2011 PACE NSC Tournament Format

Dormant threads from the high school sections are preserved here.
User avatar
Gautam
Auron
Posts: 1413
Joined: Sun Feb 11, 2007 7:28 pm
Location: Zone of Avoidance
Contact:

2011 PACE NSC Tournament Format

Post by Gautam » Mon Apr 04, 2011 7:12 pm

Good Evening,

After deliberating about potential field sizes, PACE has decided to run a sixty-team format for the NSC this year. There will be three stages of play.

Stage 1: 10 brackets of 6 teams, 5 games for all teams. To be played Saturday morning.

The sixty teams will be broken down into 10 evenly seeded brackets of 6 teams. The top two teams in each bracket will automatically advance to the top tier during the second stage of play. In addition, four wildcard teams will advance to the top tier. Each of the ten third-place teams from the preliminary brackets will be eligible to advance as wildcards. All ties (including ties for third place, and thus, eligibility for the wildcard procedure) will be broken on minimatches.

Stage 1a: The wildcard procedure (10 third-place teams). To be played Saturday during lunch.

The ten third-place teams will be seeded 1-10 based on bonus conversion. Any ties will be broken using bounceback percentage (or a coinflip if still tied). Four teams (seeds 1 v. 8, 2 v. 7) will play single matches on a full packet to determine two of the wildcards. Four teams (seeds 5 v. 10, 6 v. 9) will play half-packet matches. The winners of those half-packet matches will play the 3 and 4 seeds to determine the other two wildcards. The 3 seed will play the weakest seed that advances, while the 4 seed will play the strongest seed that advances.

Stage 2: 10 brackets of 6 teams; 4 championship-eligible brackets, 3 second-tier brackets, 3 third-tier brackets; 5 games for all teams. To be played Saturday afternoon.

The top 24 teams will be broken into 4 brackets of 6 teams in a snaking fashion, so that all brackets are of even strength. The remaining 16 third and fourth seeds, plus the two strongest fifth seeds in PPB, will be placed in 3 even brackets of 6 teams to make up the second tier, which will consist of teams eligible for 25th-42nd place. The remaining fifth seeds and the sixth seeds will make up the 3 even brackets of the third tier, which will be eligible for 43rd-60th place.

Teams will then play a 5-game round robin within their bracket. In all tiers, the teams will advance in pairs (i.e. the top two seeds in each bracket will advance to the top playoff bracket in each respective tier). In Stage 2, there are 4 top-tier and 3 second- and third-tier brackets; consequently, the brackets in Stage 3 will consist of 3 top-tier brackets of 8 teams, 3 second-tier brackets of 6 teams, and 3 third-tier brackets of 6 teams. All necessary ties will be broken on minimatches.

Stage 3: 3 brackets of 8 teams (top tier), 3 brackets of 6 teams (second and third tiers); 6 games (top tier) or 4 games (second and third tier). To be played Sunday morning.

Team will not replay the game against the common opponent who advances from their Stage 2 bracket; that game will count as already played. Teams will play a full round robin, which will consist of 6 games for top-tier teams or 4 games for second- and third-tier teams.

Finals

The top two teams in the top Stage 3 bracket will play a final unless the top team clears the field by two or more games. The final will be advantaged if one team has a one-game lead on the other, or a single-game, winner-takes-all format if the teams are tied in record.

RULES:

The rules for the 2011 NSC can be found in a DOCX format and a PDF format on the PACE website.

Thank you,
Gautam Kandlikar
Last edited by Gautam on Mon May 23, 2011 1:32 pm, edited 1 time in total.
Gautam - ACF
Currently tending to the 'quizbowl hobo' persuasion.

User avatar
etchdulac
Rikku
Posts: 355
Joined: Thu May 29, 2003 6:02 am
Location: Texas, for better or worse

Re: 2011 PACE NSC Tournament Format

Post by etchdulac » Mon Apr 04, 2011 7:44 pm

So... in a three-way 4-1 tie in the initial group phase, the top PPB team would advance automatically, while the other two would play a half-packet? Would the tiebreaker winner be seeded second going forward from the bracket?
Stephen Fontenot
Texas Quiz Bowl Alliance Deputy Director
Communications, UT Dallas
Strake Jesuit '96 -+-+- Southwestern '00

tabstop
Wakka
Posts: 122
Joined: Tue Mar 20, 2007 10:48 am
Location: NNVA

Re: 2011 PACE NSC Tournament Format

Post by tabstop » Mon Apr 04, 2011 8:21 pm

etchdulac wrote:So... in a three-way 4-1 tie in the initial group phase, the top PPB team would advance automatically, while the other two would play a half-packet? Would the tiebreaker winner be seeded second going forward from the bracket?
I expect in that situation that the top two would play, with the winner taking the top seed; the loser would play the third team for the second spot. (IANAL, IANAMOP, etc.)
Andrew Feist
Treasurer, ACF | VP of Technology, PACE

User avatar
Haaaaaaaarry Whiiiiiiiiiite
Auron
Posts: 1123
Joined: Tue Dec 04, 2007 8:46 pm
Location: Fairfax, VA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Haaaaaaaarry Whiiiiiiiiiite » Mon Apr 04, 2011 8:36 pm

tabstop wrote:
etchdulac wrote:So... in a three-way 4-1 tie in the initial group phase, the top PPB team would advance automatically, while the other two would play a half-packet? Would the tiebreaker winner be seeded second going forward from the bracket?
I expect in that situation that the top two would play, with the winner taking the top seed; the loser would play the third team for the second spot. (IANAL, IANAMOP, etc.)
That is what has happened in the past, so I would assume it would still be the case.
Harry White
TJHSST '09, Virginia Tech '13
Member, PACE
Tournament Database Search by Team
Will run stats for food

User avatar
dtaylor4
Auron
Posts: 3733
Joined: Tue Nov 16, 2004 11:43 am

Re: 2011 PACE NSC Tournament Format

Post by dtaylor4 » Tue Apr 05, 2011 10:03 am

etchdulac wrote:So... in a three-way 4-1 tie in the initial group phase, the top PPB team would advance automatically, while the other two would play a half-packet? Would the tiebreaker winner be seeded second going forward from the bracket?
It would depend on the spots contested by the three teams. If it's for spots 3-4-5, 3 plays 4, with the winner taking the 3. The loser of that plays 5, for the 4-5 spots. If it's 2-3-4, 4 plays 3, loser takes 4, and 2 and 3 play off.

EDIT: These would be on half-packets specifically designated for this purpose. Tiebreaker rules (3/3, then SD) still apply.
Last edited by dtaylor4 on Tue Apr 05, 2011 12:57 pm, edited 1 time in total.

User avatar
Gautam
Auron
Posts: 1413
Joined: Sun Feb 11, 2007 7:28 pm
Location: Zone of Avoidance
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Gautam » Tue Apr 05, 2011 11:43 am

Donald is correct, I'm pretty sure.
Gautam - ACF
Currently tending to the 'quizbowl hobo' persuasion.

User avatar
jonpin
Forums Staff: Moderator
Posts: 2031
Joined: Wed Feb 04, 2004 6:45 pm
Location: BCA NJ / WUSTL MO / Hackensack NJ

Re: 2011 PACE NSC Tournament Format

Post by jonpin » Tue Apr 05, 2011 12:56 pm

dtaylor4 wrote:
etchdulac wrote:So... in a three-way 4-1 tie in the initial group phase, the top PPB team would advance automatically, while the other two would play a half-packet? Would the tiebreaker winner be seeded second going forward from the bracket?
It would depend on the spots contested by the three teams. If it's for spots 3-4-5, 3 plays 4, with the winner taking the 3. The loser of that plays 5, for the 4-5 spots. If it's 2-3-4, 4 plays 3, loser takes 4, and 2 and 3 play off.
In the case of a 3-way tie at 4-1, it'd have to be for spots 1-2-3, but in general yes. My understanding of PACE's tiebreaking regulations are essentially (a) no team is eliminated on statistical tiebreakers and (b) no team is advanced on statistical tiebreakers.

Was any thought given to a system where the lower brackets finish up a little quicker, enable all the players to watch the top bracket's final rounds, as happened last year, or would that have required to significant a change in schedule?
Jon Pinyan
Coach, Bergen County Academies (NJ); former player for BCA (2000-03) and WUSTL (2003-07)
HSQB forum mod, PACE member
Stat director for: NSC '13-'15, '17; ACF '14, '17, '19; NHBB '13-'15; NASAT '11

"A [...] wizard who controls the weather" - Jerry Vinokurov

User avatar
cvdwightw
Auron
Posts: 3446
Joined: Tue May 13, 2003 12:46 am
Location: Southern CA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by cvdwightw » Tue Apr 05, 2011 2:04 pm

jonpin wrote:Was any thought given to a system where the lower brackets finish up a little quicker, enable all the players to watch the top bracket's final rounds, as happened last year, or would that have required to significant a change in schedule?
Speaking only for myself as the primary designer of this system and not in any official capacity:

It is true that a potential alternative to the superplayoff is to hold a three-game round robin among teams with equal placement in each playoff bracket. However, it is my opinion that teams who come to the PACE NSC generally prefer to play more quizbowl rather than less quizbowl. Therefore, time and packets permitting, it is better to increase from 15 guaranteed rounds to 16 rather than to decrease from 15 to 13 (I do recognize that this schedule decreases the guaranteed games for the top eight from 17 to 16, but that also means that the tournament as a whole is one round shorter).

Slightly off topic to that, my favorite part about this format is that it streamlines all tournament logistics: by the time teams head to lunch on Saturday, they should know what playoff bracket they are in; by the time teams head back to the hotel on Saturday, they should know what superplayoff bracket they are in. There is no more awkwardly milling around waiting for tiebreaker games/slow brackets to finish to start the next phase, because this will all be solved over breaks that would have to happen anyway.
Dwight Wynne
socalquizbowl.org
UC Irvine 2008-2013; UCLA 2004-2007; Capistrano Valley High School 2000-2003

"It's a competition, but it's not a sport. On a scale, if football is a 10, then rowing would be a two. One would be Quiz Bowl." --Matt Birk on rowing, SI On Campus, 10/21/03

"If you were my teammate, I would have tossed your ass out the door so fast you'd be emitting Cerenkov radiation, but I'm not classy like Dwight." --Jerry

Aaron Goldfein
Wakka
Posts: 109
Joined: Wed Jan 20, 2010 4:33 pm

Re: 2011 PACE NSC Tournament Format

Post by Aaron Goldfein » Wed Apr 06, 2011 12:37 am

Plus, teams should still be free to view the finals matches.
Aaron Goldfein
Niles West ('10)
Carnegie Mellon ('14)

User avatar
Gautam
Auron
Posts: 1413
Joined: Sun Feb 11, 2007 7:28 pm
Location: Zone of Avoidance
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Gautam » Wed Apr 06, 2011 2:04 pm

In this format, should finals matches be necessary, they will only take place after the superplayoffs are done. As such, we can create (and more importantly stick to) a schedule that allows all participants to convene in the central auditorium to watch the finals matches and collect their hard-earned prizes during the prize-ceremony, etc. A playoff system in which the top contenders play in a different format compared to the rest of the teams could be feasible. However, Dwight's astute observation that this format allows the streamlining of logistics is important, and it feels safer to stick to a format in which all brackets have similar structures.
Gautam - ACF
Currently tending to the 'quizbowl hobo' persuasion.

User avatar
Auroni
Auron
Posts: 2999
Joined: Thu Nov 15, 2007 6:23 pm
Location: Brooklyn

Re: 2011 PACE NSC Tournament Format

Post by Auroni » Wed Apr 06, 2011 2:11 pm

This isn't my position on this format, but I have heard it from some others: this feels like an abnormally low number of prelim games to decide if teams are ultimately in contention or not -- what is Dwight's take on this matter?
Auroni Gupta
UIUC
ACF

User avatar
jonpin
Forums Staff: Moderator
Posts: 2031
Joined: Wed Feb 04, 2004 6:45 pm
Location: BCA NJ / WUSTL MO / Hackensack NJ

Re: 2011 PACE NSC Tournament Format

Post by jonpin » Wed Apr 06, 2011 2:23 pm

cvdwightw wrote:It is true that a potential alternative to the superplayoff is to hold a three-game round robin among teams with equal placement in each playoff bracket. However, it is my opinion that teams who come to the PACE NSC generally prefer to play more quizbowl rather than less quizbowl. Therefore, time and packets permitting, it is better to increase from 15 guaranteed rounds to 16 rather than to decrease from 15 to 13 (I do recognize that this schedule decreases the guaranteed games for the top eight from 17 to 16, but that also means that the tournament as a whole is one round shorter).

Slightly off topic to that, my favorite part about this format is that it streamlines all tournament logistics: by the time teams head to lunch on Saturday, they should know what playoff bracket they are in; by the time teams head back to the hotel on Saturday, they should know what superplayoff bracket they are in. There is no more awkwardly milling around waiting for tiebreaker games/slow brackets to finish to start the next phase, because this will all be solved over breaks that would have to happen anyway.
gkandlikar wrote:In this format, should finals matches be necessary, they will only take place after the superplayoffs are done. As such, we can create (and more importantly stick to) a schedule that allows all participants to convene in the central auditorium to watch the finals matches and collect their hard-earned prizes during the prize-ceremony, etc. A playoff system in which the top contenders play in a different format compared to the rest of the teams could be feasible. However, Dwight's astute observation that this format allows the streamlining of logistics is important, and it feels safer to stick to a format in which all brackets have similar structures.
Sounds good. The streamlining is a very nice benefit, and one I hadn't thought of but always like. Would tiebreakers likely be before dismissal to breaks or via an early call back after lunch or Sunday morning?
Jon Pinyan
Coach, Bergen County Academies (NJ); former player for BCA (2000-03) and WUSTL (2003-07)
HSQB forum mod, PACE member
Stat director for: NSC '13-'15, '17; ACF '14, '17, '19; NHBB '13-'15; NASAT '11

"A [...] wizard who controls the weather" - Jerry Vinokurov

User avatar
cvdwightw
Auron
Posts: 3446
Joined: Tue May 13, 2003 12:46 am
Location: Southern CA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by cvdwightw » Wed Apr 06, 2011 6:19 pm

every time i refresh i have a new name wrote:This isn't my position on this format, but I have heard it from some others: this feels like an abnormally low number of prelim games to decide if teams are ultimately in contention or not -- what is Dwight's take on this matter?
We could have gone with the following format:

9x8 prelims, split 3/3/3 (7 rounds + TB)
12x6 playoffs, split 2/2/2 (5 rounds + TB)
12x6 superplayoffs, play every team you didn't play before (4 rounds)

It doesn't change the number of teams in championship contention after the prelims. It doesn't change the number of required packets or overall games. It gives every team two more prelim games. And you know what? Adding an extra prelim bracket and seeing how things would have to shake out from there was probably the single most obvious thing anyone could have done.

Why didn't we start by doing the obvious thing? I honestly don't have the faintest idea. But the idea of 9 prelim brackets of 8 was never once brought up in internal discussions before I proposed what largely became this format.

Why didn't we change to the obvious thing? Two reasons. The first was that internal discussions resulted in a consensus that the superplayoffs needed to have eight or more teams. The second was that I was so enamored with the logistics streamlining that I didn't bother considering any other format, once proposed, until literally less than an hour ago.
Dwight Wynne
socalquizbowl.org
UC Irvine 2008-2013; UCLA 2004-2007; Capistrano Valley High School 2000-2003

"It's a competition, but it's not a sport. On a scale, if football is a 10, then rowing would be a two. One would be Quiz Bowl." --Matt Birk on rowing, SI On Campus, 10/21/03

"If you were my teammate, I would have tossed your ass out the door so fast you'd be emitting Cerenkov radiation, but I'm not classy like Dwight." --Jerry

User avatar
Mewto55555
Tidus
Posts: 708
Joined: Sat Mar 13, 2010 9:27 pm
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Mewto55555 » Wed Apr 06, 2011 6:49 pm

How do you split brackets of 8 into 3/3/3?
Max
formerly of Ladue, Chicago

mtimmons
Wakka
Posts: 114
Joined: Mon Oct 05, 2009 8:25 pm

Re: 2011 PACE NSC Tournament Format

Post by mtimmons » Wed Apr 06, 2011 9:31 pm

cvdwightw wrote:
every time i refresh i have a new name wrote:This isn't my position on this format, but I have heard it from some others: this feels like an abnormally low number of prelim games to decide if teams are ultimately in contention or not -- what is Dwight's take on this matter?
We could have gone with the following format:

9x8 prelims, split 3/3/3 (7 rounds + TB)
12x6 playoffs, split 2/2/2 (5 rounds + TB)
12x6 superplayoffs, play every team you didn't play before (4 rounds)

It doesn't change the number of teams in championship contention after the prelims. It doesn't change the number of required packets or overall games. It gives every team two more prelim games. And you know what? Adding an extra prelim bracket and seeing how things would have to shake out from there was probably the single most obvious thing anyone could have done.
I don't think that really works. Not only can you not split 8 teams 3/3/3, it also gives 27 teams in championship contention after the first bracket, which messes up the playoffs. 8x9 prelims, split 3/3/3 would work but would require 2 more rounds.

EDIT: Are prelims supposed to be 2/3/3? That seems like it would actually work.
Max Timmons
St. Paul Central High School 2012
MIT 2016

User avatar
AlphaQuizBowler
Tidus
Posts: 695
Joined: Mon Dec 03, 2007 6:31 pm
Location: Alpharetta, GA

Re: 2011 PACE NSC Tournament Format

Post by AlphaQuizBowler » Wed Apr 06, 2011 10:18 pm

So here's my concern: it's bad enough to be relegated to at best 17th place after 7 games. Being relegated to at best 25th after only 5 games? This requires near-perfect bracketing, which, based on the statistics from previous years, PACE (or really any other organization) just really isn't capable of.
William
Alpharetta High School '11
Harvard '15

User avatar
dtaylor4
Auron
Posts: 3733
Joined: Tue Nov 16, 2004 11:43 am

Re: 2011 PACE NSC Tournament Format

Post by dtaylor4 » Wed Apr 06, 2011 10:24 pm

AlphaQuizBowler wrote:So here's my concern: it's bad enough to be relegated to at best 17th place after 7 games. Being relegated to at best 25th after only 5 games? This requires near-perfect bracketing, which, based on the statistics from previous years, PACE (or really any other organization) just really isn't capable of.
Teams are never out of contention until they lose two games.

Also, did you even take a look at last year's preliminary standings?

User avatar
AKKOLADE
Sin
Posts: 15261
Joined: Thu Apr 24, 2003 8:08 am

Re: 2011 PACE NSC Tournament Format

Post by AKKOLADE » Wed Apr 06, 2011 10:42 pm

dtaylor4 wrote: Also, did you even take a look at last year's preliminary standings?
Gonna take a wild guess here and say yes, the competitor from last year's NSC did look at the standings for that tournament.
Fred Morlan
PACE President, 2018-19
International Quiz Bowl Tournaments, co-owner
University of Kentucky CoP, 2017
hsqbrank manager, NAQT writer (former subject editor), former hsqb Administrator/Chief Administrator, 2012 NASAT TD

User avatar
AlphaQuizBowler
Tidus
Posts: 695
Joined: Mon Dec 03, 2007 6:31 pm
Location: Alpharetta, GA

Re: 2011 PACE NSC Tournament Format

Post by AlphaQuizBowler » Wed Apr 06, 2011 10:43 pm

dtaylor4 wrote: Also, did you even take a look at last year's preliminary standings?
I was there. Somehow I don't think that Lawson Bracket and Shinn Bracket were anywhere near as difficult as Davies or Sloan, and I think that's shown by the low finishes of St. Ignatius and Walt Whitman in the top bracket. Or I could put it this way: the top 3 teams in first consolation bracket each had higher PPB than 5 of the bottom 6 teams in the top bracket. Bracketing at the 2010 NSC was far from perfect.
William
Alpharetta High School '11
Harvard '15

User avatar
AKKOLADE
Sin
Posts: 15261
Joined: Thu Apr 24, 2003 8:08 am

Re: 2011 PACE NSC Tournament Format

Post by AKKOLADE » Wed Apr 06, 2011 10:49 pm

Twelve preliminary groups will spread the strength of the teams out even more than last year's format.

Of course, if only quiz bowl tournaments were seeded using Gray's Academic Competition Almanac...
Fred Morlan
PACE President, 2018-19
International Quiz Bowl Tournaments, co-owner
University of Kentucky CoP, 2017
hsqbrank manager, NAQT writer (former subject editor), former hsqb Administrator/Chief Administrator, 2012 NASAT TD

User avatar
cvdwightw
Auron
Posts: 3446
Joined: Tue May 13, 2003 12:46 am
Location: Southern CA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by cvdwightw » Thu Apr 07, 2011 4:13 am

mtimmons wrote:EDIT: Are prelims supposed to be 2/3/3? That seems like it would actually work.
I think I meant splitting 2/2/2/2, giving 3 brackets of 6 in each playoff tier (and thus making the "play the other four teams in your superplayoff bracket you haven't already played" part make sense).
AlphaQuizBowler wrote:So here's my concern: it's bad enough to be relegated to at best 17th place after 7 games. Being relegated to at best 25th after only 5 games? This requires near-perfect bracketing, which, based on the statistics from previous years, PACE (or really any other organization) just really isn't capable of.
Everything I say from here on out should be viewed as me responding personally and not on behalf of PACE or anyone associated with the tournament in any official capacity, because this statement really hit a nerve with me. Maybe it's the fact that it's late and I'm not feeling particularly passive-aggressive right now, but I don't get at all what your point is here. I guess it sucks to be the 22nd best team at the PACE NSC and get eliminated from championship contention after five rounds because the 21st best team at the PACE NSC was underseeded and beat you, or because you got upset by the 38th best team and couldn't pull out a tiebreaker. Guess what? You're not winning the championship if you're the 22nd best team at the PACE NSC. There's no way to sugar-coat it. You're just not.

Basically, the way I interpret your posts is that you want PACE to retroactively scrap the entire 2010 format because places 11-20 didn't perfectly reflect those teams' ability and performance. And you want to scrap the 2011 format because places 21-30 probably won't reflect your Platonic ideal of those teams' ability and performance either. I don't know what the hell you want me to do. I don't. Because your argument boils down to "bracketed round robin tournaments are unfair because some brackets are harder than others." So great. I guess we can't use bracketed round robin anymore. And as much as I'd love to run a 72-team round robin tournament, it would take about four or five more days and, oh, somewhere around fifty more packets than currently budgeted. So I guess a full round robin's out. What's left? Swiss pair, which has its own set of strength-of-schedule issues? Double elimination? I know, let's just cancel the NSC and declare placement by having every registered team rank the rest of the field.

Seriously: how many preliminary games is acceptable to you? Nine? Thirteen? Eighteen? As many as it takes until your team isn't in the eliminated-from-the-championship range? I know it's more than seven. So come up with that number first. And then devise a format that perfectly ranks all the teams in the tournament given that number of prelim games. And then make sure that format is perfectly robust against upsets and seeding deviations. Once you do that, then maybe I'll actually defend to you the tradeoff of more playoff teams (both in absolute numbers and by percentage) against fewer prelim games.
Dwight Wynne
socalquizbowl.org
UC Irvine 2008-2013; UCLA 2004-2007; Capistrano Valley High School 2000-2003

"It's a competition, but it's not a sport. On a scale, if football is a 10, then rowing would be a two. One would be Quiz Bowl." --Matt Birk on rowing, SI On Campus, 10/21/03

"If you were my teammate, I would have tossed your ass out the door so fast you'd be emitting Cerenkov radiation, but I'm not classy like Dwight." --Jerry

User avatar
Mechanical Beasts
Banned Cheater
Posts: 5673
Joined: Thu Jun 08, 2006 10:50 pm

Re: 2011 PACE NSC Tournament Format

Post by Mechanical Beasts » Thu Apr 07, 2011 4:41 am

cvdwightw wrote:
AlphaQuizBowler wrote:So here's my concern: it's bad enough to be relegated to at best 17th place after 7 games. Being relegated to at best 25th after only 5 games? This requires near-perfect bracketing, which, based on the statistics from previous years, PACE (or really any other organization) just really isn't capable of.
Everything I say from here on out should be viewed as me responding personally and not on behalf of PACE or anyone associated with the tournament in any official capacity, because this statement really hit a nerve with me. Maybe it's the fact that it's late and I'm not feeling particularly passive-aggressive right now, but I don't get at all what your point is here. I guess it sucks to be the 22nd best team at the PACE NSC and get eliminated from championship contention after five rounds because the 21st best team at the PACE NSC was underseeded and beat you, or because you got upset by the 38th best team and couldn't pull out a tiebreaker. Guess what? You're not winning the championship if you're the 22nd best team at the PACE NSC. There's no way to sugar-coat it. You're just not.
It always sucks to be on the threshold. But here's a big difference between the scenarios that you're not mentioning, William: in the former case, only 16 teams are still alive. In the latter case, 24 teams are still alive. We're keeping more teams in championship contention for longer. We shuck off a fraction of the field after a shorter period of time, but it's a smaller enough fraction shucked that it merits less time.

To make Dwight's point even more ironclad, you could even make the argument that we actually have a far easier time bracketing in this scenario because we only have to get the top three places right, and all three teams of the dreaded "1, 2n, 2n+1, ..." bracket make it through.
Andrew Watkins

Romero
Wakka
Posts: 157
Joined: Thu Jun 05, 2003 11:05 pm
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Romero » Thu Apr 07, 2011 7:02 am

cvdwightw wrote:I don't know what the hell you want me to do.
I cannot speak for anyone else but I would like you to come up with a better format that is fair to all teams and does not rely upon PACE's ability to seed teams before the tournament.

I think the key issue here is that you need to get beyond the idea of starting with n brackets of m teams.
dtaylor4 wrote:Teams are never out of contention until they lose two games.
Last year my team lost 1.5 and was out. If you are going to insist upon this bracket system, I insist on full game tie breakers so that at least Donald's statement is technically true.
Chris Romero
Texas Quiz Bowl Alliance & Louisiana Quiz Bowl Alliance
Quiz Bowl Coach
Former Player, Texas A&M University ('96-'07)
Former Player, Catholic High School-New Iberia, LA ('92-'96)

User avatar
Down and out in Quintana Roo
Auron
Posts: 2907
Joined: Wed Apr 09, 2008 7:25 am
Location: Camden, DE
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Down and out in Quintana Roo » Thu Apr 07, 2011 7:21 am

Romero wrote:
dtaylor4 wrote:Teams are never out of contention until they lose two games.
Last year my team lost 1.5 and was out. If you are going to insist upon this bracket system, I insist on full game tie breakers so that at least Donald's statement is technically true.
This. If you're going to have a national tournament, ties should not be broken on anything but a full game on a full packet. You don't want a skewed distribution (say, several Literature and History questions in that half packet, and very little if any Science or Math) to determine the championship ranking for a team.
Mr. Andrew Chrzanowski
Caesar Rodney High School
Camden, Delaware
CRHS '97-'01
University of Delaware '01-'05
CRHS quizbowl coach '06-'12
http://crquizbowl.edublogs.org

User avatar
Sen. Estes Kefauver (D-TN)
Chairman of Anti-Music Mafia Committee
Posts: 5640
Joined: Wed Jul 26, 2006 11:46 pm
Location: Columbia, MO

Re: 2011 PACE NSC Tournament Format

Post by Sen. Estes Kefauver (D-TN) » Thu Apr 07, 2011 7:27 am

There won't be a PACE half packet that is "lit and history-heavy, with no science." It's not a random bunch of 10 tossups and bonuses that are cobbled together, it's an internally evenly distributed half of a regular PACE packet that is basically as close to reflecting "half" of a packet's distribution as it can.
Charlie Dees, North Kansas City HS '08
"I won't say more because I know some of you parse everything I say." - Jeremy Gibbs

"At one TJ tournament the neg prize was the Hampshire College ultimate frisbee team (nude) calender featuring one Evan Silberman. In retrospect that could have been a disaster." - Harry White

User avatar
Down and out in Quintana Roo
Auron
Posts: 2907
Joined: Wed Apr 09, 2008 7:25 am
Location: Camden, DE
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Down and out in Quintana Roo » Thu Apr 07, 2011 8:46 am

Jeremy Gibbs Freesy Does It wrote:There won't be a PACE half packet that is "lit and history-heavy, with no science." It's not a random bunch of 10 tossups and bonuses that are cobbled together, it's an internally evenly distributed half of a regular PACE packet that is basically as close to reflecting "half" of a packet's distribution as it can.
Then why not have every game on half packets if they're just as good? My point is that by using a half packet, the game has the appearance of "half the importance" as well. If each game is equally factored (and equally important), then each game should be played under the same circumstances.
Mr. Andrew Chrzanowski
Caesar Rodney High School
Camden, Delaware
CRHS '97-'01
University of Delaware '01-'05
CRHS quizbowl coach '06-'12
http://crquizbowl.edublogs.org

User avatar
Galstaff, Sorceror of Light
Forums Staff: Administrator
Posts: 2241
Joined: Sat Aug 18, 2007 5:40 pm
Location: Richmond, VA

Re: 2011 PACE NSC Tournament Format

Post by Galstaff, Sorceror of Light » Thu Apr 07, 2011 10:28 am

Carangoides ciliarius wrote:
Jeremy Gibbs Freesy Does It wrote:There won't be a PACE half packet that is "lit and history-heavy, with no science." It's not a random bunch of 10 tossups and bonuses that are cobbled together, it's an internally evenly distributed half of a regular PACE packet that is basically as close to reflecting "half" of a packet's distribution as it can.
Then why not have every game on half packets if they're just as good? My point is that by using a half packet, the game has the appearance of "half the importance" as well. If each game is equally factored (and equally important), then each game should be played under the same circumstances.
Speaking not as a member of PACE but as someone who has played in 4 NSCs, I have absolutely no idea where you're getting this idea that half packet games have somehow "half the importance." Consider for example MW A's half packet game against Stuyvesant A at the 2007 NSC, which I'm pretty sure was the first half packet game I witnessed. That was pretty important, and these games are played because they are important. Having to play a pre-made half packet is unfortunate, but it's certainly a better alternative than a) having to use a statistical tiebreaker or b) arbitrarily splitting a packet in half if there are more ties than expected, creating an unfair distribution. I can understand your desire for all games to be the same, but I also think there's a point at which that just might not be practical.
Sam (Sarah Angelo) Luongo,
Maggie L. Walker Governor's School 2010 / UVA 2014 / VCU School of Education 2016
President, PACE
Member, ACF

User avatar
dtaylor4
Auron
Posts: 3733
Joined: Tue Nov 16, 2004 11:43 am

Re: 2011 PACE NSC Tournament Format

Post by dtaylor4 » Thu Apr 07, 2011 10:57 am

Carangoides ciliarius wrote:
Jeremy Gibbs Freesy Does It wrote:There won't be a PACE half packet that is "lit and history-heavy, with no science." It's not a random bunch of 10 tossups and bonuses that are cobbled together, it's an internally evenly distributed half of a regular PACE packet that is basically as close to reflecting "half" of a packet's distribution as it can.
Then why not have every game on half packets if they're just as good? My point is that by using a half packet, the game has the appearance of "half the importance" as well. If each game is equally factored (and equally important), then each game should be played under the same circumstances.
To clarify: In my statement, I counted the half-packet tiebreakers as games. Also, I am not the TD, so I am in no position to alter this.

Half-packet games, to me, are a good balance between breaking off the tie, and keeping the tournament running smoothly. Let's look at what happens in a three-way tie. Two full games (plus any OT) must be played, which takes at least an hour. Given that this is during a break in the format, what do the other 60+ teams do? If it's during lunch, do the teams in the tiebreaker simply not eat lunch? There are logistical implications that come into play when playing off ties, especially at a large tournament like the NSC.

User avatar
Deviant Insider
Auron
Posts: 4603
Joined: Sun Jun 13, 2004 6:08 am
Location: Chicagoland
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Deviant Insider » Thu Apr 07, 2011 11:03 am

If you finished 3rd in a round robin pool, one of the following three things happened:
A) You lost at least two full matches.
B) You lost a full match and two half matches.
C) You lost a full match, had the lowest PPG of the three teams in a circle of death, and lost a half match.

In other words, if you got eliminated while only losing a full match and a half match, then you also had weaker stats than two teams in your pool through five full matches against common opponents, and you got a lesser chance on the half packets because of it. If your stats were better, then you would have gotten two chances on half packets.
David Reinstein
PACE VP of Outreach, Head Writer and Editor for Scobol Solo and Masonics (Illinois), TD for New Trier Scobol Solo and New Trier Varsity, Writer for NAQT (2011-2017), IHSSBCA Board Member, IHSSBCA Chair (2004-2014), PACE Member, PACE President (2016-2018), New Trier Coach (1994-2011)

User avatar
Mechanical Beasts
Banned Cheater
Posts: 5673
Joined: Thu Jun 08, 2006 10:50 pm

Re: 2011 PACE NSC Tournament Format

Post by Mechanical Beasts » Thu Apr 07, 2011 11:19 am

Carangoides ciliarius wrote:Then why not have every game on half packets if they're just as good?
Because it's possible to write a tournament at this difficulty level that contains 23-25 packets of non-repeating material, as we did last year. Your strawman isn't convincing. No one said a sample of 10 questions is as good as a sample of 20, but you were saying "but a half packet is surely not actually randomized in a balanced way" which is flat-out wrong.
Andrew Watkins

User avatar
Down and out in Quintana Roo
Auron
Posts: 2907
Joined: Wed Apr 09, 2008 7:25 am
Location: Camden, DE
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Down and out in Quintana Roo » Thu Apr 07, 2011 12:11 pm

I do trust that half-packets are as good as they can be - for what they are. I just think it's slightly unfair for "half a loss" to eliminate a team from championship contention too, like the Seven Lakes coach mentioned earlier.

I realize that i'm making the situation more difficult here but i'm just trying to get people to think about all possible logistical alternatives. If there are no others that are sensible, then this is what we have. But we should try to make it as fair as we can, in the meantime.

I trust that PACE is doing all they can and once again i laud them for what is, by far, the best national championship in all of high school quizbowl.
Mr. Andrew Chrzanowski
Caesar Rodney High School
Camden, Delaware
CRHS '97-'01
University of Delaware '01-'05
CRHS quizbowl coach '06-'12
http://crquizbowl.edublogs.org

User avatar
jonpin
Forums Staff: Moderator
Posts: 2031
Joined: Wed Feb 04, 2004 6:45 pm
Location: BCA NJ / WUSTL MO / Hackensack NJ

Re: 2011 PACE NSC Tournament Format

Post by jonpin » Thu Apr 07, 2011 12:43 pm

Carangoides ciliarius wrote:I do trust that half-packets are as good as they can be - for what they are. I just think it's slightly unfair for "half a loss" to eliminate a team from championship contention too, like the Seven Lakes coach mentioned earlier.

I realize that i'm making the situation more difficult here but i'm just trying to get people to think about all possible logistical alternatives. If there are no others that are sensible, then this is what we have. But we should try to make it as fair as we can, in the meantime.

I trust that PACE is doing all they can and once again i laud them for what is, by far, the best national championship in all of high school quizbowl.
This is true, but it's just not at all practical to do any better given the resources of time and packets. What I would recommend (and I recommended last year) would be that any two-way ties just get a full packet.
In the absence of that, I would suggest that in compiling the tiebreaker packets, you make them as discrete packets of 11/10, as I know last year some people reading tiebreakers got confused and either read a full game or read a 3-cycle OT rather than sudden death. Also this way, if you publish say 5 tiebreaker half-packets, and only one is used for the first break, you can start with TB packet 2 at the second break. (Is this making any sense?)
Jon Pinyan
Coach, Bergen County Academies (NJ); former player for BCA (2000-03) and WUSTL (2003-07)
HSQB forum mod, PACE member
Stat director for: NSC '13-'15, '17; ACF '14, '17, '19; NHBB '13-'15; NASAT '11

"A [...] wizard who controls the weather" - Jerry Vinokurov

User avatar
Sen. Estes Kefauver (D-TN)
Chairman of Anti-Music Mafia Committee
Posts: 5640
Joined: Wed Jul 26, 2006 11:46 pm
Location: Columbia, MO

Re: 2011 PACE NSC Tournament Format

Post by Sen. Estes Kefauver (D-TN) » Thu Apr 07, 2011 12:51 pm

I realize that i'm making the situation more difficult here but i'm just trying to get people to think about all possible logistical alternatives.
No you aren't, you're getting people to think about situations that literally will never happen at the NSC. I'm not crazy about the small prelim size either, but that doesn't mean I'm going to invent made up non-existant half-packet distributions to argue against it, or otherwise stir the pot.
Charlie Dees, North Kansas City HS '08
"I won't say more because I know some of you parse everything I say." - Jeremy Gibbs

"At one TJ tournament the neg prize was the Hampshire College ultimate frisbee team (nude) calender featuring one Evan Silberman. In retrospect that could have been a disaster." - Harry White

Romero
Wakka
Posts: 157
Joined: Thu Jun 05, 2003 11:05 pm
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Romero » Thu Apr 07, 2011 2:08 pm

My students and I like the PACE questions better but the format is still a huge issue for us, particularly after last year. PACE is not the best national for two reasons: Its format does not afford each team an equal opportunity to reach the playoffs and it has bounce back bonuses. The bounce back thing is for another discussion.

You should not break some ties with half packets and other ties with full packets. Two losses should not get one team from one division into the next round while two losses eliminates a team from another division. Some playoff spots should not be determined on full packets while others are determined on half packets.

In general smaller sample sizes should be avoided to minimize sensitivity to incorrect initial seeding and avoid tiebreaker necessities. With a larger initial group size, we can eliminate the need for the first break (and the first tiebreaker) and we allow more divisional results to be based on games played rather than "luck of the seeding draw."

By making smaller initial groups, you are relying more heavily upon the accuracy of the initial seeds. PACE is incapable of fairly seeding teams. In truth given the lack of quiz bowl homogeneity across regions, it is probably an impossible task.

William alluded to statistics that demonstrate the faults of last year's format. I believe a smaller sample size will result in more tiebreakers and moving the controversy from teams 16/17 last year to teams 24/25 this year. Dwight asked what do we expect you to do? There are a number of options, but addressing concerns of your "customers" would be a good step.

Why not 6 groups of 12 with all teams with 3 losses or fewer moving on? Why not introduce some wild card spots to adjust for the unequal initial groups? Why not write a few more questions to allow full packet tiebreakers? None of the issues brought up here are impossible to address.
Last edited by Romero on Thu Apr 07, 2011 2:34 pm, edited 1 time in total.
Chris Romero
Texas Quiz Bowl Alliance & Louisiana Quiz Bowl Alliance
Quiz Bowl Coach
Former Player, Texas A&M University ('96-'07)
Former Player, Catholic High School-New Iberia, LA ('92-'96)

User avatar
Not That Kind of Christian!!
Yuna
Posts: 847
Joined: Mon Feb 26, 2007 10:36 pm
Location: Manhattan

Re: 2011 PACE NSC Tournament Format

Post by Not That Kind of Christian!! » Thu Apr 07, 2011 2:31 pm

Romero wrote:Why not write a few more questions to allow full bracket tiebreakers?
Speaking as someone heavily involved in writing last year's and this year's NSCs, this is not a trivial task. Editors have been striving for better difficulty control than the NSC has had in previous years. While I think there's been a good amount of success on this front, unfortunately, it also constricts the available answer space somewhat. Expanding those half packets to full packets with quality and difficulty-appropriate answers might not yield the results we'd and you'd like to see your students playing a tiebreaker on.
Hannah Kirsch
Brandeis University 2010
NYU School of Medicine 2014

"Wow, those Scandinavians completely thorbjorned my hard-earned political capital."

User avatar
cvdwightw
Auron
Posts: 3446
Joined: Tue May 13, 2003 12:46 am
Location: Southern CA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by cvdwightw » Thu Apr 07, 2011 2:34 pm

Romero wrote:
cvdwightw wrote:I don't know what the hell you want me to do.
I cannot speak for anyone else but I would like you to come up with a better format that is fair to all teams and does not rely upon PACE's ability to seed teams before the tournament.

I think the key issue here is that you need to get beyond the idea of starting with n brackets of m teams.
Um...what? I can't devise a 72-team round robin. And as far as I know, these are the only other scheduling options:

A. Swiss pair/card system. This is just as dependent on, and arguably even more dependent on, PACE's ability to accurately seed every team in the tournament, because a mis-seed or upset affects a whole lot more teams a whole lot faster than in a bracketed round robin.
B. n-elimination. Seriously?
C. Random prelim games similar to early HSNCTs. I'm pretty sure there was a good reason NAQT got rid of this, namely that strength of schedule was completely unbalanced.
D. Ladder system. Not only do I not have any idea how to do this with 72 teams, it was also scrapped for being unfair to many teams. In 2003.

So unless there's some newfangled tournament format that neither I nor Wikipedia have heard of, you're essentially asking me to design an entirely new tournament format. And make sure it is more fair than what we have now.

Bracketed round robins have been upheld for some time now as the best we currently have to offer, not because it's what's always been done, but because they provide a fairer chance for all teams than any other plausible current format. Now that's not good enough. Look, I don't have a problem saying that "five games before you get eliminated from contention is too few for a tournament of this prestige." The rebuttal to that is that (1) the more brackets you have in a bracketed round robin, the more spread out the ability levels are and the less exact you have to have your seeds to get a fair ranking at the end of the phase; (2) because of that, you only have to really worry about getting knocked out by an underseeded but better team if you're one of those teams possibly in the 21-28 seed range or the 45-52 seed range; (3) if you're going to get knocked out by getting upset by a lower seed, or losing a circle of death because someone else pulled an upset, that's going to happen whether it's 5 games or 7 games or 35 games. And if that argument still doesn't convince anyone, I guess we can go back into internal PACE discussions and see if the 9x8/12x6/12x6 format can work.

What I do have a problem with is that people whose teams arguably got screwed over last year by bad seeding are now coming out and clamoring for the elimination not of this specific format but of bracketed round robins in general. I looked through the past 2 NSC format threads, and this is what I found:

2009: No complaints about the tournament format as a whole. Complaints about the possibility of a team finishing in third place with one loss were addressed in 2010 with the addition of superplayoffs.
2010: Brief complaints about the switch in match format and distribution had nothing to do with tournament format. Brief complaints about seeding once seeds were revealed. No complaints about the format as a whole.

In other words, no one complained about using a bracketed round robin in 2009 or 2010. In fact, the only complaint about the tournament format at all was about the single-elimination at the top four in 2009. And now, the two people that are complaining about bracketed round robins are, coincidentally, two people associated with the two teams that were most affected by bad seeding in 2010.
Romero wrote:Why not 5 groups of 12 with all teams with 3 losses or fewer moving on?
Because this means that we do not have a pre-determined idea of the number of teams that make the playoffs, which means that either we have to run n-elimination playoffs (which have been attacked as unfair as recently as last year) or we have to eliminate some teams from contention in an arbitrary and shady way (which NAQT specifically changed the HSNCT format to avoid).
Romero wrote:Why not introduce some wild card spots to adjust for the unequal initial groups?
Because this either requires determining playoff spots on statistical tiebreakers (which is a universal no-no in PACE orthodoxy) or adding additional play-in games beyond the tiebreakers, which is unfair to the teams not involved in the tiebreakers; because that makes putting any semblance of balance in the playoff brackets a lot harder; because once my team got knocked out of the ACF Nationals top bracket (because of an upset that did not involve my team) in favor of a wild card team that had two fewer wins in its prelim bracket, and I have just as much right to irrationally complain about wild cards as you do about the current format.
Romero wrote:Why not write a few more questions to allow full bracket tiebreakers?
Because we're already pushing the boundaries of accessibility with the number of packets we have; because every tournament produced ever by a large group is inevitably going to run into last-minute question requests and can't afford to expand beyond what the editors initially budgeted for; because once again we need to balance being fair to the teams involved in the tiebreaker with being fair to the teams not involved in the tiebreaker.
Dwight Wynne
socalquizbowl.org
UC Irvine 2008-2013; UCLA 2004-2007; Capistrano Valley High School 2000-2003

"It's a competition, but it's not a sport. On a scale, if football is a 10, then rowing would be a two. One would be Quiz Bowl." --Matt Birk on rowing, SI On Campus, 10/21/03

"If you were my teammate, I would have tossed your ass out the door so fast you'd be emitting Cerenkov radiation, but I'm not classy like Dwight." --Jerry

User avatar
jonpin
Forums Staff: Moderator
Posts: 2031
Joined: Wed Feb 04, 2004 6:45 pm
Location: BCA NJ / WUSTL MO / Hackensack NJ

Re: 2011 PACE NSC Tournament Format

Post by jonpin » Thu Apr 07, 2011 2:41 pm

Romero wrote:Why not 5 groups of 12 with all teams with 3 losses or fewer moving on? Why not introduce some wild card spots to adjust for the unequal initial groups? Why not write a few more questions to allow full bracket tiebreakers? None of the issues brought up here are impossible to address.
(obvious pre-disclaimer: I am not in any sense at all affiliated with PACE; also Dwight may have said what I'm saying here more eloquently and in a more informed manner)
Assuming you mean 6 groups of 12: The problem with advancing "all teams with 3 losses or fewer" is that it leaves you with a wide range of how many teams would be available in the next round, such that the only possible next step would be a bracket of some sort. In addition, playing in groups of 12 would take 11 rounds, which means that's all day Saturday. With approximately 24 teams (if the brackets are true) and only Sunday left, again, the only way to go from that many to a winner is a bracket of some sort. Thus this idea is just not going to happen, because PACE believes that given the opportunity, round-robin with advantaged-final is the best way to crown a champion. Starting with few but giant groups eliminates that opportunity, and would by and large result in a less fair ranking of teams at the end of the weekend.

Another problem is that 8-3 in one bracket doesn't mean the same thing as 8-3 in another bracket for the reason you so often point out--the groups are not identical in strength. The reason why W-L shouldn't generally be the first criteria for wild cards is that very reason.

Now as to wild cards, in theory there's nothing bad about them, but based on the numbers, I'd be hard pressed to find a situation where advancing 2 teams from each group plus some number of wild cards produces a better tournament format than what they have.

Part of the concern here comes from a lofty but impossible-to-achieve goal that PACE sets--giving each team that shows up a huge number of games and an exact rank. Can I say for certain that the 26th place team at PACE 2010 was better than the 27th place team? Absolutely not. In fact, in certain cases, it's even possible that they had nearly identical results against their common opponents, and the timing of their losses put them in those positions. There's no conceivable system that we could use to, in a day's worth of play, say with certainty "These are the top X teams", where 12<X<24 depending on how you wanted to do playoffs. So the top team of the second tier is likely going to be better than the worst team of the first tier. And there's nothing wrong with that. I don't know for sure, but I think in past years, there's been a modest prize (plaque or trophy or something) for the best team in each playoff tier. In any event, you can have almost as much fun and learn almost as much if not more by running through the second tier and finishing 25th, then getting crushed in the first tier and finishing 24th.

I really do not think your concerns merit a rejection of the currently proposed format.
Jon Pinyan
Coach, Bergen County Academies (NJ); former player for BCA (2000-03) and WUSTL (2003-07)
HSQB forum mod, PACE member
Stat director for: NSC '13-'15, '17; ACF '14, '17, '19; NHBB '13-'15; NASAT '11

"A [...] wizard who controls the weather" - Jerry Vinokurov

Romero
Wakka
Posts: 157
Joined: Thu Jun 05, 2003 11:05 pm
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Romero » Thu Apr 07, 2011 3:29 pm

cvdwightw wrote: the two teams that were most affected by bad seeding in 2010.
Ok, assuming the a bracketed round robin is the best option and given that seeding is necessary and you acknowledge that teams did get "screwed over" last year ,what steps will PACE take to assure a more appropriate seeding this year?

How might you suggest avoiding the midwest/mid-atlantic bias that has been seen in the past?

Will the seeding discussion be made public before the event? Will a seeding committee be formed in a way that each region is equally represented? Will each region have at least one representative on the committee? Who is in charge of the seeding process this year?
Chris Romero
Texas Quiz Bowl Alliance & Louisiana Quiz Bowl Alliance
Quiz Bowl Coach
Former Player, Texas A&M University ('96-'07)
Former Player, Catholic High School-New Iberia, LA ('92-'96)

User avatar
cvdwightw
Auron
Posts: 3446
Joined: Tue May 13, 2003 12:46 am
Location: Southern CA
Contact:

Re: 2011 PACE NSC Tournament Format

Post by cvdwightw » Thu Apr 07, 2011 3:31 pm

I missed a few things here, and figured I'd reply in a new post rather than tacking more onto the last behemoth.
Romero wrote:In general smaller sample sizes should be avoided to minimize sensitivity to incorrect initial seeding and avoid tiebreaker necessities. With a larger initial group size, we can eliminate the need for the first break (and the first tiebreaker) and we allow more divisional results to be based on games played rather than "luck of the seeding draw."
This isn't actually right. Given a certain initial field size, and the stipulation of a full round robin (and only one) within each bracket, you either shrink the number of prelim brackets or shrink the number of teams within each bracket. By shrinking the number of prelim brackets, the average set of three consecutive overall seeds within a bracket has a smaller standard deviation, meaning that we would expect more circle-of-death scenarios as these teams would have a closer to 50% chance of beating each other (e.g. seed line #3 in the bracket with the #13 overall seed has a better chance of the upset in the 8-team brackets than in the 6-team brackets). This means that increasing the "sample size" by switching to a prelim structure of fewer teams/more games would result in more "first tiebreakers," not less. I am around 95% confident that a Monte Carlo simulation, given reasonable win probabilities for each team, will back me up on this.
Romero wrote:By making smaller initial groups, you are relying more heavily upon the accuracy of the initial seeds. PACE is incapable of fairly seeding teams. In truth given the lack of quiz bowl homogeneity across regions, it is probably an impossible task.
The fewer brackets you have, the more sensitive to initial seeding you are, because the further you are off for any given team, the more likely it is that the team is off by one or more seed lines. In other words, seed #11 being the actual 17th-best team probably doesn't make a whole lot of difference with 12 prelim brackets, but it does with 8 and with 6. The more brackets you have, the higher your average seeding-from-reality deviation can be without affecting overall results.

Seeding teams requires making predictions about how each team will perform during the tournament, based on their past performance. Inevitably, these predictions are wrong, because some teams will play better than predicted during the prelims and others will play worse. That's why you play the tournament. If PACE could "fairly" seed all the teams in the tournament according to your definition of "fair," there would be no point in actually playing the tournament, since every team would finish exactly in the place pre-determined by PACE seeding.
Romero wrote:William alluded to statistics that demonstrate the faults of last year's format. I believe a smaller sample size will result in more tiebreakers and moving the controversy from teams 16/17 last year to teams 24/25 this year. Dwight asked what do we expect you to do? There are a number of options, but addressing concerns of your "customers" would be a good step.
These "faults" were not publicly addressed by either you or William when tournament seedings and full stats had been publicly released. Maybe you didn't have internet until the thread died off; maybe you were too busy to bother about it; maybe you contacted other people in PACE, who conveniently either ignored your complaints or didn't bother to address them during early format discussions.

If you run any sort of bracketed round robin, regardless of whether you use wild cards or not or how many prelim matches you have, there is going to be controversy over the break between playoff and non-playoff teams. I don't see how moving the controversy from a higher overall tournament seed (by absolute number and percentile of teams) to a lower overall tournament seed is worse.

EDIT:
Romero wrote:you acknowledge that teams did get "screwed over" last year
Stop turning my words around. I said they were "arguably screwed over." Notably, you are arguing that. Do I really believe that? No, but I was trying to throw you a bone there. I guess I know better than to do that next time. Did teams receive seeds that were too low or too high, relative to how they played in the tournament? Yes. Mike Sorice even acknowledged that "it's easy to see in retrospect that there were a couple seeds too low and correspondingly a couple too high...The example most prominently brought to my attention was that Seven Lakes A was a low 3/the top 4 (due to the bracketing procedure, those positions are equivalent) and ought to have been a high 3." Were teams affected by this? Yes. Was Alpharetta affected by geographical diversity moving Northmont to their bracket instead of Quince Orchard or Lisle? Yes. Was Alpharetta going to make the top bracket after losing to Stevenson regardless of whether they were seeded 16th or 17th? Yes. Could they have lost to any other two-seed in the tournament on one of the prelim packets? Yes. Could Seven Lakes have made the top bracket if they were given a high 3 seed? In Mike Sorice's opinion, no. Were seeds affected by MSJ dropping out after seeding assignments? Yes. Will seeds be affected by another team doing that even if every seeding reform Chris Romero advocates for happens? Yes.
Last edited by cvdwightw on Thu Apr 07, 2011 3:48 pm, edited 1 time in total.
Dwight Wynne
socalquizbowl.org
UC Irvine 2008-2013; UCLA 2004-2007; Capistrano Valley High School 2000-2003

"It's a competition, but it's not a sport. On a scale, if football is a 10, then rowing would be a two. One would be Quiz Bowl." --Matt Birk on rowing, SI On Campus, 10/21/03

"If you were my teammate, I would have tossed your ass out the door so fast you'd be emitting Cerenkov radiation, but I'm not classy like Dwight." --Jerry

User avatar
Mechanical Beasts
Banned Cheater
Posts: 5673
Joined: Thu Jun 08, 2006 10:50 pm

Re: 2011 PACE NSC Tournament Format

Post by Mechanical Beasts » Thu Apr 07, 2011 3:43 pm

Romero wrote:How might you suggest avoiding the midwest/mid-atlantic bias that has been seen in the past?
PACE tends to use things like "statistics" to seed teams. Statistics don't inherently have a midwest/mid-atlantic bias, though I suppose if regions other than those tend to post statistics less frequently or go to fewer tournaments, we would actually be incapable of ranking those teams as accurately.

What positive effects could be had by making especially sure to put people from different regions on the "seeding committee?" They can't increase the amount of stats available to us, so all we could do is rely on anecdotes of moxie and pluck, which is somewhat useless. It's not like I affirmatively want to keep people from region X to be on the seeding committee; I just don't see a reason to make sure that we have one there--so if no one from region X wants to do it, we don't have to cry or try to admit new members into PACE to compensate.
Andrew Watkins

User avatar
Sen. Estes Kefauver (D-TN)
Chairman of Anti-Music Mafia Committee
Posts: 5640
Joined: Wed Jul 26, 2006 11:46 pm
Location: Columbia, MO

Re: 2011 PACE NSC Tournament Format

Post by Sen. Estes Kefauver (D-TN) » Thu Apr 07, 2011 4:51 pm

Man, I don't want to think about how PACE could be doing any worse of a job of interacting with the public and its customers than this thread! Jesus Christ, what is wrong with all of you?
Charlie Dees, North Kansas City HS '08
"I won't say more because I know some of you parse everything I say." - Jeremy Gibbs

"At one TJ tournament the neg prize was the Hampshire College ultimate frisbee team (nude) calender featuring one Evan Silberman. In retrospect that could have been a disaster." - Harry White

User avatar
theMoMA
Forums Staff: Administrator
Posts: 5686
Joined: Mon Oct 23, 2006 2:00 am

Re: 2011 PACE NSC Tournament Format

Post by theMoMA » Thu Apr 07, 2011 5:46 pm

As one of the tournament directors, I will certainly be open to changing the format if the issues that those in this thread have brought up. From what I can see, these appear to be four issues:

1. Bracketed tournaments in general.
2. There are only five games in the prelims.
3. Teams that lose 1.5 games can be eliminated from contention.
4. Seeding should be better.

To point 1, I think Dwight is right to say that brackets are the fairest way to run the tournament. The other formats that exist are either demonstrably less fair or even more dependent on initial seedings. If anyone has suggestions, I'd love to hear them.

To point 2, I believe Dwight is also correct to say that more prelim brackets increases the chance that the best teams advance, not the other way around.

To point 3, this seems to me to be an area where we can improve the format. David Reinstein is correct to say that one of three things has happened if you got third in your bracket, and I believe that none of them are outright unfair.
A) You lost at least two full matches.
B) You lost a full match and two half matches.
C) You lost a full match, had the lowest PPG of the three teams in a circle of death, and lost a half match.
But I can understand the idea that each team should have to lose two full games before being eliminated from contention. I am currently asking whether it would be feasible to play the elimination game of a three-way tie on a full packet. Chris Romero said that "You should not break some ties with half packets and other ties with full packets." But I don't see why that is the case. If we play the second match on a full packet, it will require that all teams lose two games before falling from contention without taking undue amounts of time or straining the tournament editors to come up with lots more questions. The two teams who play the first half-packet match earn the chance to win that game by playing better in the round robin phase, so it is not to their detriment that their game is only a half-packet.

To point 4, I speak for all of PACE when I say that we will be as diligent and thorough as possible when making our seeding decisions. We are just as acutely aware of their importance as the teams themselves.
Andrew Hart
Minnesota alum

User avatar
etchdulac
Rikku
Posts: 355
Joined: Thu May 29, 2003 6:02 am
Location: Texas, for better or worse

Re: 2011 PACE NSC Tournament Format

Post by etchdulac » Thu Apr 07, 2011 7:45 pm

Thanks, Andrew; your answers pretty well wrap up sufficient answers for the above concerns IMHO. It seems to me that more pools of fewer teams does create a situation where a seeding inaccuracy has to be more drastic to have a distinct negative effect, but that certainly doesn't make them impossible. It's just the nature of the competition that these will arise, and they cannot be eliminated entirely by any degree of diligence.
theMoMA wrote:To point 4, I speak for all of PACE when I say that we will be as diligent and thorough as possible when making our seeding decisions. We are just as acutely aware of their importance as the teams themselves.
How far ahead of the event does PACE expect to have its brackets (or at least its seedings) finalized? Would it be any trouble to have these announced, say, 48 hours ahead of the event?
Stephen Fontenot
Texas Quiz Bowl Alliance Deputy Director
Communications, UT Dallas
Strake Jesuit '96 -+-+- Southwestern '00

User avatar
Gautam
Auron
Posts: 1413
Joined: Sun Feb 11, 2007 7:28 pm
Location: Zone of Avoidance
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Gautam » Thu Apr 07, 2011 9:36 pm

We are still organizing a seeding committee and stuff, but the goal is to have seeding/brackets done the weekend of or immediately after HSNCT. Whatever data gathering stuff we need to will likely occur before/around mid-May, hopefully.
Gautam - ACF
Currently tending to the 'quizbowl hobo' persuasion.

User avatar
Deviant Insider
Auron
Posts: 4603
Joined: Sun Jun 13, 2004 6:08 am
Location: Chicagoland
Contact:

Re: 2011 PACE NSC Tournament Format

Post by Deviant Insider » Thu Apr 07, 2011 9:37 pm

Two factors that should be considered when answering Andrew's question about a half packet followed by a full packet:
1) If it is a three-way tie for 2nd place (as opposed to 1st place), then both extra matches are elimination matches. On the one hand, you could justify just using half packets to break it since all teams have already lost two full matches. On the other hand, you might be able to justify playing it off the same way, since the second match would involve the team that won the statistical tiebreaker and would be advancing if no extra matches were played.
2) You could have one or two staffers act as gofers to get lunch for the teams in the 2nd extra match. Have a few menus on hand, have teams pick food right before the match, and get the food for them during the match (with the expectation that they would pay for the food). If you bring those teams food, then you only need to give them a half hour to eat rather than a full hour.
David Reinstein
PACE VP of Outreach, Head Writer and Editor for Scobol Solo and Masonics (Illinois), TD for New Trier Scobol Solo and New Trier Varsity, Writer for NAQT (2011-2017), IHSSBCA Board Member, IHSSBCA Chair (2004-2014), PACE Member, PACE President (2016-2018), New Trier Coach (1994-2011)

Albox
Kimahri
Posts: 4
Joined: Tue Mar 11, 2008 4:32 pm

Re: 2011 PACE NSC Tournament Format

Post by Albox » Thu Apr 07, 2011 11:11 pm

Speaking as someone who felt the disappointment of Seven Lakes's missing the top bracket, I think there is an issue with deciding top 24 from 5 preliminary games. Here are some beliefs I have:

1. We don't need to try all that hard to make sure the best handful of teams make it to the superplayoffs. A lot of the arguments here sound like "If you lose two games, you should be eliminated from championship contention," which is really beside the point. It's really easy to decide the champion because teams are going to play 10 games and then get to play all of the other teams who could possibly be good enough to be champion. Nobody believes PACE is bad at ranking the top few teams, because the round robin system at the end is great.

2. What PACE isn't so good at is sorting out the middle of the pack. Ideally, the 1st through 24th best team will make it to the next bracket after 5 games, but this requires 2 of those teams to be in each of 12 brackets. I bet some brackets will have 1 of those teams and some brackets will have 3 of those teams, which is where the current proposal fails.

3. That said, it's less important to get the teams towards the bottom right. So for instance if we have a bracket of 6 where two belong in 1-24, and three belong in 25-48, it'd be great if someone can figure out a way to make it so those three teams make it to that group, but this is a difficult problem and I think we should prioritize giving teams in the 13-24 range a chance to play each other on Sunday.

One idea is we could have the first team in each of the 12 brackets of 6 get a pass to the top 18. It's likely they all belong there, because of how much information we have about these teams. Now we just have to worry about the teams that come out in 2nd and 3rd in the first brackets. (24 teams)

Now we break everyone into groups of four and play a round robin. I'd seed these by bonus conversion, with 2 teams that finished in 2nd and 2 that finished in 3rd in each bracket. Top team makes top 18, bottom 3 fight for 19-36 for the rest of the tournament. Meanwhile, the 3 groups of teams that finished 1st in the first bracket play round robins with four teams.

The next round is between a team that was bottom 2 in the first-seed teams, and a top team in the second/third bracket games. Winner makes top 12, loser doesn't.

We have played 9 games and two sets of tiebreakers to determine top 12 at this point. This would be a convenient place to choose a different system to determine the rank of the top 12 if you don't like mine.

Break into three groups of four, top two make top six, then round robin top six to determine champion. third set of tiebreakers, and finals. I think there might be teams that play a second game against each other, but I don't think it's a guarantee, so we can't assume the match has already happened and skip it.

Overall, this is seventeen games, three rounds of tiebreakers, and finals. I didn't flesh it out entirely but I think if you guys like it you are more qualified to do that than me. The big advantage is it gives teams that finished third in the first bracket a chance to play three quality games against teams around their level to determine more precisely where in the top 36 they belong. We have to make relatively arbitrary cutoffs at some points in the tournament but I think we should spend more time on teams in this 13-36 area of the field who are very good at quiz bowl and would love to play more meaningful games overall and I think we can do this without sacrificing the quality of the top 12.

Some weaknesses: teams playing each other twice?, traditional pace format doesn't have to assume exactly 6 non-1-seeds make top 18 but we do, more tiebreaker phases, less symmetrical (what do other teams do during the single-elimination round?), only top 6 guaranteed to play each other.

If anyone thinks there is promise, I can help think about this again some more later.
Albert R. Bayer
LASA '08, Texas '10

User avatar
theMoMA
Forums Staff: Administrator
Posts: 5686
Joined: Mon Oct 23, 2006 2:00 am

Re: 2011 PACE NSC Tournament Format

Post by theMoMA » Thu Apr 07, 2011 11:42 pm

Albert, I think your format is a creative way to sort out the middle of the pack, but there appear to be a few problems with it. Chiefly, I wouldn't want to have a single-elimination round like you propose. Secondarily, it appears that many teams will be sitting around during some of the rounds, and I don't want that to happen either. Finally, I think the most important part of any format has to be giving every team with a reasonable chance of winning the tournament a fair shot to get into the top tier. I believe that the current format does that, but I'm not sure that your format does because of the single-elimination phase.

Reinstein brings up a good point:
1) If it is a three-way tie for 2nd place (as opposed to 1st place), then both extra matches are elimination matches. On the one hand, you could justify just using half packets to break it since all teams have already lost two full matches. On the other hand, you might be able to justify playing it off the same way, since the second match would involve the team that won the statistical tiebreaker and would be advancing if no extra matches were played.
My opinion is that either option is fair. It seems harsh to say that losing one game, tying for first in your bracket, and losing a half-packet tiebreaker eliminates your team from contention. It seems less harsh to say that losing two games, tying for second in your bracket, and losing any fair tiebreaker knocks your team from the first tier. I don't know the implications on packet production that breaking second-place, two-loss circles of death on full packets would have. But if it puts additional strain on the writers, I'd be inclined to go with half packets for this scenario.
2) You could have one or two staffers act as gofers to get lunch for the teams in the 2nd extra match. Have a few menus on hand, have teams pick food right before the match, and get the food for them during the match (with the expectation that they would pay for the food). If you bring those teams food, then you only need to give them a half hour to eat rather than a full hour.
Gautam and I are already working on lunch logistics that would allow for efficient tiebreaking. We'll release more info when we figure everything out.
Andrew Hart
Minnesota alum

User avatar
theMoMA
Forums Staff: Administrator
Posts: 5686
Joined: Mon Oct 23, 2006 2:00 am

Re: 2011 PACE NSC Tournament Format

Post by theMoMA » Thu Apr 07, 2011 11:45 pm

Also, I will add that if Albert or anyone else wants to pass along a proposal for an amended or different tournament format, please email me (hartam@gmail.com). We'll consider every proposal and respond to any inquiries, but I don't want to turn this thread into a constant stream of proposed NSC formats. Thanks.
Andrew Hart
Minnesota alum

User avatar
AlphaQuizBowler
Tidus
Posts: 695
Joined: Mon Dec 03, 2007 6:31 pm
Location: Alpharetta, GA

Re: 2011 PACE NSC Tournament Format

Post by AlphaQuizBowler » Fri Apr 08, 2011 10:08 am

This thread has certainly gone places since my last post. Let me just clarify what I was saying:
I'm not advocating getting rid of rebracketed formats. I don't think there's any superior alternative to rebracketing, and I've used rebracketing at all 3 high school tournaments that I've TDed. I realize that there is always the possibility that teams n+1 and and even n+2 (where n is the size of the top playoff bracket) get screwed by seeding; it just happened that my team was that team at the NSC last year.

What I'm trying to say primarily is that I hope PACE realizes how important the initial seeding of the tournament is. There are two things from last year that should be corrected: first, if a team drops out this year, PACE should adjust the brackets accordingly rather than inserting the replacement team into their spot. Also, PACE should try, to the best of its ability, to determine whether teams will be missing players--according to this post, duPont Manual, the #10 seed last year, was shorthanded, which perhaps caused some of the unevenness in the brackets.

I do realize, as Dwight is saying, that having 12 prelim brackets makes it easier for PACE to spread out the top teams and thus makes it less likely that teams get screwed over by the bracketing. And, now that I think about it, there probably is an advantage to moving the "bubble" from 16/17 to 24/25, because it gives a clearer picture of those teams in the 13-22 range, the range that was perhaps not so clear at the last NSC. I think that all the work Fred is putting into his rankings will make the seeding process more accurate as well as more transparent this year. The proposed format seems fine, and it was probably just my initial reaction "What? 5 prelim games to determine the top 24?" that caused my last post to come off the way it did.

EDIT:
cvdwightw wrote:I guess it sucks to be the 22nd best team at the PACE NSC and get eliminated from championship contention after five rounds because the 21st best team at the PACE NSC was underseeded and beat you, or because you got upset by the 38th best team and couldn't pull out a tiebreaker. Guess what? You're not winning the championship if you're the 22nd best team at the PACE NSC. There's no way to sugar-coat it. You're just not.
I was going back through the thread and this comment caught my eye. I hope this is not the attitude that PACE takes toward running their tournament, because teams don't attend the NSC just to see who's "winning the championship": teams want to see where they rank among the best teams in the nation, whether they are #1 or #61. The NSC format does this better than, say, the HSNCT format, but it's disappointing to see someone involved in designing that format express this idea.
William
Alpharetta High School '11
Harvard '15

User avatar
millionwaves
Auron
Posts: 1360
Joined: Sun Feb 25, 2007 12:35 pm
Location: Urbana, Illinois
Contact:

Re: 2011 PACE NSC Tournament Format

Post by millionwaves » Fri Apr 08, 2011 11:25 am

AlphaQuizBowler wrote: EDIT:
cvdwightw wrote:I guess it sucks to be the 22nd best team at the PACE NSC and get eliminated from championship contention after five rounds because the 21st best team at the PACE NSC was underseeded and beat you, or because you got upset by the 38th best team and couldn't pull out a tiebreaker. Guess what? You're not winning the championship if you're the 22nd best team at the PACE NSC. There's no way to sugar-coat it. You're just not.
I was going back through the thread and this comment caught my eye. I hope this is not the attitude that PACE takes toward running their tournament, because teams don't attend the NSC just to see who's "winning the championship": teams want to see where they rank among the best teams in the nation, whether they are #1 or #61. The NSC format does this better than, say, the HSNCT format, but it's disappointing to see someone involved in designing that format express this idea.
I'll briefly assume the burden of speaking for PACE in an NSC-related capacity to affirm that this is not the attitude of the organization as a whole. We're proud of the fact that the NSC provides an unambiguous ranking (this year) from 1-72, and we're currently discussing a change in the tiebreaker procedure to ensure that no team is eliminated after only 1.5 games.
Trygve Meade
Illinois, ACF

Above the Star-Apple Kingdom

Locked