European Go Congress, How to find the strongest european?

[RobertJasiek]

Your versus, e.g., my opinion on the topic have been stated frequently enough; no need to state mine again:)

[Harleqin]

I think that Javaness is right in principle. However, since the top players do not want to play an additional tournament, neither during the congress nor at a different time, a combined tournament for both the Open European Champion and the European Champion seems to be the only way to go.

We already have such a combined tournament, in which the European Champion is simply the "best" european player, but due to the fact that in recent years the European Champion is almost always determined by SOS lottery, the current system has caused some dissatisfaction.

One step that could be taken is to let the tournament organizers determine the MacMahon bar, so that they can raise it to 6 dan (or perhaps just 5 dan). This would alleviate some of the SOS pain, since the influence of the 4 dans and 5 dans on the supergroup would be much lower.

Other proposals, of which many have already been discussed, are always in severe danger of having a negative impact on the main tournament. Such negative impact can be, for example, suboptimal pairings or top asian players not getting to play the top european players.

I currently believe that, unless a separate tournament is accepted, the current system (but with a higher MacMahon bar) is the least evil.

One modification might be feasible, but it would also incur some costs: let the best european player not be the European Champion, but the challenger for that title, who will play the then-current European Champion in a best-of-3 series on three additional dates. This would give some value back to that title.

[bass]

The only argument against moving the championship outside of the two weeks is some false sense of historical integrity. That's my decision and I'm sticking with it.

That is nearly the only thing the AGM has ever had a consensus on, so you may be correct that this is attempting the impossible. Or then again, you might be being reasonable, which is almost as bad as giving up

http://www.eurogofed.org/egf/agm2009sum.txt says:

15d Proposal to Adopt new European Championship Rules: No discussion, just votes on a few principles:
a. Hold it at EGC: 21-0
b. EC not affected by non-European players 16-0
c. The number of rounds to be played on the EC might be less than 10; 4-7

[Matti]

Not all the proposals are Finnish. I have collected them from the EGF executive discussions.

That there are 7+ proposals shows that it is possible. One just has to vote which of them is preferred and then work out its details.

Finnish 3: Looks odd to me. Needs more explanation before I would make any comments on it

Goals:

Define unique champion
Define other top place(about until 10th), possibly shared.

Stage 1:

16 round robin groups of four players.
Three rounds: Sunday, Monday, Tuesday
Each group may include one non-European player.
Two best European players qualify for the second stage. Only games between Europeans count. Non-European and unqualified European players return to the main tournament.
Tie breaks: to be defined

Stage 2:
Players will be paired within same number of wins.

Round 1: 32 players
Round 2: 16 players with 1 win; 16 players with 0 wins
Round 3: 8 players with 2 wins, 16 players with 1 win; 8 players will return to main
Round 4: 4 players with 3 wins; 12 players with 2 wins; 8 players will return to main
Round 5: 2 players with 4 wins; 8 players with 3 wins; 6 players with 2 wins

Final rounds:

Round 6: 1 player with 5 wins; 5 players with 4 wins; (more players). The player with 5 wins will be paired with a player with 4 wins even, if it requires a repeated encounter.
Round 7: 1 player with 6 wins; players with 5 wins; (more players): The players with 5 wins play each other even, if it requires a repeated encounter.
or Round 7: no player with 6 wins: 4 players with 5 wins: (more players). The players with 5 wins will be paired with each other even, if it requires a repeated encounter.
Round 8: 1 player with 7 wins; 1 player with 6 wins or 2 players with 6 wins. The top two players play a final for the championship, winner is the champion and loser if the second.

Definitions for places 3-10: to be added.

Pros:

The one game against non European doe not affect the place in the Championship
The number of wins in the second stage and the place are consistent (one cannot get ahead with less number of wins).
Losing one game, one still has a chance to get into the final.
A definite final

Cons:

It takes 11 rounds, one more than currently. However, if we use places 1 and 2 from the fist stage we can start with tow groups of 16 players and save one round.
The player who loses only in the final does not get a second chance (we could give him another chance, but we would need an optional game).

[topazg]

Controversial opinion: Why is it we need a European champion in the first place?

There's no Asian, Chinese, Japanese, or Korean champions ..

[RobertJasiek]

The purpose of a championship per se are clear:
a) To determine the currently strongest player.
b) To honour him.
c) To enable him to receive honour.
d) To create interesting games for the players themselves and, if games are broadcasted, for everybody.
e) In principle, one could use a championship for promoting the game in the media, but so far this has hardly been done successfully.

Now to the more specific question whether to have an EGC: The top players and many ordinary players want it for all or some of the above reasons, where "European" is inserted in (a).

***

Japan, China, South Korea have a modern tradition of sponsored tournaments. Prize moneys roughly equal prestige. Asia now has Asian Games.

It is up to other continents, regions or countries whether they want a championship for their top players.

The amateur scene has in Japan, I think, an All Japan Best Player tournament (or whatever that is called). I do not know about other Asian countries. The world has, obviously, the WAGC.

[tj86430]

Controversial opinion: Why is it we need a European champion in the first place?

This is as valid question as why do we need a world champion? Or a national champion (well, if your examples are true, then all nations don't)? Or any champions at all?

[Javaness]

Not all the proposals are Finnish. I have collected them from the EGF executive discussions..

Thank you for explaining this proposal.
Ok, on reflection I would go for Board proposal 1 if I had to make a choice from these 7. A European Championship taking place on rest days would be a good crowd pleaser. However that's just my personal selection and I still wonder if strong players will say No

Board 3 looks just too complex, I mean it is probably ok, but since nobody reads the rules before arriving at a tournament I think that it will create a nightmare.

[RobertJasiek]

The following table may be of interest to help evaluation of some of the system proposals.

The column Points denotes the number of wins of the European Champion. The column #EuropeansTied states the number of Europeans sharing MMS and Points with the European Champion; this includes himself.

Code: Select all

Year Points #EuropeansTied

2009    7   4
2008    7   3
2007    7   2
2006    8   1
2005    8   1
2004    9   1
2003    7   2
2002   10   1
2001    9   1
2000    9   1

[RobertJasiek]

The same just after round 9, but players outside the supergroup are not counted:

Code: Select all

Year Points #EuropeansTied

2009    7   1
2008    7   1
2007    7   1
2006    7   2
2005    7   1
2004    8   1
2003    6   4
2002    9   1
2001    8   1
2000    8   1

[HermanHiddema]

This suggests that shortening the EGC to 9 rounds would improve the selection of the EC. But since one of the main complaints of the European top players is that they don't play each other enough, and are instead playing weaker Europeans and strong non-Europeans, that hardly sounds like the right way to go.

I think this also supports that Dinerchtein's suggestion to make the top group smaller can really work. With a smaller top/super group, fewer people will get to 7 points after 10 rounds, and ties should become more unusual. At the same time, making the top group smaller should avoid some of the pairings against weaker European players, another benefit. I think there is something to be said for simply putting the top bar at 6 dan.

[willemien]

Hi Robert can you publish some more data?

number of topgroupplayers ( all players above the bar)
number of european contestants (all players of the above that are eligable for the best european title)

and if possible what would hapen if the bar was raised with these numbers

[RobertJasiek]

willemien, you can find out and post yourself:
http://www.eurogofed.org/results/congress.htm

Herman, I have not meant to imply that 9 rounds would be better than 10. It depends on tournament system, number of strong non-Europeans, strength distribution of Europeans in the super/top group. That during the last 10 years under the current system (with slightly varying tiebreakers) 9 rounds might have been better on average than 10 rounds might simply be coincidence. It would not easy to prove likelihood of an expected over-determination effect; there are too many input paramaters to be considered.

One reason for my little study is to get a priliminary estimate of how it would be for system proposal #2, which suggests modified Swiss and 9+ rounds. A denser strength distribution and Europeans only make a direct comparison also difficult though.

Making the top group smaller inside the current system means obviously fewer weakish opponents but it might lead to over-determination after a smaller number of rounds. To make this less likely, my suggestion for system proposal #2 is to start with 32 players; thereby one gets a similar distribution of numbers of wins for the top players during the late rounds because 9+ rounds Swiss is similar to the current system, which hides a Swiss-style supergroup on top of the McMahon heap.

Making the top group smaller also has a disadvantage: It makes pairing during late rounds harder, unless one allows repeated pairs, which is disliked. (The alternative, to pair down 2+ McMahon groups because of otherwise missing new opponents, is even worse.)

Presumably one can tune and test a lot with the exact top group size. Values between ca. 24 and 64 all might be feasible. But ca. 16 players for 10 rounds Swiss??? You would always need Andreas Ensch to plan all pairing alternatives for the next 3 rounds fairly. I doubt that it can be done during the last 3 rounds though. Starting with ca. 32 players avoids such problems, even if more and more players are thrown out, so that finally maybe only 8 to 10 remain. If there are ca. 16 before the last 3 rounds, then that should suffice for getting new opponents and fair pairings, I'd guess.

Putting top bar to 6d is too inflexible; don't forget about years like Dublin. The current system of allowing to set the top bar to either 4d, 5d, 6d or 7d is much more flexible.

[HermanHiddema]

willemien, you can find out and post yourself:
http://www.eurogofed.org/results/congress.htm

Herman, I have not meant to imply that 9 rounds would be better than 10. It depends on tournament system, number of strong non-Europeans, strength distribution of Europeans in the super/top group. That during the last 10 years under the current system (with slightly varying tiebreakers) 9 rounds might have been better on average than 10 rounds might simply be coincidence. It would not easy to prove likelihood of an expected over-determination effect; there are too many input paramaters to be considered.

I did not mean to imply that you did, it was simply an observation. Of course 10 data points is insufficient to make any definite statement about such a thing.

One reason for my little study is to get a priliminary estimate of how it would be for system proposal #2, which suggests modified Swiss and 9+ rounds. A denser strength distribution and Europeans only make a direct comparison also difficult though.

Making the top group smaller inside the current system means obviously fewer weakish opponents but it might lead to over-determination after a smaller number of rounds. To make this less likely, my suggestion for system proposal #2 is to start with 32 players; thereby one gets a similar distribution of numbers of wins for the top players during the late rounds because 9+ rounds Swiss is similar to the current system, which hides a Swiss-style supergroup on top of the McMahon heap.

Making the top group smaller also has a disadvantage: It makes pairing during late rounds harder, unless one allows repeated pairs, which is disliked. (The alternative, to pair down 2+ McMahon groups because of otherwise missing new opponents, is even worse.)

Presumably one can tune and test a lot with the exact top group size. Values between ca. 24 and 64 all might be feasible. But ca. 16 players for 10 rounds Swiss??? You would always need Andreas Ensch to plan all pairing alternatives for the next 3 rounds fairly. I doubt that it can be done during the last 3 rounds though. Starting with ca. 32 players avoids such problems, even if more and more players are thrown out, so that finally maybe only 8 to 10 remain. If there are ca. 16 before the last 3 rounds, then that should suffice for getting new opponents and fair pairings, I'd guess.

On what do you base your minimum of 24, and your dismissal of 16? Christoph Gerlach, in his doctoral thesis, suggests 11-33 players in a top group for 10 rounds McMahon. The BGA suggests 11-30 in their McMahon guide. The AGA TD guide gives no specific recommendation for 10 rounds, but their minimum for other sizes (up to 8 rounds) is always 1 more than the number of rounds, so extrapolation gives an 11 minimum again.

In general, given a number of rounds N, as the size of the top group approaches N (from above), the top group behaves more and more like round robin. As the size moves towards 2^N, it behaves more and more like knock-out. Round robin is is more desirable than knock-out, IMO, therefore we should not move too far away from N.

I think 16 would be an excellent top group size. Yes, you will have a harder time making pairings, and players will be paired down 2+ McMahon groups, but that is good. If you have 6/7 and have played all players at 5+, then a 4/7 opponent is fine. It is better than having a 0/0 weak opponent in the first round.

Putting top bar to 6d is too inflexible; don't forget about years like Dublin. The current system of allowing to set the top bar to either 4d, 5d, 6d or 7d is much more flexible.

6d was a rough suggestion based on recent years. Of course in far fewer strong players show up, the bar should be lower.

[Harleqin]

I believe that simply raising the top bar is a step in the right direction, and that this should also be "politically" feasible.

I support the query for data on the size of the top groups in recent years.

I do not think that extrapolating this data to "what if the top bar was at 6 dan" is possible, because the top bar decision also leads to completely different pairings.