EFT 2018 - Detailed stats

Old college threads.
Locked
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

EFT 2018 - Detailed stats

Post by ryanrosenberg »

Here are links to detailed stats spreadsheets for EFT. If you see any errors, please let me know via email or PM.

Online
Maryland
Michigan
Minnesota
WUSTL
Yale
UCF
Georgia Tech
Carleton
Boise State
Oxford
Edinburgh
Stanford
Virginia Tech

Combined all buzzes
Combined tossup summary
Combined bonuses

Tossup graphs

Overall and category BPA
Team BPA plots

Subcategory PPBs

Please use this thread for comments and questions about the stats, or for additional statistical analysis using this data.
Last edited by ryanrosenberg on Tue Nov 13, 2018 10:36 pm, edited 12 times in total.
Ryan Rosenberg
North Carolina '16
ACF
Mnemosyne
Wakka
Posts: 174
Joined: Tue May 31, 2011 12:11 am

Re: EFT 2018 - Detailed stats

Post by Mnemosyne »

You might want to make sure all the spreadsheets are view-only. I think I changed (and unchanged) something while trying to ctrl + f.
Nick Collins
C. E. Byrd '12 (Shreveport, LA)
Louisiana Tech '16, '17
University of Virginia
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

I've updated the combined files above with data from this past weekend's mirrors (Minnesota and WUSTL).
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

I've also added a link to a folder with graphs like the one below for each tossup.

Image
Ryan Rosenberg
North Carolina '16
ACF
User avatar
a bird
Wakka
Posts: 164
Joined: Sun Feb 26, 2012 3:50 pm
Location: College Park, MD

Re: EFT 2018 - Detailed stats

Post by a bird »

Geriatric trauma wrote: Tue Oct 09, 2018 12:51 pm I've also added a link to a folder with graphs like the one below for each tossup.
Thanks for making these! They seem like a really useful tool for seeing which questions generated lots of negs, which were powered very often, etc. Hopefully this will help writers and editors better anticipate how questions will play in future.

It might also be interesting to seem similar plots aggregated by subcategory to see how difficulty changed from category to category. (I'm not sayin this needs to happen, but it might be good for someone looking for a data visualization project.)
Graham R.

Maryland
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

a bird wrote: Wed Oct 10, 2018 5:07 pm It might also be interesting to seem similar plots aggregated by subcategory to see how difficulty changed from category to category. (I'm not sayin this needs to happen, but it might be good for someone looking for a data visualization project.)
World History, European History, Short Fiction, and Current Events were the easiest categories over the course of a tossup (mostly in the second half), and Biology and Music were the hardest (Music mostly in the first half of the tossup, Bio in the late-middle clues).

Image
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

I made graphs of how many correct buzzes a team got (as a percentage of possible buzzes) at each point in the tossup. Here's a graph highlighting the top 4 teams by average tossup conversion at each site using these graphs. You can see, for instance, the gap between the top 6 teams (highlighted ones plus Delaware and Virginia A) and the rest of the field at Maryland, or the gap between Ohio State A and Michigan State A and everyone else.

Image
Ryan Rosenberg
North Carolina '16
ACF
User avatar
AGoodMan
Rikku
Posts: 372
Joined: Sat Dec 20, 2014 10:25 pm

Re: EFT 2018 - Detailed stats

Post by AGoodMan »

I played the entirety of the set (save for one bye round) at the Yale site yesterday, and I had a lot of fun. I'm looking forward to taking a peek at the detailed stats for our site when they become available.
Jon Suh
Wheaton Warrenville South High School '16
Harvard '20
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

AGoodMan wrote: Sun Oct 14, 2018 6:31 pm I played the entirety of the set (save for one bye round) at the Yale site yesterday, and I had a lot of fun. I'm looking forward to taking a peek at the detailed stats for our site when they become available.
I've updated the OP with links to this past weekend's site-specific stats. The combined files are also now updated.
Ryan Rosenberg
North Carolina '16
ACF
Votre Kickstarter Est Nul
Rikku
Posts: 365
Joined: Wed May 06, 2015 2:09 pm

Re: EFT 2018 - Detailed stats

Post by Votre Kickstarter Est Nul »

Geriatric trauma wrote: Sun Oct 14, 2018 9:28 pm
AGoodMan wrote: Sun Oct 14, 2018 6:31 pm I played the entirety of the set (save for one bye round) at the Yale site yesterday, and I had a lot of fun. I'm looking forward to taking a peek at the detailed stats for our site when they become available.
I've updated the OP with links to this past weekend's site-specific stats. The combined files are also now updated.
No rush at all but just wondering if the TU graphs are getting updated as well.
Emmett Laurie
East Brunswick '16
Rutgers University '21
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

The Billiards Fool wrote: Sun Oct 14, 2018 11:46 pm
Geriatric trauma wrote: Sun Oct 14, 2018 9:28 pm
AGoodMan wrote: Sun Oct 14, 2018 6:31 pm I played the entirety of the set (save for one bye round) at the Yale site yesterday, and I had a lot of fun. I'm looking forward to taking a peek at the detailed stats for our site when they become available.
I've updated the OP with links to this past weekend's site-specific stats. The combined files are also now updated.
No rush at all but just wondering if the TU graphs are getting updated as well.
Yep, they should be updated now.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

I've calculated BPA and category BPA for EFT, you can find the stats here.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

I've also created team BPA plots like the one below, and uploaded them here.

Image
Ryan Rosenberg
North Carolina '16
ACF
User avatar
t-bar
Tidus
Posts: 671
Joined: Sun Jan 25, 2009 4:12 pm

Re: EFT 2018 - Detailed stats

Post by t-bar »

The Midwest, it seems, is the land of generalists. Here are the top 6 players across all sites ranked by their third-best category BPA:

1. Jakob Myers (Michigan State): History 38.979, Other 29.000, Literature 23.896
2. Matthew Lehmann (Chicago): History 24.932, Literature 23.091, Thought 21.955
3. Chris Ray (Ohio State): History 32.312, Other 22.417, Thought 21.583
4. Alston Boyd (Chicago): Thought 29.000, Beliefs 22.045, Literature 19.500
5. Eric Wolfsberg (Delaware): Thought 28.636, History 19.182, Other 19.182
6. Kai Smith (Chicago): Fine Arts 33.256, Science 24.096, Literature 19.154
Stephen Eltinge
Then: TJ, MIT, Yale, PACE, NAQT
Now: ACF
A Dim-Witted Saboteur
Yuna
Posts: 973
Joined: Tue Aug 02, 2016 12:31 pm
Location: Indiana

Re: EFT 2018 - Detailed stats

Post by A Dim-Witted Saboteur »

I made a list of teams and people who BPA indicates are very good at various categories (players with single category BPA above 25, teams with single category BPA above 30) (apologies for the formatting; this is copy-pasted from Excel).

Brad McLain (Illinois A) Illinois A 32.955 Beliefs
Kai Smith (Chicago Orange) Chicago Orange 33.256 Fine Arts
Jakob (Michigan State A) Michigan State A 38.979 History
Chris R (Ohio State A) Ohio State A 32.312 History
Geoffrey Chen (Dragonball Z) Dragonball Z 37.75 Science
Kevin Wang (Amherst A) Amherst A 28.023 Science
Brad McLain (Illinois A) Illinois A 30.682 Thought
Alston Boyd (Chicago Purple) Chicago Purple 29 Thought
Eric Wolfsberg (Delaware) Delaware 28.636 Thought
Caleb Kendrick (Maryland A) Maryland A 27.269 Thought
Jakob (Michigan State A) Michigan State A 29 Other
Brice Fincher (UGA B) UGA B 27.545 Other

Illinois A 32.955 Beliefs
Harvard A 30.545 Beliefs
UGA B 30.182 Beliefs
Chicago Orange 34.179 Fine Arts
Ohio State A 33.417 Fine Arts
Harvard A 31.818 Fine Arts
Michigan State A 44.312 History
Penn A 34.083 History
Ohio State A 33.479 History
Chicago Purple 33.205 History
Illinois A 39.295 Literature
Ohio State A 39.208 Literature
Michigan State A 33.958 Literature
UF Guanaco (A) 33.125 Literature
UGA A 30.205 Literature
Michigan State A 33.167 Other
Hopkins A 32 Other
Toronto J 31.182 Other
Dragonball Z 39.278 Science
Amherst A 32.409 Science
Ohio State A 47.208 Thought
Illinois A 35.5 Thought
Maryland A 35.346 Thought
UF Guanaco (A) 34.65 Thought
Chicago Purple 33.045 Thought
Michigan State A 30.958 Thought
Maryland B 30.864 Thought
Jakob M. (they/them)
Michigan State '21, Indiana '2?
"No one has ever organized a greater effort to get people interested in pretending to play quiz bowl"
-Ankit Aggarwal
User avatar
a bird
Wakka
Posts: 164
Joined: Sun Feb 26, 2012 3:50 pm
Location: College Park, MD

Re: EFT 2018 - Detailed stats

Post by a bird »

An Economic Ignoramus wrote: Thu Oct 18, 2018 3:19 pm I made a list of teams and people who BPA indicates are very good at various categories (players with single category BPA above 25, teams with single category BPA above 30) (apologies for the formatting; this is copy-pasted from Excel).

Brad McLain (Illinois A) Illinois A 32.955 Beliefs
Kai Smith (Chicago Orange) Chicago Orange 33.256 Fine Arts
Jakob (Michigan State A) Michigan State A 38.979 History
Chris R (Ohio State A) Ohio State A 32.312 History
Geoffrey Chen (Dragonball Z) Dragonball Z 37.75 Science
Kevin Wang (Amherst A) Amherst A 28.023 Science
Brad McLain (Illinois A) Illinois A 30.682 Thought
Alston Boyd (Chicago Purple) Chicago Purple 29 Thought
Eric Wolfsberg (Delaware) Delaware 28.636 Thought
Caleb Kendrick (Maryland A) Maryland A 27.269 Thought
Jakob (Michigan State A) Michigan State A 29 Other
Brice Fincher (UGA B) UGA B 27.545 Other

Illinois A 32.955 Beliefs
Harvard A 30.545 Beliefs
UGA B 30.182 Beliefs
Chicago Orange 34.179 Fine Arts
Ohio State A 33.417 Fine Arts
Harvard A 31.818 Fine Arts
Michigan State A 44.312 History
Penn A 34.083 History
Ohio State A 33.479 History
Chicago Purple 33.205 History
Illinois A 39.295 Literature
Ohio State A 39.208 Literature
Michigan State A 33.958 Literature
UF Guanaco (A) 33.125 Literature
UGA A 30.205 Literature
Michigan State A 33.167 Other
Hopkins A 32 Other
Toronto J 31.182 Other
Dragonball Z 39.278 Science
Amherst A 32.409 Science
Ohio State A 47.208 Thought
Illinois A 35.5 Thought
Maryland A 35.346 Thought
UF Guanaco (A) 34.65 Thought
Chicago Purple 33.045 Thought
Michigan State A 30.958 Thought
Maryland B 30.864 Thought
I don't want to be overly negative, since BPA is an innovative way to use advanced stats data, but I want to point out the shortcomings of lists like this. Ryan did a good job explainings the shortcomings of BPA in this thread, but I want to emphasize that category dependence is also a limitation of the stat in its current form.

Not only does BPA share the field dependence and opponent dependence of ppg, but the category specific BPA's depend on the relative difficulty of each category. Ryan's average buzzpoint plots upthread show that how difficulty changed significantly between categories (at least relative to the collective field). Because of this category dependence, it doesn't necessarily mean anything to compare Brad McLain's belief BPA to Kai Smith's on art. Comparing players within a category is more meaningful, especiialy if they played the same field (though there could still subcategory dependence issues).
Graham R.

Maryland
A Dim-Witted Saboteur
Yuna
Posts: 973
Joined: Tue Aug 02, 2016 12:31 pm
Location: Indiana

Re: EFT 2018 - Detailed stats

Post by A Dim-Witted Saboteur »

Sorry if this got confused; I wasn't trying to imply that you can make meaningful cross- category comparisons with this; using a single threshold across all categories was a somewhat arbitrary decision.
Jakob M. (they/them)
Michigan State '21, Indiana '2?
"No one has ever organized a greater effort to get people interested in pretending to play quiz bowl"
-Ankit Aggarwal
User avatar
Lake Winnipesaukee Mystery Stone
Wakka
Posts: 169
Joined: Tue Nov 06, 2012 7:24 am

Re: EFT 2018 - Detailed stats

Post by Lake Winnipesaukee Mystery Stone »

Could I get a link to the Oxford site stats?
Oliver Clarke
King Edward's School, Birmingham '11
Oxford '16
St Andrews '18
Oxford '23
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

The Abydos Helicopter wrote: Sat Oct 20, 2018 7:13 pm Could I get a link to the Oxford site stats?
Oxford and Edinburgh have been added. If someone knows what the Briticizing process was for these sites and can let me know, I'd appreciate it; I'm going to hold off on including that version in combined stats until that's squared away.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

Geriatric trauma wrote: Sun Oct 21, 2018 8:10 pm
The Abydos Helicopter wrote: Sat Oct 20, 2018 7:13 pm Could I get a link to the Oxford site stats?
Oxford and Edinburgh have been added. If someone knows what the Briticizing process was for these sites and can let me know, I'd appreciate it; I'm going to hold off on including that version in combined stats until that's squared away.
The combined stats have been updated with British sites (less the questions that were changed in Briticizing), let me know if anything seems missing.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
Lake Winnipesaukee Mystery Stone
Wakka
Posts: 169
Joined: Tue Nov 06, 2012 7:24 am

Re: EFT 2018 - Detailed stats

Post by Lake Winnipesaukee Mystery Stone »

Geriatric trauma wrote: Fri Oct 26, 2018 11:54 am
Geriatric trauma wrote: Sun Oct 21, 2018 8:10 pm
The Abydos Helicopter wrote: Sat Oct 20, 2018 7:13 pm Could I get a link to the Oxford site stats?
Oxford and Edinburgh have been added. If someone knows what the Briticizing process was for these sites and can let me know, I'd appreciate it; I'm going to hold off on including that version in combined stats until that's squared away.
The combined stats have been updated with British sites (less the questions that were changed in Briticizing), let me know if anything seems missing.
Thank you!

Some of the links appear to have gotten mixed up - Tossup graphs leads to team bpa plots now, and Team BPA plots to team buzz graphs with only one file in?
Oliver Clarke
King Edward's School, Birmingham '11
Oxford '16
St Andrews '18
Oxford '23
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

The Abydos Helicopter wrote: Fri Oct 26, 2018 1:11 pm
Geriatric trauma wrote: Fri Oct 26, 2018 11:54 am
Geriatric trauma wrote: Sun Oct 21, 2018 8:10 pm
The Abydos Helicopter wrote: Sat Oct 20, 2018 7:13 pm Could I get a link to the Oxford site stats?
Oxford and Edinburgh have been added. If someone knows what the Briticizing process was for these sites and can let me know, I'd appreciate it; I'm going to hold off on including that version in combined stats until that's squared away.
The combined stats have been updated with British sites (less the questions that were changed in Briticizing), let me know if anything seems missing.
Thank you!

Some of the links appear to have gotten mixed up - Tossup graphs leads to team bpa plots now, and Team BPA plots to team buzz graphs with only one file in?
Should be fixed now.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

Links to Stanford and VT sites have been added in; combined stats will take a day or two on my end.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
ryanrosenberg
Auron
Posts: 1890
Joined: Thu May 05, 2011 5:48 pm
Location: Palo Alto, California

Re: EFT 2018 - Detailed stats

Post by ryanrosenberg »

Geriatric trauma wrote: Mon Nov 12, 2018 10:37 pm Links to Stanford and VT sites have been added in; combined stats will take a day or two on my end.
These should be updated; let me know if anything looks off.
Ryan Rosenberg
North Carolina '16
ACF
User avatar
hftf
Lulu
Posts: 62
Joined: Sat Apr 26, 2008 7:28 pm

Re: EFT 2018 - Detailed stats

Post by hftf »

Static visualizations are now available.

Because EFT isn’t clear, access is restricted (username: alberto, password: nepomuceno). You must obtain access via this private forum.

Please continue using this private forum for all discussion of EFT. Note that there are likely many mistakes in the data (esp. at UK sites).


If you find this tool valuable and would like to support me creating more quizbowl technology, please consider donating to my Patreon.
Ophir
Maryland
Walter Johnson
Locked