Jump to content

Talk:Chessmetrics

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

Self-promotion ?

[edit]

I believe this guy is just trying to promote his chessmetrics idea. He has already gave himself 5 free plugs on every chess biography I read here. —Preceding unsigned comment added by 70.221.226.3 (talk) 21:09, 22 December 2007 (UTC)[reply]

"Not notable" template removed

[edit]

Chessmetrics is a widely-used rating system, and is cited in many chess articles. Moul & Nye (cited) selected for their analysis of whether the Soviets really did collude at the 1962 Candidates' Tournament, and explained why they considered it superior to similar comparison techniques. There is no doubt that it is notable. -- Philcha (talk) 21:31, 15 September 2008 (UTC)[reply]

Thanks, I assumed I used the template correctly. I would like to discuss the idea of demoting chessmetrics from a mathematical tool, to a research paper done by someone with no degree in statistics or any math field. Compare to Petrodollar warfare, it has been repeatedly agreed to treat this as a book, not as a hypothesis. However, the people who are trying to make the hypothesis more noteable and more popular, keep changing the article back to its current state. I do not want the article deleted in 5 days, I hoped someone would pull it down, so that a discussion can take place about removing chessmetrics from every chess article on wikipedia, except for the 3-4 that it relevantly serves. I'm also a member of a chess community, and have heard unfair things about Sonas, which I would like resolved. The article needs to say that Sonas isn't some self-proclaimed expert, that he has a degree in statistics, and isn't trying to just create stuff like another Eric Schiller, who makes stuff for its own sake. Also, its a horrible wile, if true that he created his formula to "predict" that bobby had a higher peak than kasparov, but that kasparov had a better long-term chessmetrics rating. This is the most troubling part of it all, to use ex post facto data to give credibility to your formula when tested on that same ex post facto data. I think the article should not be about the hypothesis, but should be about his research paper, and will check back on this page in a week. Sentriclecub (talk) 22:13, 15 September 2008 (UTC)[reply]
Yes I read the thing about Soviet collusion, I'm a huge chess fan. I like following some of the "dirt" in the chess world too, like the Nigel Short handshake debate, etc... Simply put though, I hope you don't think I believe chessmetrics isn't notable, its just not (as a hypothesis) worthy to be included in an encyclopedia (as a hypothesis) but should be included as a research paper. I know some people have strong opinions on chess matters, so I hope we'll both devote extra effort into smooth communications and trying to focus on the facts, as they are related to making wikipedia a better encyclopedia. Sentriclecub (talk) 22:18, 15 September 2008 (UTC)[reply]
I'm not sure how much of your comment I understood. If there are things you think the article should say and you can provide WP:RS sources for them, then insert them with citations in the right places. -- Philcha (talk) 22:26, 15 September 2008 (UTC)[reply]
Sorry I wasn't clear. I have heard bad things about Sonas from people I know who are Fide masters, and IMs and "in the know" such as that he is a self promoter and self proclaimed expert, and he has upset a lot of people in the USCF. The burden of proof is on you. I want the article deleted, unless reliable sources say that he didn't create a formula using ex post facto data, to boost and help sell its credibility. A google search shows that 500 pages of chessmetrics are in the en.wikipedia.org subdomain. Can we allow him to contaminate every chess article that exists? For us to rely on chessmetrics, and use it on tons of wikipedia pages, and we don't even know who jeff sonas is? Please read the entire page about WP:VERIFIABILITY is way more important than WP:CITATION. Just because his work is cited, and published in chessbase, does not validate his hypothesis. This article needs to be about his research paper, and all the chess articles (except the 3-4 which chessmetrics relevantly serves) need adhere to the principle of being an encylopedia article. I for one have tremendous open-mindedness about this guy. I hope that chess rumors are just that. Hopefully it will be like the 61 memorable games facade. I want a good outcome, and if this guy is an actual gem to society, great. However, there are some strong opinions of people I look up to, that accuse him of chesspolitics and are extremely upset with his method. I'm probably the only person who can discuss it here civilly. I am acquaintance to some strong minded people, and I want chessmetrics cleared up. This is the talk page for the article, so I'm in the right place. Ultimately, an encyclopedia is not a place to throw stuff at a wall, and see if it sticks. Chessmetrics got into wikipedia with zero resistance. There are a lot of people in wikiproject:chess who are at odds with a core wikipedia policy. Please expand this article, and accept the burden of proof. I'll put up a invitation at the wikiproject chess, in a few days. I'm in no rush at all, this could take weeks to figure out more info about Jeff Sonas and incorporate it into this article. Sentriclecub (talk) 23:48, 15 September 2008 (UTC)[reply]
Can you explain what you mean by "ex post facto data"? Bubba73 (talk), 01:28, 16 September 2008 (UTC)[reply]
As the article stands, it should be up for deletion as non-notable. The tag was added properly. Removing it was done absolutely improperly. Once up for deletion, a discussion must take place before removal of the tag. My impression is that Philcha is closely related to the author of the page (if not the author) and is attempting to bypass the deletion process for a page he knows will not pass any scrutiny. So, unless Philcha can provide some citations from someone who did not create the chessmetrics process, I suggest it be nominated for speedy deletion, not just non-notable deletion. It is very simple, refusal to follow the rules indicates that your article is less notable, not more notable. -- kainaw 00:15, 16 September 2008 (UTC)[reply]
I've just checked WP:CSD and Chessmetrics meets none of the criteria.
Two academic economists, Moul & Nye, found Chessmetrics the most reliable system for their purpose, I suggest you read their article. A number of online chess mags have found the system notable and reliable enough for them to publish articles based on its results. For a non-academic subject, that's as good as it gets.
Kainaw, your statement "My impression is that Philcha is closely related to the author of the page (if not the author) and is attempting to bypass the deletion process for a page he knows will not pass any scrutiny" is grossly uncivil. I expect an immediate apology.
Sentriclecub, this article is not about Jeff Sonas or his relationships with members of the USCF, it is about a statistical system for rating chess-players. As I've already pointed out, two academics are on record as preferring Chessmetrics over the Elo system for assessing whether the Soviets colluded.
Sentriclecub, I don't know what your statement "There are a lot of people in wikiproject:chess who are at odds with a core wikipedia policy" was intended to mean, but it sounds very like some sort of grudge. It might be a good idea for you to explain that statement clearly, concisely and with evidence. -- Philcha (talk) 01:53, 16 September 2008 (UTC)[reply]
Chessmetrics is widely used in the chess coummunity, e.g. it has had several articles in the major internet chess publication Chessbase, e.g. this one[1] I agree that Chessmetrics ratings are overused in Wikipedia and have written about their overuse on a number of occasions. However Chessmetrics is widely enough used and written about that it is definitely notable. Peter Ballard (talk) 03:20, 16 September 2008 (UTC)[reply]
p.s. I agree that Kainaw's accusation is uncivil and warrants an apology. It is also bizarre. Just check the edit history - Philcha's contributions to this article have been minimal. Peter Ballard (talk) 03:56, 16 September 2008 (UTC)[reply]
Also, contrary to Kainaw's assertion, Philcha's removal of the template was perfectly proper. Philcha removed the template, noted it in the edit summary, and explained reasons on the Talk page. That was exactly what is required according to Wikipedia:Proposed deletion#Contesting a proposed deletion. Peter Ballard (talk) 03:56, 16 September 2008 (UTC)[reply]
Please wait, Don't group me together with everyone else. I don't want to disrespect the editors who have made all the hundreds of chessmetrics contributions. My realistic goal is that there's no way, I can single handedly "cleans" chessmetrics from wikipedia. I want the article to be about the research paper and not about the hypothesis.
Also, to Philcha, is it okay if I remove the blue ink from my name and kainaw? At first I thought it was my signature. You are mistaking (although I hesitantly doubtfully suspect intentionally, because of the inherent chesspolitics nature of the problem) and abusing number theory. 1 can fairly be the integer, specifying a rank or order, in a set. However, you made the straw man fallacy, in the fact that it is the only system which tries to remove the inflation out of elo. Is there a single other system that has been or is cited by academia to try to take out the inflation from elo? Also, I have a couple of acquaintances in academia and you are misinterpreting the signficance of it being included as a citation. When a grad thesis is done, usually you use about 50-100 sources, and since the thesis is precisely 100% exactly related to what chessmetrics is about, it would be absurd to not include it. If I was doing a grad thesis on ways to calculuate water in underwater aquifers on mars, and there is a popular article written exactly on my narrow narrow subject, then I'm forced to cite it. Yes that adds to notability, but inclusion as a citation does not equate with verifiability. If we made an article for every citation in every academia thesis, we would cover every topic.
Please let me escape with brevity. I did not adequately respond to your good points. As you may know, since you are familiar with chessmetrics and maybe know or have heard or know those who know about Mr. Sonas. There's always drama in the professional chess bodies, like Fide and USCF, and we are both aware, even if only slightly, about some of the different groups of people that have different opinions about strong and heated issues. There are lots of unfair things said about Mr. Sonas, especially in chess circles that intersect the wikipedia chess portal community and its hardworking editors, but also chess-gossip-websites (for lack of a less humble term) such as redhotpawn, 65thsquare, chessgames, etc... I would like permission to keep this stuff, which I have overly generalized, completely out of this discussion, since its purely hearsay. I hope you accept my request to keep all this outside stuff sequestered from our discussion about chessmetrics only. Addtionally, I think Kainaw's response was a little bit heated, but if he reviewed your contribs, he'd see the pattern that you are part of the group of editors which make lots of edits in the chess articles. The chess editors are a very polarized group, and you have a lot of power in the way chess articles are edited. I apologize for stepping on your feet, because I mainly edit math and science articles, but as we will discuss in the future, there are fatal flaws with the fact that chessmatrics has spread through pages which you help edit, and has met absolutely ZERO editorial resistance. In my math and science wikiportals, we wouldn't spread a controversial hypothesis from an unknown author (Mr. Sonas is redlinked--how do I know if he has a PhD in statistics from an ivy leagure school, or as sadly well quite not the case). He never sent his hypothesis to peer review he just had some chesspolitical connections with chessbase, and it was an entertaining article, but its not a verified hypothesis. You seem to treat notability as a permit. Notability You can't just use it like a backstage pass to have unlimited rights to post this to 500 pages in the .en.wikipedia.org subdomain. In the math and science portals, we have this problem with creationism. We point to WP:VERIFIABILITY and it usually ends in flaming and its such a chaotic mess. If you want to include Jeff Sonas's original research, then you had better explain that Jeff Sonas wrote the research paper and it was because of his keen mathematical vigor, that it got great reception by chessbase (and not through chesspolitical networking in the good-ole-boys network). You see that by dumping the burden of proof on me, I have to skirt the touchy subject about all this outside gossip, which I'm committed to keeping sequestered from this discussion about chessmetrics. The article is an insta-shrink. It should only exist on about 3-4 relevant articles (where it complements and adds value to the articles and to the encyclopedia as a whole). The only way you can win this argument is to try to win it on the argument itself. The facts are clear. Creationism has been published in many magazines with greater noteability and citationpower than chessbase. Creationism is noteable and has been included in many citations, but it does not verify the hypothesis. (and trust me, I'm not even saying that chessmetrics is even remotely close to valid). Philca, please accept the burden of proof. I will admit my own bias. I have acquaintances that put me in an extremely delicate position, about Mr. Sonas. However, I'd like you to realize the bias that the group of editors who have helped spread chessmetrics to 500 google hits in the subdomain en.wikipedia.org, also have a source of bias. We need to focus strictly on chessmetrics. We need to focus strictly on the policies and rules which govern the rest of wikipedia and not let a group of chess editors run their niche under a more lax set of rules. Chessmetrics shall be judged by the same standard as a hypothesis put forth by a no-name self-proclaimed scientist, self appointed expert who has enough power in his networking circle, to have it published by a hobby company like chessbase. Please own up to the burden of proof, that it is not my responsibility to defend chessmetrics. Wikipedia is straightforward, honestly run organization. This chess hypothesis should be judged fairly, without bias, and we should discuss all the pertinent issues. I'm an excellent listener, and will try to fully understand your points, and you make good points, so I'm optimistic about our future communications, and I know we'll reach accross the aisle and hopefully arrive at a better informed resolution, since we all want wikipedia's interests to exceed our personal interests, and that is why we shall be pleasant collaborators with the same goal. Sentriclecub (talk) 04:21, 16 September 2008 (UTC)[reply]

addendum 1 before we go further, can you guys add some internet links where I can read more information about Mr. Sonas's mathematical background and what qualifies him to write a formula which improves elo (by getting rid of inflation) more than if I wrote a formula? Afterall, my father is CEO of a publishing company, surely you wouldn't allow me to use his friend's or fratboy old college roommate's magazine to slip in my article for it to gain inclusion on wikipedia right? There are very few hypothesis in chess. Most of the chess articles are descriptive (such as biographies) while most of math/science articles are inherently more complicated. Chessmetrics, as it stands now, presents and reads as a hypothesis. I would be thrilled if we could skip the weeks of debate, and just downgrade chessmetrics article to being about a research and fully describe the research paper, but not to have the aura and prestige of a hypothesis. Its a minefield for me to try to come in and threaten the scope of chessmetrics on articles on the wikipedia mainspace. Just as I would fear if you wanted to rescind the theory evolution do just a few mentions on a half dozen or less pages. This discussion is inherently a very sensitive issue, and I really hope from the start, we can all set aside personal interests, and simply use the same standards that are used in the science or humanities subspace of wikipedia. I am a chess editor on wikibooks (currently developing the Richter Rauzer attack of the sicilian defense) so I do have a love of chess, I'm not trying to just post a delete tag to stir up hornet's nest. I know that for whatever reason, chessmetrics has creeped onto virtually every chess article, with zero resistance, nobody has stood up and applied editorial restrain and said "you know what, lets end this now, literally now, I mean today".

Lastly, so you know where I'm coming from. I spent a lot of time today getting permissions to fix a lot of blatant inaccuracies in a subset of wiki:finance articles. They surprised me with their willingness to not take it as an offense, that I saw dire need to fix a problem. I have been wanting to understand the risks that I must take to try and take on a drastic challenge, like cleansing the chess subspace of wikipedia, and having chessmetrics only mentioned a few times, and only in articles were it is in the article's best interests, not in Mr. Sonas and his pet hypothesis and their best interests. Sentriclecub (talk) 04:40, 16 September 2008 (UTC)[reply]

Did Elo have a PhD in statistics from an Ivy League school? Did Harkness have a PhD in statistics from an Ivy League school? Did Walt James? Did Kenneth Williams? Did William Wilcock? Bubba73 (talk), 04:52, 16 September 2008 (UTC)[reply]
Please, that is clear strawman. I'm not arguing that you have to a degree to author a notable research paper. Why do I feel like its going to be 1-vs-100 now that its just me?. Sentriclecub (talk) 05:10, 16 September 2008 (UTC)[reply]
I'm not talking about a research paper, I'm talking about designing a chess rating system. Bubba73 (talk), 12:36, 16 September 2008 (UTC)[reply]

You might want to look at the discussion at Talk:Emanuel Lasker#Using Chessmetrics for rankings. The consensus I believe was (a) Chessmetrics is the work of a single person and must never be presented as an absolute conclusion, but (b) Chessmetrics is the best (sometimes only) online resource for ranking players pre-1970. So as a guide - and as nothing more than a guide - it is useful to say, "X was ranked #4 in the world according to Chessmetrics at his peak in 19xx". Since Chessmetrics is linked, the user can then click on this article and make up their own mind as to how useful that assessment is. But it is certainly better than none.

A secondary use of Chessmetrics is that Sonas has made all past major tournaments accessible online, so it is a reference to something like, "Paul Keres came 3rd in the 19xx USSR championship".

On 3rd party users: I have seen it used on Chessbase, The Week in Chess and Mig Greengard's blog - which are IMHO 3 of the top english language web resources for chess news (certainly my personal top 3).

So there are plenty of places where chessmetrics could and should be used, and it needs to be kept. (Although I concede there are still plenty of places it is misused and needs to be cleaned up). Peter Ballard (talk) 06:22, 16 September 2008 (UTC)[reply]

p.s. I previously said Chessmetrics was mentioned by Mig Greengard in his blog. I retract that - it often comes up in his blog responses, but I can't find Mig ever writing on chessmetrics himself. Peter Ballard (talk) 00:17, 17 September 2008 (UTC)[reply]
Peter Ballard alluded to Talk:Emanuel Lasker#Using Chessmetrics for rankings, where I expressed incredulity at Chessmetrics' claim that in 1914 reigning World Champion Lasker, a veritable chess god at the peak of his powers, was #12 in the world. (That presumably made his winning of the super-strong St. Petersburg 1914 tournament shortly thereafter, ahead of Capablanca, Alekhine, Marshall, Tarrasch, Rubinstein, etc. a huge upset.) That said, I have not the slightest doubt that Chessmetrics meets Wikipedia's notability criteria. Lots of scientific or assertedly scientific theories, such as those of Sigmund Freud, have been questioned (and by persons much more qualified than I am as a statistician). That hardly makes them non-notable. Sonas' methods may not be perfect (nor are anyone else's), but they're unquestionably an important contribution to the subject. Krakatoa (talk) 07:12, 16 September 2008 (UTC)[reply]
Apparently, the chess world of Wikipedia is separate from the rest of Wikipedia. In the rest of Wikipedia, if I were to make up some formula, create a web page about it, and write an article about it, then my web page and my article would not be considered valid references to support notability. Also, in the rest of Wikipedia, removing a deletion template with the explanation "It is notable because I say it is" is not proper. It isn't difficult to understand - especially for those capable of playing chess with any skill level - third-party sources are used to show notability. Telling someone that they need third-party sources (which, in this case means anything not written by Sonas) does not mean that the topic is not notable, it simply means that the person has not shown notability. However, as I said, the chess world obviously doesn't care as long as they get their warm fuzzy feeling from twisting Wikipedia into their own little cozy corner of the web. -- kainaw 12:06, 16 September 2008 (UTC)[reply]
You've been told where the third-party sources are.
My statement "I expect an immediate apology" (for your incivility) was not a request. -- Philcha (talk) 12:59, 16 September 2008 (UTC)[reply]
Kainaw has twisted your words (you did not say "It is notable because I say it is", you asserted that it was widely used, and mentioned a paper). Kainaw is also wrong about Wikipedia procedure - a deletion template can be removed for any reason, go read WP:PROD. I vote we ignore this guy. Peter Ballard (talk) 13:09, 16 September 2008 (UTC)[reply]
I know very little about Chessmetrics, and would have to do a bit of research into the available sources before I can make up my mind about it. However, I have to respond Kainaw's allegation that "the chess world obviously doesn't care as long as they get their warm fuzzy feeling from twisting Wikipedia into their own little cozy corner of the web.". This is completely false, and a cheap stab at the very integrity of the chess writers at Wikipedia. Moreover, if you take a look at the various chess-related AFDs which have taken place will show that the chess contributors are just as likely to vote "delete" on chess-related stuff as other people are. All the work which has been done aims to meet the concepts of verifiability, and notability, and the reason we write these articles is not to "twist Wikipedia into our own little cozy corner", but to provide information which we believe to be of benefit to readers who are interested in chess, whether it's the game, the players, the competition, or the history. To suggest we are perverting the intent of Wikipedia by writing this type of material is totally missing the mark. Sjakkalle (Check!) 06:07, 17 September 2008 (UTC)[reply]
Obviously so, considering Peter's good work below. All the discussion above should not be necessary. Once the article was tagged for deletion as being non-notable, Philcha should have added one or two references to the article to show notability and commented on it in the deletion discussion. Then, others would have popped in and said "KEEP - Philcha did a great job of showing notability." Then, the deletion would have ended in a keep and everyone would be happy. -- kainaw 12:39, 17 September 2008 (UTC)[reply]
You could have read the article and the cited literature. Then I might be thanking you rather than Peter Ballard for pointing out the difficulty, and asking you rather than Peter Ballard to comment on the resolution I've proposed. In that case all the discussion above would have been unnecessary.
However your inital post showed no sign of your having read anything and continued, "My impression is that Philcha is closely related to the author of the page (if not the author) and is attempting to bypass the deletion process for a page he knows will not pass any scrutiny. So, unless Philcha can provide some citations from someone who did not create the chessmetrics process, I suggest it be nominated for speedy deletion, not just non-notable deletion. It is very simple, refusal to follow the rules indicates that your article is less notable, not more notable." Peter Ballard agreed with me that you were uncivil and should apologise. Even Sentriclecub, who added the "notability" tag in the first place, wrote, "Kainaw's response was a little bit heated".
Instead you wrote in a later post, "However, as I said, the chess world obviously doesn't care as long as they get their warm fuzzy feeling from twisting Wikipedia into their own little cozy corner of the web," which Sjakkalle and I both regard as a smear agaisnt Wikiproject Chess. -- Philcha (talk) 14:01, 17 September 2008 (UTC)[reply]

Use this Section for the Mr. Sonas links, which I requested, so that everyone can read them

[edit]
  • Noted chess author John L. Watson mentions chessmetrics - http://www.chesscenter.com/twic/jwatsonbkrev82.html . Interesting (for me) is that he does not advocate it, he simply assumes his readers know about it: "I was surprised to find from Chessmetrics that he (Alexei Dreev) was in the top ten players in the world for about a year (and briefly #9), and he stayed within the top 12 for almost two and a half years."

Comments

[edit]

(I've split this section out, so the above section is purely a list of links, which was Sentriclecub's original intention). Peter Ballard (talk) 01:24, 17 September 2008 (UTC)[reply]

Thanks Bubba73, for that reference at the bottom of the article. He is a statistician afterall, and I can breathe a sigh of relief. Sentriclecub (talk) 05:44, 16 September 2008 (UTC)[reply]

As you know, Elo wasn't a statistician or mathematician. Neither was Harkness. Bubba73 (talk), 00:37, 17 September 2008 (UTC)[reply]
Thanks Peter. I am very happy to see that someone understands notability here. A couple of those references, as you say, simply mention chessmetrics (I can't tell if they are mentioning the website or the program). But, there is the ability to use them as references on the article itself to demonstrate notability in the article, not just on the discussion page. Side note... I did a google search and I find Sonas' marketing speak rather self-centered: "Chessmetrics is the most popular chess tool used by all chess professionals and most amateurs!" It makes me wonder if he has anything to back his own claims. -- kainaw 02:05, 17 September 2008 (UTC)[reply]
I've added the references to the article. I agree that Sonas is quite self-promoting, but we should be able to look past that, and describe and use Chessmetrics on its merits. Peter Ballard (talk) 05:23, 17 September 2008 (UTC)[reply]
Great! Now notability is met within the article. I mention Sonas' self-promoting because he uses phrases such as "the most popular" and "used by all professionals". If he has something to back those claims, they could be used here to note that this is, in fact, the most popular used by all professionals. -- kainaw 12:28, 17 September 2008 (UTC)[reply]

Missing Quote

[edit]

The article said (until I deleted it):

In 2006 economists Charles C. Moul and John V. C. Nye used Chessmetrics to determine the "expected" results of games, and wrote, "This specification (Chessmetrics) has been optimized for predictive power for games between world-class players. ... there are three primary differences between these Sonas ratings and standard Elo-based ratings. Besides a linear framework that appears to dominate in predictive power the more common logistic specification, ..." (ref name="Moul_and_Nye">Charles C. Moul and John V. C. Nye (May 2006). "Did the Soviets Collude? A Statistical Analysis of Championship Chess 1940-64". The Social Science Research Network. Retrieved 2008-07-08. Full article freely available via links on the cited web page./ref)

I've searched the paper and I can't find that quote, or anything resembling it. Am I blind? What page is the quote on? Until this is resolved, I'll remove that section. Peter Ballard (talk) 12:32, 17 September 2008 (UTC)[reply]

Thanks, Peter B. There appear to be different versions in circulation. This one contains the quoted words. It's now hard to tell which is the "authorised version". I'd be inclined to go for the mirror labelled "SSRN" at the cited web page. This version makes the following points:
  • "Ratings in chess that make use of rigorous statistics to produce good estimates of relative player strength are now relatively common, but comparing ratings across different time periods is often complicated by idiosyncratic changes (cf. Elo, 1968 for the pioneering discussion). Sonas uses the same rating formula throughout our sample and updates this rating monthly instead of annually, as is more common"
  • ""Moreover, retrospective grading allows him to establish rankings that are unbiased estimates of the “true” relative strengths of players"
which I think are a more than adequate endorsement. What do you think? -- Philcha (talk) 13:45, 17 September 2008 (UTC)[reply]
Replacement quote added. -- Philcha (talk) 23:01, 18 September 2008 (UTC)[reply]

Lasker 1914

[edit]

In "Perceived disadvantages", he did drop to #12 while inacvtive 1910-1913, but in 1914 when we returned, he immediately went back to #1. Should the article mention that? (In fact, he dropped more than once, but went back to #1.) Bubba73 (talk), 03:15, 19 September 2008 (UTC)[reply]

I wanted to put something in to counter the rather extensive list of supposed advantages of Chessmetrics, but I'll be honest, what I put in was straight WP:OR from a Wikipedia talk page. The truth is there is not a lot of 3rd party discussion of chessmetrics in WP:Reliable Sources. I'm wondering whether it might be better to cut back on the "advantages" section; simply link to Chessmetrics and Sonas articles and let Sonas speak for himself. We can do this by splitting the "differences" from the "perceived advantages". As far as I can see, Chessmetrics differs from Elo in two main ways: it is more dynamic (ratings change more quickly), and rewards activity / penalised inactivity. Those two are not in dispute, and it is a matter of opinion whether these differences are good or bad. Then there are the claimed advantages, which are claimed by Sonas but not verified by 3rd party sources (except partly by one sentence in a single paper): that Chessmetrics is predictive, and it does not have a built in bias. We should make it clearer that these advantages are simply what Sonas claims. I think if we separated it properly, and talked less about how good Chessmetrics is - instead leaving that for the links to the Sonas articles - then we can remove the "perceived disadvantages" section which is, I freely admit, WP:OR. Peter Ballard (talk) 04:04, 19 September 2008 (UTC)[reply]
Or even better don't discuss Sonas' claimed advantages at all. Simply mention the differences I list above. Peter Ballard (talk) 05:00, 19 September 2008 (UTC)[reply]
We can't leave the article mute on Chessmetrics' pros and cons, otherwise someone who hasn't been through this debate will add material, perhaps incautiously. As Peter Ballard says, it's hard to find neutral but WP:RS comment. I suggest sections "Sonas' claims" and "Comments by others".
Re Lasker's drop to #14, that's consistent with what Sonas says. Steinitz played only 1 competitive event between 1873 and 1882, his 7-0 over Blackburne in 1876, and drops right off the radar in th elate 1870s. Both cases appear consistent with what Sonas says. -- Philcha (talk) 06:41, 19 September 2008 (UTC)[reply]

-- Philcha (talk) 06:41, 19 September 2008 (UTC)[reply]

Original Research

[edit]

Forgive me for re-opening this arguement - but has Chessmetrics yet had any official acceptance by a statistical body? Has it been accepted in the form of an academic paper? I do not count the paper used by the two economists, 1 because they are economists, and 2 because they did not compare the results given by given by FIDE ratings (at least as far as I can see)--ZincBelief (talk) 16:24, 9 March 2009 (UTC)[reply]

Their being economists is no obstacle - formal analyses of games are often done by economists, see Economics#Game_theory. They found Chessmetrics ratings (which they often refer to as "Sonas ratings") so valuable that they only considered games where both players had Chessmetrics ratings, and explain why they found these ratings useful. To turn your questions round, which alternatives receive better independent support? --Philcha (talk) 18:15, 9 March 2009 (UTC)[reply]
Elo was a physicist. Harkness also devised a rating system, but he wasn't a mathmetician either. These people did it by hand. Sonas had a big database and computers to help him out. Bubba73 (talk), 19:12, 9 March 2009 (UTC)[reply]
Neither point answers the key issue - unless this is independently validated it is original research. Who is to say that they didn't use Sonas Ratings purely because it supported their case. Any serious academic paper on the subject, and trustable academic paper on the subject would have used both methods. Otherwise it is simply flirting with an untried theory, an unproved theory. I do not trust the rigour of the approach used by these two economists. It is not a scientific approach. --ZincBelief (talk) 10:39, 10 March 2009 (UTC)[reply]
Under WP:V and WP:NOR your opinions of the 2 economists' reasons do not matter at all. If you can find another independent and reliable source that presents a different view, that would probably be well worth adding. --Philcha (talk) 12:26, 10 March 2009 (UTC)[reply]
The problem is actually that there is no paper that proves chessmetrics as a reliable and statistically valid rating scheme. That I point out the flaws in the paper quoted is of course my own 'research', normally known as common sense. I do not rule out that chessmetrics can be valid, I am simply saying that it has not shown to be valid. An economics paper using the rating scheme does not validate it. It is patently not the relevant field.--ZincBelief (talk) 13:23, 10 March 2009 (UTC)[reply]
I don't know of any independent validation of Elo's method, the method by Harkness, or the one used in Great Brittan. The chessmetrics website shows how his method is more accurate than Elo's. That stands to reason. Elo made some assumptions about the distribution. Sonas had a great deal of information (much more than Elo) and he had a computer to test various methods. It is clear to me that he came up with a more accurate system. Bubba73 (talk), 14:57, 10 March 2009 (UTC)[reply]
And as far as your point about original research, that applies to work done by editors. Bubba73 (talk), 15:12, 10 March 2009 (UTC)[reply]

2 or 4 years?

[edit]

I think the system gives a zero weight only to games 4+ years old, not 2 years. 93.96.236.8 (talk) 12:43, 11 February 2010 (UTC)[reply]

"Perceived disadvantages" John Nunn Claim Citation Needed

[edit]

I can provide a link to establish the John Nunn claim that needs citation here: http://chess.co.uk/twic/jwatsonbkrev82.html

I'll leave it so someone who knows what they are doing to make the citation edits to the actual article 75.166.58.124 (talk) 03:14, 18 February 2010 (UTC).[reply]

That just seems to have one line saying that he is surprized that ChessMetrics has Alexey Dreev in the top 10 for a year or so. Bubba73 (You talkin' to me?), 04:03, 18 February 2010 (UTC)[reply]
Bubba73 - look about nine tenths of the way down the page and you will find this paragraph:
"Returning then to the question as to how Süchting scored 11.5 points, the answer is simply that the other players were not much better. If we assume Süchting as 2100, then his score implies an average rating for the tournament of 2129 - it would not even be assigned a category today. Based on the above, readers will not be surprised when I say that my general impression of the play at Karlsbad was quite poor, but the main flaws did not show cup in the areas I expected..." 75.166.58.124 (talk) 04:14, 18 February 2010 (UTC)[reply]
Actually this is really the one:
"In order to be more specific about Karlsbad, take one player: Hugo Süchting (1874-1916). At Karlsbad he scored 11.5/13.5 or 'minus 2', as they say these days - a perfectly respectable score. Having played over all his games at Karlsbad I think that I can confidently state that his playing strength was not greater than Elo 2100 (BCF 187) - and that was on a good day and with a following wind. Here are a couple of examples of his play:"... 75.166.58.124 (talk) 04:17, 18 February 2010 (UTC)[reply]
 Done OK, I didn't see that. Bubba73 (You talkin' to me?), 04:19, 18 February 2010 (UTC)[reply]
I see that :-). Thanks and cheers. 75.166.58.124 (talk) 04:31, 18 February 2010 (UTC)[reply]

No Updates

[edit]

The Chessmetrics site has not been updated since 2005. That's important enough to be mentioned in the main article. Abenr (talk) 15:35, 26 November 2013 (UTC)[reply]

[edit]

Hello fellow Wikipedians,

I have just modified 2 external links on Chessmetrics. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 12:58, 21 November 2016 (UTC)[reply]

[edit]

Hello fellow Wikipedians,

I have just modified one external link on Chessmetrics. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 13:20, 4 August 2017 (UTC)[reply]

Moule and Nye and the reliability of chessmetrics

[edit]

It's quite a few years late for this, but I just read some interesting commentary by Mark Weeks on the reliability of chessmetrics and the Moule and Nye study mentioned prominently and repeatedly on this page: http://worldchesschampionship.blogspot.com/2010/02/calculating-collusion.html. In short Weeks argues that the Moule and Nye study is flawed precisely because it relies on dubious chessmetrics ratings. Weeks points out a few oddities in the chessmetrics ratings for Zurich 1953. Chessmetrics ranks Reshevsky 16 points higher than Smyslov. Reshevsky was an all-time great, but Smyslov would become WC and Reshevsky was not stronger than Smyslov in 1953. Also Chessmtrics ranks Stahlberg higher than Keres and within 2 points of Bronstein and Boleslavksy, also absurd. No analysis based on those numbers can be considered sound. Weeks also considers what flaws in chessmetrics might produce these odd ratings. Quale (talk) 07:25, 11 January 2018 (UTC)[reply]

A blog is not a WP:RS in any case. 107.190.33.254 (talk) 20:37, 22 June 2019 (UTC)[reply]