College Policy Debate Forums
April 20, 2018, 04:06:55 PM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
News: IF YOU EXPERIENCE PROBLEMS WITH THE SITE, INCLUDING LOGGING IN, PLEASE LET ME KNOW IMMEDIATELY.  EMAIL ME DIRECTLY OR USE THE CONTACT US LINK AT THE TOP.
 
   Home   Help Search Login Register CEDA caselist Debate Results Council of Tournament Directors Edebate Archive  
Pages: [1]
  Print  
Author Topic: Kentucky Octo pairing Frosh Qtrs pairing Doubles results  (Read 5507 times)
glarson
Sr. Member
****
Posts: 477


« on: October 08, 2012, 09:08:46 AM »

attached

* Kentucky 2012 OCTOS Frosh QTRs.pdf (9.81 KB - downloaded 1067 times.)
* Kentucky Doubles.pdf (10.1 KB - downloaded 872 times.)
Logged
ScottElliott
Full Member
***
Posts: 148


« Reply #1 on: October 08, 2012, 09:35:24 AM »

Wow. Wow. Looked at the Double Octofinal results. Can you say Affirmative side bias?
Logged
Chris Crowe
Newbie
*
Posts: 26


« Reply #2 on: October 08, 2012, 09:57:47 AM »

Not quite as bad as that damn "higher seed bias."
Logged

Chris Crowe
ScottElliott
Full Member
***
Posts: 148


« Reply #3 on: October 08, 2012, 11:35:19 AM »

True. But, it seems a lot of teams are choosing Aff.
Logged
brubaie
Jr. Member
**
Posts: 77


« Reply #4 on: October 08, 2012, 01:10:57 PM »

I often think "side bias" claims are overblown.

We can't derive that more teams flipped a particular way from the available info (Scott's conclusion.) All 16 teams that won the debate could've lost the flip, the other side selected Neg and the Aff just happened to win most of them.

What we can do is control for skill and widen the data set by looking at the prelim win-loss records for the elimination-round qualifiers.

Elimination-round qualifiers went a collective 191-65. 96 of those wins went to the Neg and 95 went to the Aff.

An identical number of teams were "very" strong on one side (5 teams won 4 Aff debates and 2 Neg debates, 5 teams won 4 Neg debates and 2 Aff debates.) There were 7 teams who were "slightly" stronger on the Neg than the Aff (4 Neg wins, 3 Aff wins or 3 Neg wins, 2 Aff wins) while there 6 of those on the Aff. There were 9 teams who won an equal number (3) on each side of the resolution.

I am as hopeful as anybody for an Aff renaissance but it doesn't appear that one is in the offing at Kentucky. That the 2A has managed to battle to an equal share of victories in the era of Neg traveshamockery is impressive enough.

There is, however, a youth renaissance. Is Peyton the only Senior in the Quarters?

* UKWinsBySide2012.xlsx (11.65 KB - downloaded 255 times.)
« Last Edit: October 08, 2012, 01:16:56 PM by brubaie » Logged
ScottElliott
Full Member
***
Posts: 148


« Reply #5 on: October 08, 2012, 01:17:32 PM »

I am glad to see the data for the people that made it to elims. What about for the whole tournament?
Logged
brubaie
Jr. Member
**
Posts: 77


« Reply #6 on: October 08, 2012, 01:26:27 PM »

I don't plan to construct a data set that large but I would guess the results would be similar.

I also think a smaller, more targeted control group is more useful. The greatest challenge in figuring out a "side bias" that exists independent of debater skill-level (unarguably the primary determinant of who wins debates, as Crowe points out) is to isolate a smaller control group of similarly-skilled debaters.  Limiting the results to elim participants accomplishes this goal.
Logged
joe
Jr. Member
**
Posts: 73


« Reply #7 on: October 08, 2012, 02:23:29 PM »

I am glad to see the data for the people that made it to elims. What about for the whole tournament?

I'd like to see it broken down by SMR's, Bataille, and Other.
Logged
ScottElliott
Full Member
***
Posts: 148


« Reply #8 on: October 08, 2012, 02:54:55 PM »

Seems to me that with power matching, a sample size of 1,000 debates (8 rounds at kentucky times 140+) would be a better determinant of side bias. Obviously I am not a statistician, but it would seem to be pretty easy to compile the data set and find the results.

That would account for the fact that elim rounds are matched power protect. Wouldn't rounds 2-6 be a better determinant of side bias?

Logged
brubaie
Jr. Member
**
Posts: 77


« Reply #9 on: October 08, 2012, 03:59:57 PM »

I agree that prelims are more useful for examination than elims because teams are more closely-matched. My data is based on preliminary debate performance against like opponents. Your original observation was based on the outcome of elim debates.

The difference is whether we gather data from EVERY team that debated. Since the skill gap in the top and bottom of the 3-5 bracket is enormous, I focused solely on elim teams, where the gaps in skill on day 2 are usually less pronounced. If the opposite model is easy to construct then someone could it but it wouldn't be easy and I don't think it would be quite as useful.  
Logged
ScottElliott
Full Member
***
Posts: 148


« Reply #10 on: October 08, 2012, 05:25:20 PM »

I disagree that the skill gas for the middle and bottom of the bracket at Kentucky or Wake is the large. But, it does not matter that much, one way or the other.
Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2013, Simple Machines
SMF customization services by 2by2host.com
Valid XHTML 1.0! Valid CSS!