No announcement yet.

2014 Richard Calkins Invitational Tab Summary

  • Filter
  • Time
  • Show
Clear All
new posts

  • 2014 Richard Calkins Invitational Tab Summary

    Thank you to all the teams that attended. A full tab summary is attached.

    Final Rankings (Balancing Weighted Partial Ballots)
    1 1548 Iowa A (37.878)
    2 1103 Cornell A (34.878)
    3 1104 Cornell B (33.722)
    4 1118 Washburn A (31.747)
    5 1549 Iowa B (31.208)
    6 1105 Cornell C (30.345)
    Honorable Mention
    7 1119 Washburn B (28.236)
    8 1366 Drake A (27.929)
    9 1106 Creighton A (27.395)
    10 1266 Loras A (26.425)

    A full explanation of the tabulation methods is also attached.
    Attached Files
    Last edited by captainbowtie; October 29th, 2014, 08:26 AM.
    Ray Barr (the Amazing)
    "Bow ties are cool."

  • #2
    I'll fully admit that I use the straight-line system I do because I don't have the mathematical know-how to do what you or Ben Graham are doing. But I am curious as to why you think that your logarithmic curve is optimal for this type of tournament, rather than Yale's. I cannot count the number of times I have seen a team get absolutely thumped and only lose a ballot by three or four points. An assumption that a -3 should be treated essentially the same as a tie, which is treated essentially the same as a +3 seems utterly counterintuitive to me.


    • #3
      I think a four-piece curve makes more sense because it treats +1, -1 scores as being more analogous to a tie as compared to a two-piece curve. Curves with a plateau in middle still give close to .5 ballots for such scores, which I think is appropriate considering that a change of one point would result in the team tieing the ballot.

      I certainly see your point about a +3/-3 ballot having the potential to be a blowout. I think the answer to that, however, is to adjust the Base Quality Win, narrowing the points needed for a complete win, so that +3 is worth more to the winning team and -3 is worth less to the losing team. Fourteen points seems a rather large requirement for a complete win, and perhaps +7 or +10 would be more appropriate.

      If the Base Quality Win was set at 7, then a ballot of +3/-3 would fall right in the high slope area of a four-piece curve, where the partial ballots won begin to rapidly drop off.

      In summary: 1) Four-piece curve makes sense because of how it treats almost ties more like ties than a two-piece curve does. 2) Treatment of +3/-3 scores suggests a need to adjust the width of the curve (by adjusting the quality win), not modifying the shape of the curve.
      Ray Barr (the Amazing)
      "Bow ties are cool."