Announcement

Collapse
No announcement yet.

2019 Agenda - TPR Bracketing for ORCS

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • 2019 Agenda - TPR Bracketing for ORCS

    "TFC-03: Motion by Bernstein that, at ORCS, AMTA will use the following pairing system designed to equalize strength of schedule: Teams will be divided by TPR into four groups: Groups A (teams ranked 1-6), B (7-12), C (13-18), and D (19-24). Each team will face exactly one team from each of the four groups. The Tabulation Advisory Committee is directed to create a detailed implementation of this policy for the Board's consideration at the 2019 mid-year meeting.

    Rationale: AMTA's entire system of determining winners -- based first on number of ballots won – is predicated on a false assumption: that teams face opponents of equivalent difficulty. In fact, AMTA's pairing system makes no effort to equalize schedule strength and, in some ways, is designed to reduce the likelihood that teams face schedules of equivalent difficulty (by pairing high-high for two rounds, and inconsistently overall). The proposed pairing system would create a more balanced playing field. TPR is not perfect, but this would make it far less likely that a team faces, say, three of the very best teams in the tournament, and it would make it far less likely that a team faces, say, none of the very best teams in the tournament. It does not advantage top-ranked teams, as they would also have to face exactly one team in Group A. I used this format at three tournaments, and it substantially reduced the standard deviation of teams' CS – in other words, it created more similar strengths of schedule. Note that this could not be used at Regionals, as it requires a field size of a multiple of eight, and (currently and historically) Regionals have uneven field sizes."


    Discuss this item here!

  • #2
    I'd be interested to see what the actual correlation between TPR and a team's actual ability is. TPR obviously only measures performances form previous years -- does it actually do a good job at predicting how well a team will do.

    Comment


    • #3
      Originally posted by The Real Mock Prodigy View Post
      I'd be interested to see what the actual correlation between TPR and a team's actual ability is. TPR obviously only measures performances form previous years -- does it actually do a good job at predicting how well a team will do.
      To give a rough sense of how well TPR tracks success, here's a quick rundown of the Top 50 from the 2018-2019 TPR with how that team did in that season.
      TPR Rank School Results
      1 Miami University A 12-4 NCT (2nd Place)
      2 Yale A 13-3 NCT (1st Place)
      3 Virginia A 11-4-1 NCT (2nd Place)
      4 Rhodes A 13-2-1 NCT (1st Place)
      5 Michigan A 5-3 ORCS
      6 Georgia Tech A 8-8 NCT (HM)
      7 UCLA A 10-5-1 NCT (7th Place)
      8 Stanford A 7-8-1 NCT
      9 Ohio State A 10-5-1 NCT (5th Place)
      10 Harvard A 4-3-1 ORCS
      11 Cornell A 11-4-1 NCT (3rd Place)
      12 NYU A 5-2-1 ORCS
      13 Chicago A 11-5 NCT (3rd Place)
      14 Indiana A 5-3 ORCS*
      15 UC Irvine A 5-3 ORCS
      16 Rhodes B 8-8 NCT (HM)
      17 Arizona A 5-3 ORCS
      18 Emory A 9-5-2 (6th Place)
      19 Howard A 7-9 NCT
      20 Northwestern A 9-7 NCT (9th Place)
      21 Columbia A 10-5-1 NCT (6th Place)
      22 Florida A 5-11 NCT
      23 California (Berkeley) A 10-6 NCT (9th Place)
      24 Patrick Henry A 11-4-1 NCT (4th Place)
      25 Rutgers A 3-5 ORCS
      26 Tufts A 8-8 NCT (HM)
      27 Yale B 4-4 ORCS
      28 Rochester A 4-12 NCT
      29 Cornell B 6-9-1 NCT
      30 Furman A 4-4 ORCS
      31 Virginia B 5-3 ORCS
      32 American A 3-5 ORCS*
      33 Washington & Lee A 3-5 ORCS
      34 Washington-St. Louis A 4-4 ORCS
      35 Richmond A 4-3-1 ORCS
      36 Brown A 4-3-1 ORCS
      36 Penn State A 4-12 NCT
      38 Wheaton A 5-3 ORCS
      39 Duke A 11-5 NCT (4th Place)
      40 Fordham-Linc Ctr A 7-8-1 NCT
      41 Florida B 4-4 ORCS
      42 Michigan B 3-5 ORCS
      43 NYU B 3-5 ORCS
      44 Georgia A 8-8 NCT (HM)
      45 Wesleyan A 9-7 NCT (8th Place)
      46 Boston University A 9-6-1 NCT (7th Place)
      47 Baylor A 3-4-1 ORCS
      48 Kansas A 4-4 Regionals
      49 UC Davis A 5-11 NCT
      50 Iowa A 4-4 ORCS

      *Where neither A nor B team qualified for the NCT and the B team received the better record, the B team's record was used.

      Comment


      • #4
        I've been giving this proposal a lot of thought. Because while decreasing the 'randomness factor' in mock trial is extremely appealing, I think this system takes away one of the things I find most fun about competing- the idea that how you do determines who you go up against. For this TPR- centered placement to work, the schedule would need to be pre-determined. And what's the fun in that? Some of the best, and most terrifying moments in mock trial come from finding out who's next. It's not like no one loses in this TPR-determined ORCs, there are still the same number of teams going home for good, the same amount of devastating missteps, of judge-fuckups, etc that are all inevitable. So is a more even CS worth taking away the 'whos next' element from ORCs? I don't know.

        Comment


        • #5
          Originally posted by EgregiousInvention View Post
          I've been giving this proposal a lot of thought. Because while decreasing the 'randomness factor' in mock trial is extremely appealing, I think this system takes away one of the things I find most fun about competing- the idea that how you do determines who you go up against. For this TPR- centered placement to work, the schedule would need to be pre-determined. And what's the fun in that? Some of the best, and most terrifying moments in mock trial come from finding out who's next. It's not like no one loses in this TPR-determined ORCs, there are still the same number of teams going home for good, the same amount of devastating missteps, of judge-fuckups, etc that are all inevitable. So is a more even CS worth taking away the 'whos next' element from ORCs? I don't know.
          I'm not certain how I will vote on this proposal, or exactly what implementation would look like if it passes, but my understanding is the schedule would not be truly pre-determined. The first round pairing could still be random (e.g. random A pod team vs. random D pod team). There could still be some measure of result-dependency in terms of which teams within each pod meet later on. In other words, perhaps the A pod and B pod are scheduled to meet in Round 3. You could have a high-high pair where the overall best team in the A group (based on wins/CS/PD after R1 and R2 results) meets the overall best team in the B group, or you could have a power-protect pair, or something else. TPR would guide who gets placed into each group or pod but the individual matchups could still be decided based on the results. If the proposal passes it would just have to be decided exactly it would work.
          I post in my personal capacity, not on behalf of AMTA.

          Comment


          • #6
            Originally posted by MizzouMock View Post

            I'm not certain how I will vote on this proposal, or exactly what implementation would look like if it passes...If the proposal passes it would just have to be decided exactly it would work.
            This, right here, is my biggest problem with this proposal (and a number of other proposals on this agenda). Ideally, we’d have a specific, detailed description of how this new system would work. The system would be tried out at a bunch of invites this fall, and a proposal to adopt it would then be submitted next year. That proposal would include detailed information on how the system had played out over a good number of tournaments. Then we’d all be in a position to comment intelligently on the merits (or demerits) of the new system.

            Instead, the board is being asked to vote on a system which isn’t yet fully specified (even the Tab Director doesn’t know all the details a week before the meeting). And even if the board is given full details at the meeting, the community at large hasn’t been given them in time to submit comments (and, frankly, if the board get the details at the meeting, they won’t have time to sit down and work the simulations to see whether those details make sense).

            This is precisely what people mean when they talk about a lack of transparency on the board. Motions without specifics are not fair to anyone.

            Comment


            • #7
              Originally posted by LogicalNitpicker View Post

              This, right here, is my biggest problem with this proposal (and a number of other proposals on this agenda). Ideally, we’d have a specific, detailed description of how this new system would work. The system would be tried out at a bunch of invites this fall, and a proposal to adopt it would then be submitted next year. That proposal would include detailed information on how the system had played out over a good number of tournaments. Then we’d all be in a position to comment intelligently on the merits (or demerits) of the new system.

              Instead, the board is being asked to vote on a system which isn’t yet fully specified (even the Tab Director doesn’t know all the details a week before the meeting). And even if the board is given full details at the meeting, the community at large hasn’t been given them in time to submit comments (and, frankly, if the board get the details at the meeting, they won’t have time to sit down and work the simulations to see whether those details make sense).

              This is precisely what people mean when they talk about a lack of transparency on the board. Motions without specifics are not fair to anyone.
              This is a misunderstanding of what the Board does in many cases. Often times a motion--like this one--is for the Board to decide whether to approve further implementation of a particular policy, idea, or concept. If that happens, the Board will often refer to a committee of the Board to develop specific language, procedures, etc to accomplish that policy or concept. That final product then comes back to the Board for final approval at the midyear meeting or the following year. There is no use in spending time debating logistical details if the Board as a whole isn't in favor of the general concept. And, it's usually not a worthwhile exercise for a board of 30 or so people to get into the weeds of those fine level details when there's a limited amount of time and other discussions to be had.
              I post in my personal capacity, not on behalf of AMTA.

              Comment


              • #8
                I will just say that I think a simpler proposal that is more straightforward is to simply equalize the first round pairing process. The way I propose to do this (and I believe I have said this before on perjuries) is to do a snake design similar to R3 pairings based soley on TPR, THEN you would take the two groups and separate them. Instead of a totally random draw, the rep from the tournament at openings would choose one name from one hat, then one name from the other. This way the TPR is evenly balanced between the Defense and P side during round 1 and round 2. I believe one of the main issues with our system is that you can theoretically have all the top teams all on the same side in Round 1 and Round 2 and be unable to hit each other. A simple coin flip (same as Round 3) would suffice to determine which side is which. I think this maintains the elements of randomness to the pairings that people seem to like, while also ensuring more balance exists in the pairing process.

                Just to explain this for more visual folks:
                Lets say I took the temple NCT division: my initial snake would look like:
                Side 1:
                1) UVA
                4) OSU A
                5) Chicago
                8) Northwestern A
                9) Tufts
                12) Duke
                13) Wesleyan A
                16) Minnesota
                17) UCLA B
                20) UNC
                21) Miami B
                24) St. Thomas

                Side 2:
                2) Rhodes A
                3) GT
                6) Emory A
                7) Howard
                10) Rochester
                11) Cornell B
                14) BU
                15) Northwood
                18) BC
                19) UCSD
                22) UGA B
                23) Texas A&M

                A random draw would create the 12 rounds, then a coin would be flipped (Heads: side 1 is defense, Tails: side 2 is defense).

                To be perfectly clear, I am not matching up UVA and Rhodes round 1, I am saying take side 1 and side 2 and do random draws from those piles.
                Last edited by Adevans; July 12th, 2019, 02:58 AM.

                Comment

                Working...
                X