Da NAQT Code

Dormant threads from the high school sections are preserved here.
Locked
User avatar
Stained Diviner
Auron
Posts: 4841
Joined: Sun Jun 13, 2004 6:08 am
Location: Chicagoland
Contact:

Da NAQT Code

Post by Stained Diviner »

For all you geeks out there, here is a partial explanation of how the NAQT system worked.

During the first five rounds, each team played against teams that matched their record. Not only did records match, but order of wins and losses also matched.

For example, our (New Trier) first match was against Novi. For the second match, our first match was paired with James Island defeating Eleanor Roosevelt, so we played James Island while Novi played Eleanor Roosevelt.

For the third match, this foursome was combined with Minnetonka, Kellenberg, Brookwood, and San Mateo. Minnetonka and we had WW, James Island and Kellenberg had WL, Brookwood and Novi had LW, and San Mateo and Roosevelt had LL.

For the fourth match, this octet was grouped against a similar octet as follows: Wilmington Charter def New Trier in the WWW match, Columbus def Minnetonka in the WWL match, James Island def Broken Arrow in the WLW match, Kellenberg def Menlo-Atherton in the WLL match, Brookwood def Half Hollow Hills in the LWW match, Novi def Cedar Shoals in the LWL match, San Mateo def Orono in the LLW match, and Roosevelt def MLK B in the LLL match.

For the fifth match, this group of sixteen was grouped against a similar group of sixteen. When it was all said and done, every permutation of a total of 5 W’s and L’s was occupied by one team of the thirty-two. Namely:
WWWWW State College
WWWWL Wilmington Charter
WWWLW New Trier
WWWLL East Lansing
WWLWW Danville
WWLWL Columbus
WWLLW Raleigh Charter
WWLLL Minnetonka
WLWWW James Island
WLWWL Arcadia B
WLWLW Broken Arrow
WLWLL Mustang
WLLWW Kellenberg
WLLWL Wausau West
WLLLW Malvern
WLLLL Menlo-Atherton
LWWWW Brookwood
LWWWL Georgetown Day
LWWLW George Mason
LWWLL Half Hollow Hills West
LWLWW Novi
LWLWL EO Smith B
LWLLW Cedar Shoals
LWLLL Sulphur
LLWWW San Mateo
LLWWL Loveless
LLWLW Orono
LLWLL Breck
LLLWW E Roosevelt
LLLWL Livonia Churchill
LLLLW MLK B
LLLLL Wayzata C

There were five such groups of 32 in the tournament. If you look at results of subsequent matches, the group above is very average on the whole. I don't know if there are any groups that turned out to be significantly above or below average.

For this group, winning early was a good thing. The four teams that won their first three matches were the only four teams that finished 7-3 or better. Of the twelve teams that started off 2-1, nine of them finished exactly 6-4, and three missed the playoffs. Of the sixteen teams that had a losing record through three matches, nobody made the playoffs.

For the sixth round, pairings were a little more haphazard. This was necessitated by the fact that there were now an odd number of teams with perfect records (perfect in both the good sense and the bad sense). Santa Monica, who won their first five matches, played 4-1 Gonzaga in their sixth match. Almost every team played somebody with the same record in their sixth match, but the orders of wins and losses no longer matched.

One oddity with the sixth round was that many 3-2 and 2-3 teams played opponents from within their group of 32. This opened up the possibility for repeats in the sixth round even though that possibility could have been postponed for another round.

After five rounds, there were 24 teams that had a chance to end up with the #1 card. They were, in bracket order:
New Trier & Livingston
Garfield Heights & Dorman A
Eden Prairie & Maggie Walker*
Russell & Danville
R Montgomery & Wheaton North
Wilmington Charter & St Ignatius
Whitman & Hunter
Cave Spring & Brookwood
-----
State College & MLK
Maggie Walker & Eden Prairie
Stuyvesant & James Island
TJ B & Central Gwinnett
MLK & State College*
Gonzaga & Santa Monica

The 5-0 teams were State College, MLK, Maggie Walker, Eden Prairie, and Santa Monica. The asterisks are the places in the bracket where the losing team of the first match got placed. There are fewer teams in the bottom bracket because the loser of the match between 6-0 teams got added to the very bottom of this bracket in Round 7.

There were other 4-1 teams that did not have a chance at the 1 and 2 cards no matter what they did the last five matches: Dorman B, Dunbar, TJ A, Brindlee Mountain, Pingry, and Rancho Bernardo. These selections seem somewhat random; they were based more on the order of wins and losses, and NAQT had only limited control on who these teams were. In any respect, this placement is at least as much as an advantage as a disadvantage. Five of the six made the playoffs.

To NAQT's credit, teams in the groupings above played teams outside the group if they lost, though depending on position they sometimes had to lose two or three times. I am fairly certain that it was possible for all of them to make the playoffs.

Personally, I would have done the pairings a little differently in the first few rounds. After two rounds, WL teams are better on average than LW teams. Therefore, it is fairer if the WL teams play the LW teams rather than having WL teams playing other WL teams. Similar discrepancies exist for all rounds. This flaw only applied to the first five rounds as far as I could see, and the pairings in the final rounds generally are more important in terms of who lives and who dies anyways.

I also think that some pairings could have done more to decrease the likelihood of repeat matches. Any power system with this number of teams and matches gives some possibility of repeats, and I can't guarantee that it would be possible to decrease the number of repeats a lot, just that it would be possible to decrease them.

Even with these flaws, NAQT did a great job. Furthermore, this is the way all tournaments should be run. There is no point in having Maggie Walker and State College play a bunch of average and weak teams all day Saturday—it is not enjoyable for either side (State College was fairly bored beating up on us, and we were 8-1 going into the match)—and it would be unfair to place good teams against them in a division system. Furthermore, this system corrects teams being improperly ranked better than a division system.

I have not yet figured out how the last few rounds were matched up for bubble teams—if you are 5-4 going into your last match, then your match-up is a matter of life and death. Anybody who wants to look further into these things should be warned—the order matches are listed in is not always the same as the order they were played in. Pay attention to the round number on the left side of the page.
User avatar
BuzzerZen
Auron
Posts: 1517
Joined: Thu Nov 18, 2004 11:01 pm
Location: Arlington, VA/Hampshire College
Contact:

Post by BuzzerZen »

Wow. How long were you figuring this out? Now that the innards have been exposed, I'm even more impressed with the card system.
Evan Silberman
Hampshire College 07F

How are you actually reading one of my posts?
User avatar
Stained Diviner
Auron
Posts: 4841
Joined: Sun Jun 13, 2004 6:08 am
Location: Chicagoland
Contact:

Post by Stained Diviner »

I run a tournament on a similar system, so I'm pretty quick at sorting through this. I also have a one-year-old who insists on falling asleep each night over my shoulder, so I am forced to sit still for long periods.
Locked