[bd71]

I have almost zero experience in knockouts NOT having 2^x teams.

One of the mini-Spingolds starting tomorrow will go from 30 to 16. Questions:

1. I assume this will likely be 12 head-to-heads and two "2 of 3 qualify" round robins. Does this sound right? Is it automatic, or is there director discretion?

2. Assuming #1 is right, is there a standard way to set it up with seeds? Will seeds 1/2 get into the round robins, and if so will they play the 4 teams seeded 27-30? Or maybe the 1/2 seeds face the teams seeded 15-18? Or is this all non-standardized and up to director discretion?


Edit: In you you're wondering, the conditions of contest are no help at all: "The DiC will create a bracket to the best playing advantage of the field."

[Mbodell]

Yes, the standard way is to play 3-ways with 2 surviving, 4-ways with 3 surviving, and head-to-head as much as possible. In theory there could be byes (I heard the full spin has 74 teams, I'd expect 40 teams playing 4-ways with 3 survive and 34 teams with a bye).

For the micro spingold especially I doubt very much that they want byes. Having a small field of 30 will make day 1 much more pleasant than the times they've had to go from ~60 teams to 16 in one day. So 2 3-ways and 12 heads up sounds right. Traditionally the headsup matches are considered advantaged for the top seeds, instead of the 3-way 2 survive (a long match with a worse team is better than two short matches with not as bad teams). I think that is a little odd, but it is the way it works. So the top 12 seeds should play the bottom 12 seeds in heads up matches. And then there should be 2-3 way matches with, I think 13, 16, and 17 in one and 14, 15, and 18 in the other. Your idea of 1 + 27 + 30 or what ever in the 3 way is also really unfair to teams 26/27. 26 would need to beat the 3 seed, while 27 just needs to be one of 2 to survive with the 30 seed and 1 seed (in essence beat the 30 seed assuming that the 1 outclasses both).

In actual fact the micro-spingold especially is poorly seeded because the seeds are often reversed. There are exceptions, but a team that has an average near 1500 but still chooses to play the 1500 is often worse than a team of up and comers with an average of only a couple of hundred points. Certainly not always, but often.

[TimG]

There should be a prescribed format for a field of 30 teams which says exactly which seeds are in which matches. If mini-Spingold CoC don't specifically state such things, then there should be a general ACBL document for the DiC to reference in order to consistently serve the "best playing advantage of the field".

There is no easy answer to the seeding/matchup problem. Take Mbodell's solution of the top 12 seeds playing the bottom 12 seeds and

13 v 16 v 17
14 v 15 v 18

Would you rather be the 12 seed who has to beat the 19 seed in a head-to-head match or the 13 seed who has to survive a 3-way with 16 & 17? It would depend upon the difference in strength amongst the seeds in the middle of the field, of course, but assuming nearly equal strength of teams in the middle, the 13 seed has an easier path to the second round than does the 12 seed. The 13 seed has at least a 2/3 chance of surviving the 3-way (they are favorites against each of 16 and 17), meaning that the 12 seed would have to be at least a 2:1 favorite to beat the 19 seed in order for their path to be easier. 16 & 17 also have an easier path than 12. I could understand 12 not being happy with their seed!

I think for 30 teams you should have something like:

1 v 10 v 24
2 v 9 v 23
3 v 30
4 v 29
5 v 28
6 v 27
7 v 26
8 v 25
11 v 22
12 v 21
13 v 20
14 v 19
15 v 18
16 v 17

(You do not want 1 v 9 v 24 & 2 v 10 v 23 in the first round because that could lead to a 1 v 10 or 2 v 9 rematch before the final.)

[TimG]

Mbodell, on 2012-July-15, 23:33, said:

(I heard the full spin has 74 teams, I'd expect 40 teams playing 4-ways with 3 survive and 34 teams with a bye).

It appears there are 36 teams playing 4-ways with three survivors and 2 teams playing a head-to-head. That's 28 teams surviving from play today. Added to 36 with byes gets to 64.

Today's matches are:

37 v 74

38 55 56 73
39 54 57 72
40 53 58 71
41 52 59 70
42 51 60 69
43 50 61 68
44 49 62 67
45 48 63 66
46 47 64 65

I don't know the rationale behind a single head-to-head instead of all 4-ways.

[Mbodell]

I heard the rational for the one heads-up match in the full spingold was so that the seeds worked out where there will need to be no playback until the final amongst the 4-way teams. I didn't look into it though.

In the mini-spin there were 38 teams and when there are slightly more than the 32 they need to cut it to on day one they do a fairly sensible 10 3-way with 2 survive and 4 heads up (eliminate 14 teams to leave 24 teams), then tomorrow it is 8 3-way with 2 survivors to get to 16. Even though full day matches are nice, it is nicer to cut 14 and then 8 rather than 6 and then 16.

[gnasher]

TimG, on 2012-July-16, 09:04, said:

I don't know the rationale behind a single head-to-head instead of all 4-ways.

It seems to treat team 74 rather unfairly.

[TimG]

Mbodell, on 2012-July-16, 20:59, said:

I heard the rational for the one heads-up match in the full spingold was so that the seeds worked out where there will need to be no playback until the final amongst the 4-way teams. I didn't look into it though.

With three survivors from each four-way, you can't place those three in three different halves of the draw. The best you can do is make sure they don't playback until the semi-final.

For instance, the 38 v 55 v 56 v 73 match from Monday places the #38 seed in the #3 quarter and the #55 seed in the #2 quarter; if they both keep surviving, they will meet in the semi-final.

Perhaps the objective was to prevent playbacks until the semi-final, but that could have been accomplished with some seed flipping (as I did for the example for 30 teams). That would seem to me to be more fair to the #74 seed (especially given that the bottom group of seeds are likely shuffled anyway -- #74 didn't get there purely through seeding, but through a random draw from amongst the bottom few seeds).

[paulg]

TimG, on 2012-July-17, 04:44, said:

Perhaps the objective was to prevent playbacks until the semi-final, but that could have been accomplished with some seed flipping (as I did for the example for 30 teams). That would seem to me to be more fair to the #74 seed (especially given that the bottom group of seeds are likely shuffled anyway -- #74 didn't get there purely through seeding, but through a random draw from amongst the bottom few seeds).

Except for the top two seeds, all the teams are shuffled in increasingly large pools. From the CoC:

The teams will be ordered by their average seeding points. In cases where a tie broken by lot involves teams in two groups, the loser(s) of the tie-break will be given the high seed(s) in the next group.

• The defending champion will be seeded number one provided that at least four original members are playing together. Otherwise the team with the highest average seeding points will be seeded number one.
• The next ordered team will be seeded number two.
• The next two teams will be assigned seed numbers three and four by lot. Then, the next four teams will be assigned seed numbers five through eight by lot. Similarly, seed numbers will be assigned by lot for positions 9-12, 13-16, 17-20, 21-24, 25-28, 29-32, 33-40, 41-48, 49-56, 57-64, 65-80, 81-96, and so on in groups of 16.

[TimG]

I guess I shouldn't have said "likely". I knew the seeds were shuffled, just didn't know off hand how big the groups were. In this case, seeds 65 through 74 were shuffled, so the eventual 74 got a raw deal in having to play a head-to-head instead of compete in a 4-way with 3 survivors.

[aguahombre]

TimG, on 2012-July-17, 07:18, said:

I guess I shouldn't have said "likely". I knew the seeds were shuffled, just didn't know off hand how big the groups were. In this case, seeds 65 through 74 were shuffled, so the eventual 74 got a raw deal in having to play a head-to-head instead of compete in a 4-way with 3 survivors.

Does this mean they were shuffled to arrive at their seeding number, or they were shuffled to decide which one got stuck with the head-to-head and 74 was unlucky in addition to being the original bottom seed?

[TimG]

aguahombre, on 2012-July-17, 11:11, said:

Does this mean they were shuffled to arrive at their seeding number, or they were shuffled to decide which one got stuck with the head-to-head and 74 was unlucky in addition to being the original bottom seed?

This means that the bracket was pre-determined, it was known that 74 would play 37 in a head to head. Then the teams that were seeded 65 through 74 were put in a hat and drawn randomly to fill the 65 through 74 slots on the bracket sheet.

[Mbodell]

TimG, on 2012-July-17, 04:44, said:

With three survivors from each four-way, you can't place those three in three different halves of the draw. The best you can do is make sure they don't playback until the semi-final.

For instance, the 38 v 55 v 56 v 73 match from Monday places the #38 seed in the #3 quarter and the #55 seed in the #2 quarter; if they both keep surviving, they will meet in the semi-final.

Perhaps the objective was to prevent playbacks until the semi-final, but that could have been accomplished with some seed flipping (as I did for the example for 30 teams). That would seem to me to be more fair to the #74 seed (especially given that the bottom group of seeds are likely shuffled anyway -- #74 didn't get there purely through seeding, but through a random draw from amongst the bottom few seeds).

With 4-ways you don't actually play most of the other teams. You only play one team headsup, and then one team will play both. So it isn't quite as clear to me. Of course nearly every team from the 4-ways should be gone by the round of 8.

[TimG]

Mbodell, on 2012-July-17, 21:59, said:

With 4-ways you don't actually play most of the other teams. You only play one team headsup, and then one team will play both. So it isn't quite as clear to me. Of course nearly every team from the 4-ways should be gone by the round of 8.

So, in a four-way with A, B, C, & D, A plays B and C plays D, then the losers playoff.

So, any combination of teams can advance, but some pairs can't both advance and have played each other. A & C can't both advance, for instance, if they have played each other. That would seem to make it easier to set the initial matches in such a way that avoids playbacks.

Tim