Tournament pairing methods


#1

Hi there.

Is there any documentation on how the pairing methods work? I can’t find anything in the FAQ, nor forum, nor internet search. Of course there is plenty of information on the net about Swiss and McMahon, but there’s nothing on “Slide”, “Slaughter”, etc.


#2

Hi there, you can find some basic information here:

http://senseis.xmp.net/?GroupPairing


#3

Slide is flip and Slaughter is fold, paring method.


#4

Sorry for reviving this old thread, but I don’t get it.

In https://senseis.xmp.net/?McMahonPairing I find this: “Pair players within the bands according to one of the methods outlined in Group Pairing.”

When looking at https://online-go.com/tournament/29263 the slide pairing is seemingly applied across all players, not within one band. In how far is this still a MCMahon tournament? What am I missing???


Is There Documentation About The Different "Create Tournament" Settings?
#5

Does nobody know or is my question just too silly? :slight_smile:


#6

More likely no-one knows :wink:


#7

Or the right people haven’t seen the question yet :wink:


#8

Looks like a bug. At least the description of the tournament does not match the pairings.


#9

I still see the same same problem I have mentioned before. Look at https://online-go.com/tournament/31547

The first game is between a 9d and a 17k, The 9d has a starting score of 0 the 17k of -25. This is not McMahon pairing, it is swiss pairing. And worse that that: It is swiss pairing that additionally gives weaker players a handicap. (Yes, McMahon is meant to do that, but in exchange everyone gets similarily ranked players as opponents from the start, which is NOT happening here.) This makes no sense at all and should be fixed. I find it strange that nobody else seems to be bothered.

Well, maybe I’m missing something. In that case, would someone enlighten me? :slight_smile:


#10

Well actually…!

That tourney is set up in a funny way. It has a random pairing, why pairing does npot make any sense. It also has handicaps set to “none”


#11

I don’t think it’s set up in a funny way at all. ‘Random’ is one of the standard pairing methods WITHIN A GROUP, see https://senseis.xmp.net/?GroupPairing

The problem is that this randomness is applied ACROSS all groups. And that’s a software bug, not a problem with the setup of this specific tournament.


#12

But that ain’t simultaneus McMahon, just regular MM. So all players are in same group and they only play one game at a time. Everyone gets a a negative base MM-points, exept the 9d’s. Then they are paired randomly for a single game (even games, random colours, 5 days for each move), and when all those games are over they’ll be paired again randomly. This will last for 10 rounds, after which tournament standings are calculated on base points + gained points for wins.

That tournament is just set up in that way by 21k tournament director. It’s not a software bug, but a little funny decision for chosing a tournament format.


#13

Well, maybe I am totally misunderstanding what McMahon means. I have may information from: https://senseis.xmp.net/?McMahonPairing

Before the First Round
Order the field of players by strength. Divide the field into bands of players based upon number of rounds and the distribution of players strengths. Players in the same band will be paired against each other in the first round. There are several ways to do this, for details refer to the section on assigning initial bands. The highest group is generally referred to as the “top group”, the lowest as the “bottom group”.

Assign starting McMahon Scores
There are two main ways this is done. The first is to set the McMahon Score of the top group to 0, with all other bands starting at a negative score. So the next weaker band is -1, down to the weakest band which is -(M-1) for M bands. The other is is to set the McMahon Score of the bottom group to 0, with all other bands receiving a positive initial score. In effect, the early slaughter rounds of the Swiss pairing are replaced by assigning a number of virtual won games to stronger players. The stronger the player the greater the number of won virtual games assigned to that player.

Note that this is nowhere talking about simultaneous McMahon.

I think what is described in senseis is not at all what is happening in our tournament.

There may be 4 possibilities:

  1. I am reading it wrong and it actually is what is happening in the tournament. In that case, could someone please help me understand?

  2. What is written in senseis is just wrong. It should be changed there.

  3. There are many different definitions of McMahon and the one on OGS happens to differ from that on senseis.

  4. There is a bug.


#14

There the top group would be 9 dans and over, and bands go weaker till 25k. For the first round all the players are paired randomly regardless of their MM score. Then for the subsequent 9 rounds this same thing is repeated. After 10 rounds of randomly paired games, player who’s still on the tournament with the highest combined base and gained points will win the tournament.

There is no bug there, that’s just how the TD has decided that tournament to be.

Note: Since OGS allows users to create our own tournaments with large variety on tournament formats and settings, you shouldn’t be too surprised to find some tournaments with “funny” or “bug-like” settings :slight_smile:

So it’s the option 3 ^^


#15

…But I’m not seeing a setting for group within a McMahon bar…

Like, if there is an option for pairing without regard for McMahon band, how do tournament players tell the difference?


#16

Well if you really think it’s option 3, is the OGS definition used anywhere else in the world? I sort of doubt that tbh, because in my understanding the entire idea of McMahon is that of an improved swiss system that will ensure reasonable pairings from round 1. This is exactly what this OGS definition is not providing. So why call it McMahon?


#17

Umm tournament director has set pairing to be random, this is why pairing is done randomly. It’s not “OGS definition” but “definition by anime-link 21k” and yeah it’s not used anywhere else, because that doesn’t really make any sense.

If someone wants to make MM tournament where players are paired based on similiar MM points, then the pairing method should be “strength” instead “random”


#18

I think it is a bug. Pairing is intended to be random only within the group with the same MM points.