Life In 19x19
http://www.lifein19x19.com/

New EGC Rules
http://www.lifein19x19.com/viewtopic.php?f=45&t=2632
Page 1 of 4

Author:  RobertJasiek [ Sun Dec 19, 2010 7:01 am ]
Post subject:  New EGC Rules

The EGF has just changed the European Go Championship's tournament rules. They are at

http://www.eurogofed.org/egf/ecrules.htm

and mirrored at

http://home.snafu.de/jasiek/EuropeanCha ... Rules.html

with changes marked here

http://home.snafu.de/jasiek/EuropeanCha ... anges.html

and the old version for reference still being available here

http://www.eurogofed.org/egf/ecrulesold.htm

***

The tournament determines the European Champion and possibly also a European Open Champion. The European Champion is a European. From now on, a European Open Champion will always be a non-European (but the title might not be given if all non-Europeans get a smaller MMS than the McMahon's top Europeans).

After McMahon round 7, top Europeans enter a different title path: 8 Players shall play a KO for the European Champion title. In between McMahon and KO, there is a seeding as follows: Consider the top 16 Europeans. Top MMS players among them qualify, bottom MMS players drop out, intermediate MMS players have to proof their right to qualify in relegation games. This allows for a very meaningful seeding: All qualified players must have a greater MMS than all not qualified players.

Europeans dropping out in relegation games or the KO quarter finals re-enter the McMahon. Thereby non-Europeans still have enough interesting European opponents during the last three rounds.

Well known pairing strategies are introduced to reduce the impact of otherwise too great strength differences for some of the top players. It remains to be seen though how well pairings programs will be able to implement that. Another aspect has already been attacked: Minimization of repeated pairings in relegation or KO games.

Author:  willemien [ Sun Dec 19, 2010 10:53 am ]
Post subject:  Re: New EGC Rules

:clap: :clap: Good work :clap: :clap:

Although i found there is a strange regulation:

If a player only reads Spanish, French,Romanian (or another european - non-english language) he must sign :shock: a Japanese Chinese or other Asian language form (according to the registration rules)

I think the supergroup could be bigger (you have now 7 rounds to reduce the field from 32 to 16 participants)

The rules speak of "re-enters the McMahon tournament " while i think it should be "continues the McMahon tournament "

It should be noted that the results of the "Qualification for the Knockout" round do NOT count for the McMahon tournament.
And that the results in the "Knockout" DO Count for the McMahon tournament.

maybe you should add a 13th tiebreaker in the list for the Open european champion "lotery by nigiri" (the 12th tiebreaker is not always usable)

But still

:clap: :clap: GOOD WORK :clap: :clap:

ADDED LATER

There is something strange with the new Miscellaneous rule

Quote:
# Especially in top groups, the recommended pairing strategy is Cross Pairing (scheme 1-3, 2-4) by McMahonScore - rating in rounds 1 and 2 and Fold Pairing (scheme 1-4, 2-3) by McMahonScore - SOS in all later rounds.


Is this really what you want?
Normaly pairings are done within a McMahonScore group, do you want to change this to a system where you don't do that?

Author:  RobertJasiek [ Sun Dec 19, 2010 11:34 am ]
Post subject:  Re: New EGC Rules

willemien wrote:
:clap: :clap: Good work :clap: :clap:


Many years of thinking and discussion finally have paid off! :)

Quote:
If a player only reads Spanish, French,Romanian (or another european - non-english language) he must sign :shock: a Japanese Chinese or other Asian language form (according to the registration rules)


I guess if you read the text a hundred times more carefully, then you can find as many more :shocks: :)

Quote:
I think the supergroup could be bigger (you have now 7 rounds to reduce the field from 32 to 16 participants)


No. It is a reduction from 24~26 Europeans (32 means that the excess is non-European) to 8 KO players with an intermediate filter of size 8~16. I guess we should await practice to see for sure how well these parameters will work out.

Quote:
The rules speak of "re-enters the McMahon tournament " while i think it should be "continues the McMahon tournament "


It amounts to the same, doesn't it?

Quote:
It should be noted that the results of the "Qualification for the Knockout" round do NOT count for the McMahon tournament.
And that the results in the "Knockout" DO Count for the McMahon tournament.


Write "relegation" instead of "Qualification for the Knockout". (Rounds 1 to 7 are also part of that qualification.) Then distinguish between the KO semi-final players themselves and their McMahon opponents. Afterwards you have understood it correctly. (I think that this solution is a pragmatic way of keeping the McMahon for Open-EC as well as possible. After all, the AGM has wanted us to keep the McMahon rather than creating a second KO also for the Open-EC. - I could live also with a tied Open-EC title while not using any final placement tiebreakers but the EGF Committee is fond of cutting down the number of Open-EC title holders to one by (almost) all means.)

Quote:
maybe you should add a 13th tiebreaker in the list for the Open european champion "lotery by nigiri" (the 12th tiebreaker is not always usable)


Very clearly I did make a proposal like that, of course, but Matti pointed out that it would not have agreed to the EGF Committee's intention.

Quote:
Is this really what you want?


Yes.

Quote:
Normaly pairings are done within a McMahonScore group, do you want to change this to a system where you don't do that?


No need to worry. You just misunderstand. XY-Pairing by MMS - Tiebreaker is supposed to mean to pair within each MMS group. Ok, I get your point: The text could be more explicit about its intention.

Author:  hyperpape [ Sun Dec 19, 2010 8:08 pm ]
Post subject:  Re: New EGC Rules

Robert Jasiek wrote:
Europeans dropping out in relegation games or the KO quarter finals re-enter the McMahon. Thereby non-Europeans still have enough interesting European opponents during the last three rounds.
Why only after relegation and quarterfinals? Will the semifinals not conclude until the 10th round of the open has finished?

Author:  RobertJasiek [ Mon Dec 20, 2010 1:34 am ]
Post subject:  Re: New EGC Rules

The EGF thinks that places 3 and 4 are determined better in the KO context and that this is worth more than sending back 2 more Europeans to the McMahon for its round 10.

Author:  HermanHiddema [ Mon Dec 20, 2010 1:42 am ]
Post subject:  Re: New EGC Rules

RobertJasiek wrote:
The EGF thinks that places 3 and 4 are determined better in the KO context and that this is worth more than sending back 2 more Europeans to the McMahon for its round 10.


Which is nonsense, of course. KO is no good for determining 4 places.

Author:  RobertJasiek [ Mon Dec 20, 2010 3:22 am ]
Post subject:  Re: New EGC Rules

We do not claim that KO would be the most meaningful theoretically possible way to sort places 1 to 4 but that places 3 and 4 are more meaningfully decided in the KO-3rd-place-game than by sending back the semi-final losers to the McMahon and then using SOS-x tiebreakers for that purpose. (Place 2 is a related issue.)

OART, it is unclear how to determine places 2+ the most meaningfully for a given player field. E.g., one might hold iterative tournaments: First determine place 1 and drop out that player, next determine place 2 (winner of second stage) and drop out also that player, etc. However, one might well argue that such a system would also not be the fairest possible for places 2+. Many other systems could be imagined. E.g., when designing a system for German Championship places 1 and 2, we had already about 10 different proposals, one more convincing than the other;)

Author:  topazg [ Mon Dec 20, 2010 3:42 am ]
Post subject:  Re: New EGC Rules

What about:

A vs B = A wins
C vs D = C wins

1st/2nd place -> A vs C = A wins
3rd/4th place -> B vs D = D wins

A -> C -> B -> D

But as B and C never played, should there now be a 2nd place play off? Could B rightly feel aggrieved at getting A in the semi final and not the final if he's the 2nd strongest player?

Author:  HermanHiddema [ Mon Dec 20, 2010 3:47 am ]
Post subject:  Re: New EGC Rules

RobertJasiek wrote:
We do not claim that KO would be the most meaningful theoretically possible way to sort places 1 to 4 but that places 3 and 4 are more meaningfully decided in the KO-3rd-place-game than by sending back the semi-final losers to the McMahon and then using SOS-x tiebreakers for that purpose. (Place 2 is a related issue.)

OART, it is unclear how to determine places 2+ the most meaningfully for a given player field. E.g., one might hold iterative tournaments: First determine place 1 and drop out that player, next determine place 2 (winner of second stage) and drop out also that player, etc. However, one might well argue that such a system would also not be the fairest possible for places 2+. Many other systems could be imagined. E.g., when designing a system for German Championship places 1 and 2, we had already about 10 different proposals, one more convincing than the other;)


I don't agree. The problem with KO, as I'm sure you're aware, is that any player who lost to the winner can realistically be the second strongest player of the field, regardless of whether they lost in the first, second or third round. Therefore, KO only meaningfully decides the best player. That player has either beaten all other players, or they have been beaten by players that he has beaten, etc. Even the loser of the final might realistically be weaker than the winner's first & second round opponents, etc. So even place two is not reliably determined.

With a 3rd/4th place game, the player who gets place 4 will, from the start of the KO, have won one game (the first) and lost two. It is very possible that there are other players who have scored two wins after losing in the first KO round. I don't think there is value in saying that the no 4. player deserves a better placement. Winning the first round of the KO should not automatically qualify you for a better placement than the losers. KO divides the field into two groups: 1 winner, 7 losers. That's why it's called KO.

I think there is more value in simply having one European Champion and ranking all the other players in the Open according to tie breakers. The KO winners is EC, the player with best MMS/SOS/Whatever in the Open is the European Open Champion (possibly, this is the same player).

For prize money, I think you should always split equally among players with equal MMS.

Author:  RobertJasiek [ Mon Dec 20, 2010 3:56 am ]
Post subject:  Re: New EGC Rules

HermanHiddema wrote:
The problem with KO, as I'm sure you're aware, is that any player who lost to the winner can realistically be the second strongest player of the field, regardless of whether they lost in the first, second or third round. Therefore, KO only meaningfully decides the best player. That player has either beaten all other players, or they have been beaten by players that he has beaten, etc. Even the loser of the final might realistically be weaker than the winner's first & second round opponents, etc. So even place two is not reliably determined.

With a 3rd/4th place game, the player who gets place 4 will, from the start of the KO, have won one game (the first) and lost two. It is very possible that there are other players who have scored two wins after losing in the first KO round. I don't think there is value in saying that the no 4. player deserves a better placement. Winning the first round of the KO should not automatically qualify you for a better placement than the losers. KO divides the field into two groups: 1 winner, 7 losers. That's why it's called KO.

I think there is more value in simply having one European Champion and ranking all the other players in the Open according to tie breakers. The KO winners is EC, the player with best MMS/SOS/Whatever in the Open is the European Open Champion (possibly, this is the same player).


Yours is a reasonable alternative view but we won't come to an agreement about places 2+ soon. (For reference, players dropping out of a KO then play against a weaker European players field than the one still in the KO, so one cannot simply compare numbers of wins 1:1.)

***

topazg, many systems could be imagined. Not all of them fit into the EGC schedule of 1 game per day up to the last Saturday.

Author:  HermanHiddema [ Mon Dec 20, 2010 4:04 am ]
Post subject:  Re: New EGC Rules

RobertJasiek wrote:
Yours is a reasonable alternative view but we won't come to an agreement about places 2+ soon. (For reference, players dropping out of a KO then play against a weaker European players field than the one still in the KO, so one cannot simply compare numbers of wins 1:1.)


I don't think it is that important anyway, the important thing is to have a single winner, a single champion from a single final game. Playing for 3rd/4th place is done often enough (e.g. in the football world championship) that there is plenty of precedent to justify it, even if it is not really mathematically sound. One thing I do think would be better is to play for place 3 only, and rank the loser of the 3rd/4th match by MMS/SOS. Places 1/2/3 are the traditional "important" places (gold/silver/bronze), and the players in place 2/3 have won at least as many games as any other KO losers, so there isn't much ground for complaints from them.

Author:  gaius [ Mon Dec 20, 2010 8:58 am ]
Post subject:  Re: New EGC Rules

After all the discussion that has taken place in the past years, I think that this system will work optimally for everyone involved. Many congratulations to the designers :clap:.

Author:  willemien [ Sat Jan 01, 2011 4:40 pm ]
Post subject:  Re: New EGC Rules

topazg wrote:
What about:

A vs B = A wins
C vs D = C wins

1st/2nd place -> A vs C = A wins
3rd/4th place -> B vs D = D wins

A -> C -> B -> D

But as B and C never played, should there now be a 2nd place play off? Could B rightly feel aggrieved at getting A in the semi final and not the final if he's the 2nd strongest player?


Was thinking about this .
and it is even worse than you mention:

It is possible that B won against C outside the KO rounds (and because they did play eachother before the KO rounds they are not paired against each other inside the KO rounds)

Author:  Javaness [ Sun Feb 13, 2011 5:03 am ]
Post subject:  Re: New EGC Rules

This week I started to rewrite these rules, However I was told that it was unimportant to the rules committee that they should make sense, so I stopped.

https://docs.google.com/document/pub?id ... Vl7r_TnZFw

Author:  RobertJasiek [ Sun Feb 13, 2011 6:12 am ]
Post subject:  Re: New EGC Rules

You must be Ian Davis then. - The Rules Commission thinks that there is too little benefit from endless linguistic revisions. We collect suggestions so that they can be considered in a next update of contents but, since the latest update is rather recent, time is not urgent for resumed fine print discussion of language details.

I will try to answer your questions though:

"There is no such thing as a European rank though?"

Correct. However, informally it makes very much sense to speak of European ranks in contrast to various non-European or online ranking systems. The purpose of the related rule is that players from other continents should translate their ranks to European standard so that there will not be (too many) accidental sandbaggers. E.g., a common problem is the proverbial unaware Japanese using his Japanese rank as if it were a European rank and then scoring an impressive number of 10 losses.

"What is ‘lower rating consideration bar’?"

It is the rank declared by the Tournament Supervisors that is at least 4d but might be set higher. Only players with at least that rank X will be considered as possibly eligible for the supergroup. E.g., if we have 100 players 5d+ and 100+ players of rank 4d, then we would set X to (at least) 5d so as to be able to determine the supergroup players within a reasonable amount of time instead of delaying the start of round 1 by hours.

(to be continued)

Author:  Javaness [ Sun Feb 13, 2011 6:27 am ]
Post subject:  Re: New EGC Rules

If you create a rules document for a continental championship it should

1: reflect the opinion of the AGM concerning its direction
2: be well written, so as to present a good impression of the organisation both to its potential and existing sponsors, and to its members
3: not contain incomprehensible phrases
4: not contain ambiguous phrases
5: be open to corrections

Author:  RobertJasiek [ Sun Feb 13, 2011 6:45 am ]
Post subject:  Re: New EGC Rules

The current (i.e. the new) rules have already been discussed and adopted by the EGF Committee and the Rules Commission. Every new cycle of linguistic changes amounts to several days or weeks of work plus weeks to months of adoption. There are various styles of how to formulate a ruleset and every two persons tend to have at least two different preferences as to which the best style is.

So although I more or less agree with your criteria (and before adoption had suggested in vain to offer the text for further discussion here), now that the text is adopted, we should bear its insufficient wording.

Rewriting it as fluent common English might easily increase the text length by 50%. So I am not sure whether, as you suggest, sponsors would read such a long text at all. More appropriately, for them or others a short summary could be written. Such need not be a rules text and could therefore be written by anybody incl. you (ASA you will have well understood every aspect of the contents).

Author:  tapir [ Sun Feb 13, 2011 7:04 am ]
Post subject:  Re: New EGC Rules

HermanHiddema wrote:
RobertJasiek wrote:
Yours is a reasonable alternative view but we won't come to an agreement about places 2+ soon. (For reference, players dropping out of a KO then play against a weaker European players field than the one still in the KO, so one cannot simply compare numbers of wins 1:1.)


I don't think it is that important anyway, the important thing is to have a single winner, a single champion from a single final game. Playing for 3rd/4th place is done often enough (e.g. in the football world championship) that there is plenty of precedent to justify it, even if it is not really mathematically sound. One thing I do think would be better is to play for place 3 only, and rank the loser of the 3rd/4th match by MMS/SOS. Places 1/2/3 are the traditional "important" places (gold/silver/bronze), and the players in place 2/3 have won at least as many games as any other KO losers, so there isn't much ground for complaints from them.


It isn't about mathematics, it is about the bronze medal! If you establish a new world record in the 100m prelimaries, but finish fourth in the finals, you won't get a medal. Bad luck.

Competitions are about drama not about mathematics (only the math olympiads are about mathematics after all).

Author:  RobertJasiek [ Sun Feb 13, 2011 7:12 am ]
Post subject:  Re: New EGC Rules

(answers to questions continued)

"What about Europeans who cannot read English?"

The tournament organizers should provide translators and organizers / supervisors / Rules Commission should provide translations of texts or summaries. In practice finding verbal translators is easier than for documents. OTOH, the Japanese tend to organize their own translations and summaries of crucial information.

"Exactly how is strength distinct from rating or rank?"

Rank: kyu, dan, pro-dan.

Rating: numbers due to rating system

Strength: The true strength of a player while his rank or rating might or might not represent the strength correctly.

Strength is the ideal that rank or rating should represent correctly, but they can fail.

"The pairing has already been made, why make it again?"

There are these stages for making a pairing for the relegation games:
1. tentative pairing by MMS - SOS - rating - lottery, players are numbered from 1 to 16 accordingly.
2. filter the middle of this group of players to enter the relegation stage
3. these players are now paired while abiding the principles of a) avoid repeated pairings, b) apporach fold pairing.

Now your question might refer to either "Why proceed to (2.) at all?" or "Why apply extra principles during (3.)?" The latter is obvious to you, I suppose. So presumably you want to know the former.

Here are some reasons:
- Qualifying (among high MMS players) by greater number of Wins is a much more meaningful seeding criterion than SOS - rating - lottery tiebreaking.
- Unsportsmanlike SOS manipulation (like reported from Grenoble EGC) becomes essentially impossible.
- If there is a high percentage of strong non-Europeans, then after 7 rounds McMahon the lower end seeding candidates might have so few Wins (like only 4) that qualification for some might be too easy; e.g., 4 wins in 7 rounds is not a convincing seeding. By having to play a relegation game, the most doubtful candidates will have to show that they can make 1 more win before they will actually qualify.
- Similarly, too few wins might be too few especially when compared with the top European(s) after 7 rounds, who might have 6 or 7 wins. E.g., if a 4 or 5 wins player has to play a relegation game to then have 5 or 6 wins, then the difference of wins among the qualified players will be much more acceptable and a Eu Champion with fewer wins than the finally 2nd European will be much less likely.

Author:  RobertJasiek [ Sun Feb 13, 2011 7:21 am ]
Post subject:  Re: New EGC Rules

(continued)

"Do the top players want to be tired out like this?"

The EGF Committee voted on regular thinking time for the relegation games, so it does not consider this to be a serious problem. Players among the potential top 8 (the KO participants) preferring rest should play better during the first 7 McMahon rounds. It is like in other sports: Play perfectly and you qualify without problems, play suboptiomally and you might need to enter relegation. Players surviving relegation surely will be happy to enter the KO much more than worrying about being exhausted. After congress, it is early enough to take a year of rest;)

"Why have this strange rule?"

You seem to refer to

"# To become the European Open Champion, a player has to be non-European and in the group of one or more players with the highest MMS. If that group has more than one non-European, then the European Open Champion is determined among those non-Europeans by the tiebreakers mentioned below. If that group does not have any non-European, then there is no European Open Champion. Europeans do not become the European Open Champion."

The idea is, to exaggerate a little, that a non-European 3d does not become the EOC amidst a field of much stronger 5th and higher placed Europeans just because in the current year stronger non-Europeans do not attend. The title EOC must be worth something; reaching the top MMS groups is the worthy criterion to prove it.

"What is Direct Comparison? Why not just say that for two players, one player’s Mutual Game Score is the player’s own result from a game with the other. That way you have a simple explanation people can understand."

Language issue.

Page 1 of 4 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/