Many years ago at a MARFU playoff event, seeds were based upon apparent strength of teams. This was fine until a team surprised the MARFU administrators and upset a higher seed. In turn, the event organizer(s) changed the matches for the semi-finals because the #1 and #2 teams from the same union were going to be playing each other. This infuriated the "surprise" team because they felt the change was not to their advantage and that they had approached the tournament based upon their potential path they would take, assuming victories in their first two matches. The team eventually petitioned the MARFU Board (December 14, 1996).
The MARFU Board, based upon the meeting with the team, made a number of decisions. The decisions were as follows:
The solution should be a system that could be used to predict the seeds in the future.
Seeds or positions should not change once the tournament starts.
Because MARFU advances more than one team, cross-LAU or intra-LAU play is important. Two teams from the same LAU should not face each other in the opening round.
We should avoid playing the #1 and #2 teams from the same LAU against each other before the finals.
The MARFU Board tasked the collegiate committee to create a system based upon the criteria (December 14, 1996).
Solution - aka the "ABC System"
In an attempt to meet the criteria, Kris Kabza proposed (March 9, 1997) the following setup:
1. ARU 1
2. BRU 1
3. CRU 1
4. BRU 2
5. CRU 2
6. ARU 2
7. ARU 3
8. BRU 3
Where the ARU is the LAU that contains the team that wins MARFU from the previous year and the BRU is the LAU that contains the team that finishes the next highest.
This leads to:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
Note: It was suggested that only the higher finishing teams be used in determining the following year's seeds, that is, finishes fifth through eighth are not necessarily used for the following year's seeds.
Tests and Checks of the ABC System
To make sure we avoid intra-LAU matches in the quarterfinals and to avoid intra-LAU matches of the #1 and #2 LAU seeds prior to the finals, here are several checks and possibilities.
By observation, it is obvious that no intra-LAU matches win occur in the quarterfinals.
As for an LAU's #1 team and #2 team facing each other before the finals, while there are several combinations that can occur, the most likely way for this type of conflict to arise is if an LAU wins all its first round matches. Therefore, this extreme possibility will be used for examples:
What if the ARU wins all its first round (quarterfinal) matches?
This leads to the following semi-finals:
ARU 1
ARU 3
ARU 2
BRU 2 v CRU 2
or
ARU 1 v BRU 2/CRU 2
ARU 3 v ARU 2
Thus meeting the MARFU criteria.
What if the BRU wins all its first round (quarterfinal) matches?
This leads to the following semi-finals:
BRU 3
BRU 1
CRU 1 v ARU 2
BRU 2
or
BRU 3 v BRU 2
BRU 1 v CRU 1/ARU 2
Thus meeting the MARFU criteria.
What if the CRU wins all its first round (quarterfinal) matches?
This leads to the following semi-finals:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1
CRU 2
or
ARU 1/BRU 3 v CRU 2
BRU 1/ARU 3 v CRU 1
Thus meeting the MARFU criteria.
Changes and Tweaks to the ABC System
In the past, this setup has been changed or tweaked for a few reasons. The various changes are listed.
Drop to Two Division I Teams, Last Minute Fourth Seed Award
One of these changes was due to MARFU Men's Division I receiving four seeds to the National Championship Round of 16 (2000) and because the VRU dropped to two Division I teams. In the spring of 2000, in an attempt to better approximate the top four seeds, so that MARFU would send its best four teams, a play-in match was used. As a result of the play-in match, seeds 5 through 7 were determined. The PRU received 4 seeds, the EPRU 2, and the VRU 2. However, the #8 team did not change and PRU 4 was forced to play PRU 1 in the opening round.
Because the VRU finished in second place in MARFU that year, there was now a dilemma about the seeds for spring 2001. After some debate, it was decided that the EPRU would get the seed given up by the VRU. For spring 2001, the seed structure was:
PRU1 hosts EPRU3
VRU1 hosts PRU3
PRU2 hosts EPRU2
EPRU 1 hosts VRU2
Expanding out the various possibilities, it can be noted that the criteria, laid by the MARFU Board, is met.
Multiple High Finishes
In 2001, 2002, and 2003, the Potomac Rugby Union finished in the first and second positions in the Men and Women's Division I playoffs. In an attempt to reward the PRU for their multiple high finishes, and still meet the criteria set forth by the MARFU Board, the setup has been modified. In Men's Division I, the VRU has continued to finish "the next highest", but still has two Division I teams. To accommodate this, the setup was modified. Where the ARU is the PRU, the BRU is the VRU, and the CRU is the EPRU, the following setup was used
1. ARU 1
2. ARU 2
3. BRU 1
4. CRU 1
5. BRU 2
6. ARU 3
7. CRU 2
8. CRU 3
This leads to
ARU 1 v CRU 3
ARU 2 v CRU 2
BRU 1 v ARU 3
CRU 2 v BRU 2
For the PRU Women's Division I, where the EPRU finished second, and there was no need to worry about the "next highest" LAU not having enough seeds, the following setup was used:
This leads to
ARU1 v BRU3
ARU2 v CRU2
BRU1 v ARU3
CRU1 v BRU2
Four Seeds to an LAU
For the spring of 2002, the PRU men's Division I was awarded four seeds to the MARFU playoffs. That year, IUP, a sweet sixteen team joined the PRU. Now with 3 potential sweet 16 teams, MARFU felt the PRU should be given an extra seed to the tournament. However, in the following years, the PRU would have to earn that fourth seed.
The setup for 4 seeds for Men's Div I in 2002 was Pool A
PRU 1 v EPRU 2
VRU 1 v PRU 4
Pool B
EPRU 1 v PRU 3
PRU 2 v VRU 2
For the semifinals, the winners within each pool played each other. For the consolation round, the loser within each pool played each other. For the play in the final round:
1st - #1 Pool A v #1 Pool B
3rd - #2 Pool A v #2 Pool B
5th - #3 Pool A v #3 Pool B
7th - #4 Pool A v #4 Pool B
There was some question about the setup of this system. Should the VRU1 (which is the "next highest" LAU) have had an easier first game and be put against the PRU 4? Or should the VRU1 have been put into a bracket where the VRU1 was more likely to make it to the finals, but have a harder first match, going against PRU3? That is, should the VRU1 and EPRU1 been switched in their brackets?
In the spring of 2003, the EPRU Women's Div II finished first, second, and third. The MARFU Board awarded the EPRU Division II a fourth seed. Taken into consideration was the fact that the PRU had also dropped to one Women's Div II team. During the fall season VRU #1 soundly defeated the PRU's lone Women's Div II team.
The setup for 4 seeds for Women's Div II in 2004 was Pool A
EPRU 1 v VRU 2
EPRU 3 v PRU 1
Pool B
EPRU 2 v VRU 3
VRU 1 v EPRU 4
Pools
Beginning in 2002, the MARFU Men's Collegiate Committee and the Women's U23 Committee began to place the teams into "Pools" rather than confuse the issue in listing teams as 1-8.
In the situation where the LAUs had 3, 3, and 2 seeds, the ARU1 was put into one pool and the BRU1 and CRU1 was put into the other pools.
To avoid playing the #1 and #2 from the same LAU before the finals, the ARU2 was put into the opposite pool as ARU1. The same was done for BRU2 and CRU2, resulting in: Pool A
ARU1
BRU2
CRU2
Pool B
BRU1
CRU1
ARU2
Adding the #3 teams from the ARU and BRU results in:
Pool A
ARU1
BRU2
CRU2
BRU3
Pool B
BRU1
CRU1
ARU2
ARU3
Pitfalls of the ABC Setup
To ensure I have a good grasp on this setup, a program was written to simulate the various finishes depending upon the number of advancing seeds.
With 8 teams, there are 40320 possible combinations, however, for the program I made an assumption that the ARU #1 team was better than the ARU #2 team, which was better than the ARU #3 team. I made the same assumptions for the BRU & CRU. While I understand that this is not always the case, it is a close approximation and can easily be changed in the program. Using this assumption, this brought the number of possible combinations down to 560.
The following are pitfalls of the ABC System based upon the number of teams advancing to the National Championship Series Round of 16 or Round of 8.
1 Team Advancing
Really any setup is fine, in the end; a team needs to win all its games regardless of the setup. There are no pitfalls in this case.
2 Teams Advancing
Without knowing relative strengths of teams across MARFU the ABC setup gives excellent intra-LAU matches to determine relative strength of teams in differing LAUs. Can be problematic if the BRU and CRU truly have the #1 and #2 teams. (However, this would hold true for a system that also tried to determine seeds based upon perceived strength.) This could potentially advance the #1 and #3 teams from MARFU.
For example, given the setup:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
If BRU1 really is the #1 team and CRU1 is the #2 team, then CRU1 would be eliminated in the semis and the BRU1 (#1 team) and ARU1 (#3 team) would advance.
3 Teams Advancing
Can be problematic for the CRU, if the ARU has two strong teams. This could potentially lead to the advancement of the #1, #2, #4 teams within MARFU.
For example, given the setup:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
If the ARU has the two strongest teams, and the BRU1 really is the #4 team and the CRU1 is the #3 team, CRU1 (#3 team) would be eliminated in the quarterfinals, sending ARU1 (#1), ARU2, (#2), and BRU1 (#4).
4 Teams Advancing
Can be very problematic depending upon team strength through out MARFU. This could potentially lead to the advancement of the #1, #2, #4 & #6 teams within MARFU. The advancement of 4 seeds has happened with the Div I men, and the setup was slightly modified before the playoffs to attempt to place teams correctly. This has also happened with the women. When approached with the concern, they declined to change the setup.
For example, given the setup:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
Using the same scenario for 3 teams advancing and if the ARU3 is really the #5 team, then BRU2 or CRU2 are at best the #6 team. Since BRU2 and CRU2 face each other in the quarterfinals, the winner would advance to the nationals. Therefore, sending ARU1 (#1), ARU2, (#2), BRU1 (#4), and BRU2vCRU2 (#6).
5 Teams Advancing
Has the potential to be problematic for ARU #2 & #3 if they were truly the #4 & #5 teams in strength in MARFU (would face CRU #1 & BRU #1 in first round). This could potentially lead to the advancement of the #1, #2, #3, #6 & #4 teams within MARFU (in that order). Would definitely need consolation round.
For example, given the setup:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
If the real strengths were ARU1 (#1), BRU1 (#2), CRU1 (#3), ARU2 (#4), ARU3 (#5), then on top of the scenario for 4 teams advancing, CRU1 (#3) defeats ARU2 (#4) in the quarterfinals. ARU2 wins the consolation bracket. Therefore, sending ARU1 (#1), BRU1, (#2), CRU1 (#3), BRU2vCRU2 (#6), and ARU2 (#4).
6 Teams Advancing
Will we ever get this? Has the potential again to hurt ARU #2 & #3. This could potentially lead to the advancement of the #1, #2, #3, #6, #4, & #7 teams within MARFU (in that order).
This is simply an extension of the "5 teams advancing" scenario.
7 Teams Advancing
Okay now this is nuts. The number 8 team will definitely not advance, but this could potentially hurt ARU #2 & #3. This would lead to the advancement of the #1, #2, #3, #6, #4, #7, & #5 teams within MARFU (in that order).
For example, given the setup:
ARU 1 v BRU 3
BRU 1 v ARU 3
CRU 1 v ARU 2
BRU 2 v CRU 2
If the real strengths were ARU1 (#1), BRU1 (#2), CRU1 (#3), ARU2 (#4), ARU3 (#5), then on top of the scenario for 4 teams advancing, CRU1 (#3) defeats ARU2 (#4) in the quarterfinals. ARU2 (#4) defeats ARU3 (#5) in the consolation semifinals. ARU2 (#4) wins the consolation bracket. ARU3 (#5) plays #8 for the seventh seed. Therefore, sending ARU1 (#1), BRU1, (#2), CRU1 (#3), BRU2vCRU2 (#6), ARU2 (#4), #7, and ARU3 (#5).
BOTTOM LINE: All the "what ifs" are enough to drive you nuts.
Other Systems
Fall LAU / Spring MARFU Season
Only one other system has been presented to the MARFU Board. The system calls for a fall LAU season and a spring MARFU season. That is, each LAU determines the ranks of their teams. For the spring, each LAU sends a predetermined number of teams to the MARFU round. The teams in the MARFU round would play each other through out the season. There would be NO MARFU playoffs.
The number of teams that could advance from each LAU can be either two or three.
If two teams advance from each LAU, then the #1 team from each LAU is automatically given a 1-0 record, the #2 team is given a 0-1 record. Each team would then be required to play 4 games for their standings.
If three teams advance from each , then the #1, #2, and #3 teams from each LAU are given a 2-0, 1-1, and 0-2 record, respectively. Each team would then be required to play 6 games for their standings.
Teams would play matches against the teams from the other LAUs. Teams from the same LAU would not face each other during the MARFU season
The top MARFU advancers to the National Championship Series Round of 16 or 8 would be based upon MARFU standards. A tiebreaker system would need to be defined.
The MARFU Board never voted on the proposed system. No second motion was made to bring it to a vote.
Coaches Poll
The idea of polling the coaches to determine MARFU seeds was presented to the MARFU Board. The idea was not presented with a full proposal, but mentioned that perhaps the coaches in the playoffs should be polled to determine the seeding positions. However, not all the teams that are in the playoffs have a coach. There was no recommendation what should be done in a situation were a team has no coach. During the entire discussion, no "plan" was ever present, only ideas were expressed.