Currently regular 5-dan players on EGC can enter the supergroup, so the difference in strength between top and weak members of the super group is 3-5 stones handicap.
I propose to make the super group smaller and allow only best 16 Europeans to take part (instead of 24)
In this case only super strong 5-dans and 6-dan+ players will be able to stay in the supergroup.

I feel, that only players who have chances to win EGC must stay in the supergroup. EGC supergroup is for deciding the European Champion! It's not for getting your teaching games for free!

_________________
Go lessons: http://breakfast.go4go.net/
Russian Go news: https://vk.com/go_secrets/

I think that everybody who has have chances to become the European Champion or win the EGC tournament has a right to be in the supergroup.

I think a bar on a level of 5 dan is a good idea, but am not so in favour or just bluntly limiting the number of players.

The problem I think is not so much the number of players. The problem is the inference that strong non european players have on the pairing (and therefore on the outcome). (is the European player who just lost 2 games against Koreans weaker than an other european who just won two games against other europeans)

I think my idea to have an 7 round swiss "hidden" in the main tournament (meaning that every european will have 7 games against other europeans from the topgroup and 3 against non europeans, while the last 3 have no influence on the pairing or the european result)

Ps maybe 8 rounds against europeans and 2 against others is a better idea but in any case all european candidates play an even number of games against other european candidates.

But are you not blaming other for what you want yourself:

_________________
Promotor and Librarian of Sensei's Library

breakfast, you warm up side discussions again, so I need to state my differing opinions on those side aspects again:

- The current system is not "24+ Europeans" but "ca. 24+ Europeans". I.e., it could be less.
- You suggest that there is a handicap difference of 3 to 5 stones between top supergroup members and 5d's. This is wrong for two reasons: a) In years with 7d top players, then average nominal handicap difference would be 2; so maybe you could say about 1 to 3 handicap stones. b) In years with 6d top players, 5d are close to those 6d. To repeat, online server byoyomi only games mean nothing for real world long time games.
- Unlike your propaganda, the typical 5d does not play in the EGC to get teaching games but to try winning the title or else getting the best possible result. Stop guessing wildly of what the 5d think!

***

Elsewhere you have said that you dislike tiebreakers determining the final results ordering for the top places. I see nothing related to this in your proposal. Why?

Actually, it is "a supergroup of up to ca 32 player, with up to ca 8 Asian". Having a super group of size 16 or 24 is completely within the rules. But the reality is that it never happens.



Alex did say "currently", and currently the reality is that European Professionals and Korean ex insei play in the super group. The nominal rating difference for 3-5 stones is 250-450. For weaker 5 dans, that means players in the range 2700-2900 are 3-5 handicap stones away.

There are certainly European players in the range 2700-2800 participating, and it seems realistic to admit that the strongest Koreans are probably around in the 2800+ rating range, so 3-5 stones is realistic.



Alex has stated multiple times already that he thinks making the supergroup smaller will reduce the risk of tie breakers being needed.

It will also improve their performance if they are needed.

BTW, whose decision is this, who will be tournament supervisors this year?

I honestly think there is a 0% chance of a 5d EGC winner in the next 20 years unless the participation at the top end changes (which, if it meant they could win, I would think is a very bad thing generally for European go quality). If the 5 dan players are really aiming to win the title, they are dreaming.

That's almost like me saying I intend on being the challenger for the British title.

Herman, we do not need to recur all discussion aspects all the time.

The ratings are not linear, especially not at the top ratings. There run-away-ratings can occur.

Is there any evidence that making making the supergroup smaller reduces the risk of tiebreakers being needed? Intuitively I would guess that yes, a tiny bit. But where is the poof? Also a slight decrement does not solve the problem.

The tournament supervisors are set by the Rules Commission. Currently there have not been any applicants for doing the job. Time for their education is becoming short. So it looks like this yeear the commission members will do the job. As usual, it could happen that some EGF Committee decides to watch and participate.

First comes the dream, then the reality (winning the title). It requires motivation and determination to win. It would be sad if the 5d did not have that.

There seems to be little evidence that this has occurred. In fact the ratings of top European players like Dinerchtein, Taranu, Pop and van Zeijst have been quite stable over the years, they have not run away at all, if anything they have dropped a little.



The current system is broken. There is wide support for reducing the size of the supergroup. It is a change that could have a positive effect. Why should we stick with a broken system just because another system is unproven?



The rules commission has failed, for years now, to recruit people for this task. To sit around and do nothing, and then say "there were no applicants" is a failure. I suggest the committee use their remaining time to urgently find people for this position, it is the least you can do to make up for past failures.



First comes the delusion, then it doesn't happen.

I should have expressed more clearly that with run-away ratings I also mean an improperly great gap between top rating players and the "pack".

The current system is broken, but we must not just look at one parameter while ignoring all other parameters. You assume that the supergroup size parameter is broken. Is it? I find it easier to identify other parameters as broken:

- usage of final result placement tiebreakers at all
- from player to player varying numbers of non-European super-strong opponents
- too McMahon style pairing also for the supergroup player pairings; the players can get lower MM group opponents too easily
- usage of the axiom of equal initial pairing rights; cross pairing by average rating during early rounds might be an alternative (equal title winning chances may be a nice ideal whilst, in other championship tournaments, cross pairing in initial rounds has turned out to be of helpful assistance in giving the more likely top players more equal opponents)

The supposedly ideal Europeans in supergroup size depends on several factors: number of rounds, rank / rating distribution of the upper Europeans players field, rank / rating distribution of the upper non-Europeans players field, preference of axiom (either give more players a principle chance to win the title or give fewer players this chance). The latter is a political rather than mathematical aspect. The other aspects do not allow an easy general answer. E.g., if there are one 7d, no 6d's, 25 5d's (which have pretty similar ratings) and a rating gap to the 4d's, then I would rather include all the 5d than making an arbitrary cut amidst the 5d.

The rules commission has not looked for recruiting other / further tournament supervisors because we could do the job well. The few critics criticise mainly the supergroup size because their personal preference is a smaller size. As a consequence of their preference, they want to interpret the "ca." differently: 10 to 38 Europeans instead of, as the rules commission members things, 22 to 26 (in case of emergency 20 to 28). Then the next step of different interpretation is: "I want a smaller size, so 10 to 38 actually means 10 to 16." Such is neither it the spirit of "ca." nor of the AGM's apparent opinion when the AGM acknowledged my cautious request for allowing small variation. Note that the rules do not say "very roughly 24".

Experienced tournament organizers / EGF certified referees respecting the spirit of the rules may apply to the become tournament supervisors for the EU congress 2010. Write to the rules commission if you are interested in doing ca. 10 hours of work during the congress.

if the organisers find that the top group is to big they can chance to SwissTournament accelerated pairings, that will remove weak players in the topgroup rather quickly

_________________
Promotor and Librarian of Sensei's Library

willemien, please describe accelerated pairings!

Proposal #2 is another possibility: players with too few wins drop out of the EC tournament (and enter the Open).

***

Technical question: A few days ago, I saw 3 people voting for poll option 1. Now it is 2. How is that possible? Can one take back one's vote?

You can change your vote, but you can not cancel it.

_________________
Offending ad removed

Whether or not you can do the job well is irrelevant. Your job is to find tournament supervisors, not to do it yourself. There are dozens, if not hundreds, of people in Europe that are qualified to be tournament supervisors. You have consistently failed to approach those people, and have therefor not done your job.



Note that the rule is actually "up to 32 players" for the super group. (emphasis added)



Yes, now email the above to all EGF certified referees. Job done.

In an effort to keep the conversation moving:

The system is described at http://en.wikipedia.org/wiki/Swiss-syst ... d_pairings.

In an accelerated pairing system, the top half of the players in the field (in this case supergroup) have an extra "phantom" McMahon point added to their score for the first two rounds. Players are then paired as normal, including the effect of this phantom point. After the second round is played, the phantom point is removed and pairing continues as usual.

Assuming things go roughly as ratings predict, players in the first and third quartiles should win their first games and will be paired against each other in the second round, where the top quartile players should win. On average, the system has the effect of halving the number of undefeated players after the second round.

Not sure how this affects tiebreaks, though. It seems like there might be a discontinuity with players who span the boundaries between the initial quartiles.

I do not understand this. Wouldn't that mean that being the best of the lower half gives much better chances for the first two rounds than being the worst of the upper half?

Intuitively, I would assume that the rest of the tournament would then mostly be about "healing" this result.

_________________
A good system naturally covers all corner cases without further effort.

Accelerated pairings were designed to work well when you have large field with a wide rating range, to separate out the top players more quickly. Since McMahon is also designed (and much better at) separating out the top players, combining the two is not usually very useful. In a smaller group with a relatively narrow rating range, the system indeed gives an undue advantage to the third quartile.

Accelerated pairings would be an excellent option for the WAGC, in my opinion. There, you have a larger field (~64 players) with a wide rating range (7d-7k) starting on the same level (as it is Swiss).

We did find tournament supervisors: ourselves.

Maybe you are not aware of the history of congress supervision: Basically it was done by the rules commission. The greater flexibility of allowing tournament supervisors set by the rules commission was put into the rules to allow the job being done if not enough rules commission members would be present. Now you read the rule independent of its history as something that would require the rules commission each year to spend much time looking for other persons doing the supervision. It is not like we would be required to install other persons. Rather we may do so if think that we want to do less work ourselves and if finding and educating others actually would be less work than doing the job ourselves.

Do you have the email addresses of all EGF Certified Referees? If so, please forward them to us, thanks.

Why does accelerated Swiss (in a tournament like KPMC or WAGC) provide any advantage? Why would it better than cross-pairing in the first two rounds, then fold-pairing? In which sense better?

Better in the sense that accelerated pairings allow more games between the top players to happen, and avoid games with an overly large strength difference between the players. Currently, there are many games in the first round of the WAGC that are pretty much pointless to play, because the strength difference between the players is too large. If we look at the playing field of the latest WAGC, for example, then if we order by strength and cross-pair, the top players (6-7 dan from) play against 1-3 dan players, while the weakest of the top half (4-5 dan) are playing against kyu players.

Accelerated pairings effectively eliminate the first round by pairing the first quartile against the second and the third against the fourth, then forcibly pairing the second against the third in the second round. So now you have 6-7 dan playing 4-5 dan in the first round, and 1-3 dan playing kyu players. Then in the second round the 4-5 dan are paired against the 1-3 dan, while the 6-7 dan are already playing amongst themselves, as are the kyu players.

If most of the games in the 2nd vs 3rd quartile have their expected outcome, then only about 1/8 of the players will be undefeated after the first two rounds, as opposed to 1/4 with normal Swiss. At the WAGC, that will be ~8-10 players. Those players will get to play more of their games amongst themselves, which improves the sampling for the top places. More games between the real contestants (CN, JP, KR, KP, TW, HK) will happen.

Sending out a few emails is not much work. Finding others to do the work is a good idea, regardless of historical necessity. The EGF needs all the volunteers it can get, so any little task that can be delegated should be delegated if an appropriate volunteer can be found. It is a well known problem in many volunteer organizations that certain people tend to prefer doing jobs themselves rather than finding others to help, and then when such a person quits the job, there is a much bigger problem than needed.



Why should I do your job? I assume you have a list of who is certified? Finding email addresses for many of them should not be overly hard.

Finding all the email addresses would be more work than we would want to do any time soon.