It strikes me that two issues are actually involved in labeling the current system as "broken." The first is the compression which the 50 point scale attempts to resolve by reassigning values under the encouragement to broaden what a particular parameter of numbers might mean (which involves, obviously, getting debaters to realize that they may have actually been a 43/26 etc.). The second issue is hidden in the first of compression. The compression of points have created virtual ties in speakers and impacted breaks in part, at least, because judges may assign the same points. If judges continue to compress the 50 point scale (which I think they will do--see Larsen's posts on this subject last year or the year before regarding the norm to move to a 5 point scale whether it is 29 to 27 or A to B-) then the new scale will be readjusted in due time. I'll add a third issue, discussed last year (I think), which is that of outliers. Thus, I would advocate inverting the current system and/or creating a different means to assess speaker ability, and thereby team ranks/seedings, than simply the creation of a new (but functionally same) scale. For example, we (and by we I mean someone with expertise/comprehension in math) could create an algorithm to prioritize a combination of factors. Given the "problem" of judges awarding the same speaker points to each debater in a debate, scale compression, and outliers, ranks seems to provide an initial way out of the dilemma. Obviously, however, ranks suggest similarity across debates that are likely not similar. So, the formula could account for a variety of factors--how difficult the opposition was, how well the opposition spoke, existing speaker point scales and judge variance. Basically, each debate is ranked in certain variables--w/l, ranks (1-4), and speaker points, but each team/debater can be ranked in a number of other ways including the combination or differential preferencing (outside of rank such as h/l points, ranks, double h/l) that employs the existing data (just differently). Derek Buescher Puget Sound
________________________________ From: [EMAIL PROTECTED] on behalf of Ross Smith Sent: Thu 11/1/2007 11:34 AM To: edebate; [EMAIL PROTECTED] Subject: [CEDA-L] 50 pts. AT Hoe and Ellis Good points (no pun intended). The emporer has no clothes, Andy. Which 5-3's will clear at Wake is random as it is (at least with regard to the differences between the bottom 5-3's that do clear versus the top 5-3's who just miss clearing. We *pretend* that it is meaningful. Yes, there will be a new form of randomness introduced. At least there is some *potential* in the new system for points to have meaning in relation to performance. Making 20-30 "work" has been tired and failed, Josh. One hypothesis I have is that no judge wants to give, and debaters do not want to recieve, less than a 27. 20-26 is *percieved* as (and therefore "means") BAD. There is no reason 40-45 needs to be seen in that same way. Choosing a scale that is not a multiple of 30 is part of that. This discussion is another. At least we have an *opportunity) to create some breathing room in our evaluations. -- Ross K. Smith Director of Debate Wake Forest University 336-251-2076 (c) 336-758-5268 (o) http://groups.wfu.edu/debate/ http://www.DebateScoop.org <http://www.debatescoop.org/> _______________________________________________ CEDA-L mailing list [email protected] http://www.ndtceda.com/mailman/listinfo/ceda-l
_______________________________________________ CEDA-L mailing list [email protected] http://www.ndtceda.com/mailman/listinfo/ceda-l
