Talk:Tie-breaking in Swiss-system tournaments

Rename

If the article's only purpose is to describe chess-related tiebreaking procedures, that should be made clear in the title. I propose a move to Tiebreaks in chess or the like. (The non-hyphenated spelling is generally accepted and I prefer it.) Tiebreaker already exists but there's enough material on this page that it can stand on its own. youngvalter 03:23, 21 July 2007 (UTC)

I don't know if these systems are used other than in chess or not. I din't know about the other article. I used the hyphen because the USCF rulebook does it that way, but I'll go along without the hyphen. Bubba73 (talk), 03:27, 21 July 2007 (UTC)

"Tiebreak" (without hyphen) is not a standard word, but "tiebreaker" is. Bubba73 (talk), 03:47, 21 July 2007 (UTC)

It seems like some of these methods might be used in Go. Even so, the title as it stands now would encompass all games and sports. Maybe Tie-break systems in chess then? youngvalter 21:31, 28 July 2007 (UTC)

Probably all of these can be used in Swiss-system tournaments, so maybe it should be Swiss system tie-breaks or Swiss system tiebreakers. Some of the methods can also be used in Round Robin tournaments. Bubba73 (talk), 21:42, 28 July 2007 (UTC)

These tiebreakers are also used in Go tournaments, but the names are slightly different. I think leaving the title as it is is preferable.--ZincBelief (talk) 11:28, 25 February 2008 (UTC)

Sonneborn-Berger / Neustadtl / Gelbfuhs

The article currently states that what we commonly call the Sonneborn-Berger system was in fact invented by Oscar Gelbfuhs, citing a book by Kennth Harkness. I don't have the Harkness book, but do have the Oxford Companion to Chess (OCC). It says that Gelbfuhs' method was different to Neustadtl's (which is what we know today as Sonneborn-Berger), though it worked on the same principle. Gelbfuhs' method was designed for tournaments where participants play an unequal number of games. To quote the OCC: "Each defeated opponent's score is divided by the total number of games played by that opponent; for each opponent drawing, half the score is similarly divided. The sum of these fractions is a player's Gelbfuhs score." Of course, in cases when all players have played the same number of games (the only situation when Neustadtl can be used), the finishing order generated by Neustadtl and Gelbfuhs will be the same, but Gelbfuhs' method is clearly unnecessarily complicated for modern touraments. (The OCC asserts, incidentally, that Neustadtl made no mention of Gelbfuhs when he proposed his own system, though who knows whether this was sly appropriation or an innocent coincidence.)

I don't really have the time to fix all this up (and as I say, I don't have the Harkness book anyway, which for all I know provides absolutely certain proof that the generally reliable OCC is mistaken), so I'm just sticking this info in here in the hope it'll be of use to somebody. I might get round to it one day if nobody else does. --Camembert 17:58, 21 July 2007 (UTC)

I'm trying to sort this out and understand it. The Harkness book doesn't mention Neustadtl at all, and neither does Golombek's Encyclopedia of Chess, which I just got today. Golombek also says the S-B method came from Gelbfuhs, in 1873. BTW, I appreciate you help in this topic. Bubba73 (talk), 00:16, 22 July 2007 (UTC)

The Oxford Companion seems to be more complete than Harkness or Golombek, but I'm still a little confused. Bubba73 (talk), 00:34, 22 July 2007 (UTC)

Armageddon blitz games

"Armageddon" games (White gets more time but must win) as adopted by FIDE should be mentioned. In fact FIDE's whole tiebreaking method for the 2007 world championship could be mentioned. (it's here in Section 3.7) youngvalter 21:34, 28 July 2007 (UTC)

Kashdan

The Kashdan system is designed so that a win and a loss is better than two draws. However, it has some strange effects: three losses are better than one draw and two unplated games; five loses are better than one draw and four unplayed games! Bubba73 (talk), 22:11, 28 July 2007 (UTC)

Comment

The article as it is now is quite good, but somewhat confusing. I found a link which is imo a bit clearer, at least in structure. It also explains the differences in terminology FIDE-USCF. Perhaps this article should have a similar structure. Here is the link: [1]. If nobody objects, I might have a go at this. Voorlandt 12:21, 30 September 2007 (UTC)

I have no objection. I wrote most of the article, but I'm only familiar with the USCF and not FIDE. Bubba73 (talk), 23:38, 30 September 2007 (UTC)

Removed {Globalize/USA} tag

I removed the tag that says that the article is USA-centric, which has been there for several months. I couldn't find anything on Swiss system tie-breaks used in different countries. A book by the US Chess Federation (USCF) is used as a reference for the technical details of the tie-break systems. In addition, there is a little bit about the order in which the USCF recomends they be applied. But this isn't mandatory for a US tournament, the director can use any order he wants to (stated beforehand, of course). So if anyone has any information about the Swiss tie-breakers used outside the US, please add it. Bubba73 (talk), 00:29, 22 April 2008 (UTC)

removed material

I removed the following, for which a citation was requested.

The USCF recommends having a comprehensive list of the systems to be used in order, such as this one used by a state organization:

1. Modified Median
2. Solkoff
3. Cumulative
4. Result between players
5. Most games with the black pieces
6. Kashdan
7. Sonneborn-Berger
8. Coin flip.

I thought I had seen this too, but I couldn't find it in the USCF rulebook. Anyhow, it is not very important to the article. Bubba73 ^{You talkin' to me?} 02:22, 19 December 2010 (UTC)

FYI p. 203, TD Tip after 34E, but agreed, not very important. Badbishop (talk) 06:31, 19 February 2011 (UTC)

Good find! (I thought it was there.) If this article were specifically about chess, I think it should be included. But since it isn't, I don't think it needs to be there. Of course, I won't object if someone wants to put it in, now that it can be referenced. Bubba73 ^{You talkin' to me?} 00:11, 19 February 2011 (UTC)

Issues/omissions

How much of the following is important, and how much is nitpicking? (I've happened to have acquired a broad knowledge of tie-breaks in the past few days.)

Zeroth, if this is not specifically about chess, then remove "chess" from the first sentence. Then later explain that chess is referenced because it has well-developed methodologies for tie-breaks.

Zero-point-fifth, should the second paragraph (Golombek) be moved below, to a new "selection and ordering of tie-breaks" section that refocused the "USCF recommdended" section? This would be a place to gather the recommendations of various authors and the expectations of various groups, including usage in other sports. (Also note that Golombek probably predates adjustments to these systems, so someone would have to check the relevance of his recommendations.)

First, does this need a brief explanation of scores (0 for loss, 1/2 for draw, 1 for win, plus special circumstances for unplayed games)? Some of these tie-breaks would be invalid for other systems of scoring.

One-point-fifth, in both USCF (p. 199, 34E) and FIDE, these tie-breaks are descriptions of commonly used systems and do not have the status of rules. That seems worth discussing.

Second, both USCF and FIDE don't use the system described under Median/Modified Median/Solkoff. In fact, many online sources are incorrect. Specifically, both use opponents' adjusted scores, in which any unplayed game is re-scored as a half point.

Further distinctions: when calculating these tie-breaks, USCF and FIDE treat a player's unplayed games differently. Under USCF, a player's unplayed games are treated as an opponent with 0 score (p. 200). In FIDE, a player's unplayed games are treated as having been played against self (as a half-point tie wrt Sonneborn-Berger). This does not apply to Cumulative.

Third, what's called Cumulative is equal to FIDE's Progressive, but USCF's Cumulative = FIDE Progressive minus (sum of scores of all the player's unplayed games) (p. 200).

Fourth, should more effort should be made to better include FIDE usage of terms, and not overwhelmingly present USCF terma?

Fifth, the Median/ModMed/Solkoff structure seems odd. If it's really not chess- and USCF- centric, then it would make sense to start simplest and move on to more complex, that is, Solkoff, Median, Modified Median.

Sixth, I don't see any indication in FIDE of ever acknowledging subtracting only the highest score for Modified Median. I also can't find specifications for the FIDE implementation of Modified Median. That is, they define the different modifications, but don't define when the specific modifications are applied. Further, I'm not sure if FIDE common usage is the 50% score rule -- I haven't been able to find any relevant FIDE citation. My understanding is that colloquially, "even score" often refers to "equal wins and losses." This rule would be different from the 50% score rule. For example, win/full-point bye/loss is >50% but is equal wins and losses.

Seventh, in my reading, it's not a USCF "recommended order," it's the expected order for USCF events. That is to say, there's never a specific recommendation to use that order. Rather, they are laying out the predominant order in USCF events. In fact, the recommendation (p. 198, 34D) is to make deliberate choices based on the particular event.

Eighth, since there is the USCF recommended order, then should there also be the FIDE recommendation? They don't give a specific positive recommendation, but instead imply that USCF is philosophically wrong -- they define different categories of tie-breaks, and say that all tie-breaks for an event should belong to the same category. Thus Cumulative (cat.: own results) should never be used with ModMed (cat.: opponents' results).

Ninth, the FIDE categorization appeals to me as an organizing structure. Any comments?

Tenth, these tie-breaks are specifically for individual (not team) Swiss-System tournaments. Rename?

- - -

USCF references are Just & Berg.

FIDE source is mostly: http://www.fide.com/component/handbook/?id=127&view=article

or more generally the handbook: http://www.fide.com/handbook?option=com_handbook

Badbishop (talk) 06:25, 19 February 2011 (UTC)