European Championship Rules

[Javaness2]

Previously, the quarter-final losers re-entered the McMahon; this was considered good for top non-Europeans, who would get stronger opponents in rounds 8 to 10. The new ruling emphasises more top European-European games; this is an advantage for the top Europeans, who like to get more chances to play against other top Europeans in the sparse tournament calendars these years.<snip>
These changes and language corrections were made on behalf of the EGF Committee. I am aware that the language is still not perfect, but at least it is better than before.

This is puzzling. I'd have thought that maintaining the integrity of the Open was more important than ensuring 4(?) extra europe-europe games? I don't think 4 games is going to compensate for no Ing Cup (etc).

[speedchase]

If you do a good job of setting pairings, then swiss and KO give the same winner, the only real difference is that swiss has people who have lost continue to play, and assigns rankings based on those games as well.

[RobertJasiek]

This is puzzling. I'd have thought that maintaining the integrity of the Open was more important than ensuring 4(?) extra europe-europe games? I don't think 4 games is going to compensate for no Ing Cup (etc).

Different people have different opinions on that aspect.

I am, however, not sure about the Committee's motivation; maybe they simply want exact places 5 to 8 and have the impression that playing them out in the KO is more meaningul than the McMahon's tiebreaker lottery?

[RobertJasiek]

If you do a good job of setting pairings, then swiss and KO give the same winner,

Setting good pairings is not necessarily sufficient; one might also need a number of participants that is a power of 2. Otherwise one needs enough rounds and some extra criterion such as "the only first player during the tournament to win R games wins the tournament".

[speedchase]

If you do a good job of setting pairings, then swiss and KO give the same winner,

Setting good pairings is not necessarily sufficient; one might also need a number of participants that is a power of 2. Otherwise one needs enough rounds and some extra criterion such as "the only first player during the tournament to win R games wins the tournament".

But in KO if there isn't a power of two, there is a similar arbitrariness regarding who gets byes.

[HermanHiddema]

Swiss is better pretty much by definition. [...] The KO system, on the other hand, guarantees that place 4 (1 win, 2 losses) ends above place 5 (2 wins, 1 loss).

You mention an aspect for which Swiss is better than KO. You do not mention another aspect for which Swiss is usually worse than KO: final placement tiebreakers. If, however, you can agree on sharing Swiss final places instead of using tiebreakers and if Swiss contains an implicit KO for place 1, then we can agree on such a usage of Swiss being better than KO for a determination of places 2 to 8 as meaningful as possible within the given number of (here: 3) rounds.

I never have a problem with sharing places. If, however, we do wish to distinguish places, then even Swiss + lottery is better than the sorting provided by a continued KO.

Continued KO is just a really really really bad system. Consider the case of more than 3 rounds, e.g. suppose 5 rounds continued KO with 32 players. Then the player in place 16 has 1 win 4 losses and ends above a player with 4 wins, 1 loss. Having one game weight heavier than all other games combined is terrible. Continued KO is an absolutely indefensible system.

By removing strong players from the main tournament, you work against purpose 2 [determination of European Open Champion], because you make it easier for Asian players to be Open Champion by removing the toughest opposition.

I prefer to make a weaker claim: there is a partial conflict in aims.

I am making the claim than "player's personal preference in opponents" is a bad pairing criterium. Do you agree or disagree?

[RobertJasiek]

if there isn't a power of two

For the tournament in question, 8 as a power of 2 is guaranteed.

[RobertJasiek]

I am making the claim than "player's personal preference in opponents" is a bad pairing criterium. Do you agree or disagree?

I agree for the SPECIFIC choice of which opponent. I disagree for the GENERIC preference to play against a certain expected number of European opponents in top field of the EGC.

[Zombie]

What kind of tournament system is "KO" even referring to? A single-elimination knockout bracket or perhaps a winner stays until defeated kind of arrangement?

FWIW, elimination tournaments should probably be done as seeded double elimination with a reasonable series length (this should guarantee a pretty accurate Top 2/3 or so, rest whatever. Also guarantees two games even to the worst players)

Swiss takes pretty long, but ranks people better. In Magic, the preferred tournament format is random seeded Swiss (byes to up to 3rd round for known professionals) with best of three rounds that is played for N rounds where N depends on turnout. After N rounds the best 8 or 4 (depending again on size) players are cut to a single-elimination knockout tournament that determines the winner (Final or even whole top 8 can be played out with a longer series length than Bo3). It's reasonably accurate yet provides that finals excitement.
This system is decent in Magic but should work even better in Go because:
1. It provides weak people with more games to play.
2. In Magic, the Round 4 field is drastically different than the Round 1 field, which gives pros an advantage in that they can prepare for a narrowe field than the normal people need to. This is not an issue that exists in Go.
3. The cut to Top8 can cause matchup issues based on the chosen decks. Again, not an issue in Go.

[Javaness2]

straight knock-out

[HermanHiddema]

What kind of tournament system is "KO" even referring to? A single-elimination knockout bracket or perhaps a winner stays until defeated kind of arrangement?

FWIW, elimination tournaments should probably be done as seeded double elimination with a reasonable series length (this should guarantee a pretty accurate Top 2/3 or so, rest whatever. Also guarantees two games even to the worst players)

Swiss takes pretty long, but ranks people better. In Magic, the preferred tournament format is random seeded Swiss (byes to up to 3rd round for known professionals) with best of three rounds that is played for N rounds where N depends on turnout. After N rounds the best 8 or 4 (depending again on size) players are cut to a single-elimination knockout tournament that determines the winner (Final or even whole top 8 can be played out with a longer series length than Bo3). It's reasonably accurate yet provides that finals excitement.
This system is decent in Magic but should work even better in Go because:
1. It provides weak people with more games to play.
2. In Magic, the Round 4 field is drastically different than the Round 1 field, which gives pros an advantage in that they can prepare for a narrowe field than the normal people need to. This is not an issue that exists in Go.
3. The cut to Top8 can cause matchup issues based on the chosen decks. Again, not an issue in Go.

The system under discussion here is the one used to determine the European Championship.

The original design was 7 rounds of McMahon (a variant of Swiss), then the top 8 European players at that point enter a 3 round single elimination knockout (with, before it, some optional relegation games in case of ties around place 8). Any player knocked out returns to the McMahon.

Now, they have changed it into some kind of freaky continued knockout, where players that are "knocked out" are not actually knocked out, but play against others that were knocked out, and the top 8 is decided entirely on the order of your wins (not the number of your wins). So if you go: win-lose-lose, you are fourth. If you go lose-win-win, you are fifth.

This system goes against everything we know about sensible tournament organization. And the only reason for it is that those European players in the top 8 do not like to play against Asian players. It is an absolute disgrace.

[Javaness2]

This system goes against everything we know about sensible tournament organization. And the only reason for it is that those European players in the top 8 do not like to play against Asian players. It is an absolute disgrace.

Sorry, but is there any evidence that the top European players have asked for this system?

[HermanHiddema]

This system goes against everything we know about sensible tournament organization. And the only reason for it is that those European players in the top 8 do not like to play against Asian players. It is an absolute disgrace.

Sorry, but is there any evidence that the top European players have asked for this system?

In post #3 in this thread, Robert Jasiek wrote (emphasis mine):

Previously, the quarter-final losers re-entered the McMahon; this was considered good for top non-Europeans, who would get stronger opponents in rounds 8 to 10. The new ruling emphasises more top European-European games; this is an advantage for the top Europeans, who like to get more chances to play against other top Europeans in the sparse tournament calendars these years.

And in post #13, he writes (emphasis mine again):

2) the reason that top Europeans want to play more among themselves carries some weight (not for the quality of the ordering of places 2 to 8 but for the attractivity for the Europeans).

From that, I gather that strong player wishes played a major role in this decision. Certainly, there is no good tournament system based reason for it.

[Javaness2]

I appreciate that Robert has relayed these words here, but are they accurate? Or is this just a whim of the EGF, or 1 or 2 people who decided to lobby it.

[HermanHiddema]

I appreciate that Robert has relayed these words here, but are they accurate? Or is this just a whim of the EGF, or 1 or 2 people who decided to lobby it.

I don't know. I go by what information I have, which is what was relayed here by Robert.