times, particularly the queen. This ancient law holds good even for Bobby Fischer'." Watson 1999, p. 18.
  3. ^ Siegbert Tarrasch wrote in his classic treatise The Game of Chess, "It is very dangerous to make a raid with the Queen early on in the game. In particular the capture of the Queen's Knight's pawn with the Queen often brings its own revenge." Tarrasch, S. (1938). The Game of Chess. David McKay. pp. p. 220. ISBN 978-1880673942 (1994 Hays Publishing edition). {{cite book}}: |pages= has extra text (help)
  4. ^ Geller, Y., Gligoric, S., Kavalek, L., and Spassky, B. (1976). The Najdorf Variation of the Sicilian Defence. RHM Press. pp. p. 98. ISBN 0-89058-025-1. {{cite book}}: |pages= has extra text (help)CS1 maint: multiple names: authors list (link)
  5. ^ Sokolsky, A. (1972). The Modern Openings in Theory and Practice. Pitman Publishing. pp. p. 154. ISBN 0-273-31409-2. {{cite book}}: |pages= has extra text (help)
  6. ^ Watson 2006, p. 199.
  7. ^ Georgiev and Kolev 2007, p. 10.
  8. ^ Georgiev and Kolev 2007, p. 11.
  9. ^ "Vallejo Pons vs Kasparov (2004)". ChessGames.com. Retrieved 2008-06-28.
  10. ^ Harding, T. (2006). "The Poisoned Pawn is Still Looking Tasty" (PDF). ChessCafe.com. Retrieved 2008-06-27. Page 10 mentions the Vallejo Pons-Kasparov game.
  11. ^ Bücker, S. (2008). "Poisoned Pawn for Tarriers" (PDF). ChessCafe.com. Retrieved 2008-06-27.