Bring the national poll back to life?

[bigtrain]

I was wondering what you all thought about bringing the national poll back. I have no idea how it should be done or who specifically should/can do it. Should the poll be limited to specific regions? Any ideas or comments?

[solonqb]

I'm sure everyone would like to see it, but Weiner only got six people to respond last time. I don't know who he polled, but apparently they got inured to answering questions.

I propose polling the tournament staff at universities that host HS tournaments as to what they think are the best high school teams they have seen.

Some potential respondents:

Michigan/Michigan State
Case
BGSU
Pitt
CMU
UVA
UMD
Yale
Northwestern
Illinois
WUSTL
Iowa State/Iowa
UF
UGA
Tulane
Vanderbilt
Duke
UTC
TAMU (Cromero)
UT-Austin (Kwartler, if he even runs HS tournaments)
Rice
Caltech
Stanford
USC
UCLA
Carleton
Mr. Hentzel (he's in a category all by himself, as I don't know whether he's affiliated with Minnesota officially)
Tulsa
Oklahoma State

Those are the ones that run major HS tournaments that come to my mind right now, if there are any others, same idea. Their addresses are all easily accessible on the Maize Pages. And not all of them would have be asked, just enough to ensure a good, balanced regional distribution. For example, the limited number of universities to ask in California holds down the number that can be asked in each region to ensure balance.

Another idea, perhaps combined with this, would be to email such a poll to the staff of the main Nationals (at least PACE and NAQT, perhaps ASCN, but NAC only if we're desperate) I exclude Panasonic because of its all-star nature.

Just some thoughts.

[AKKOLADE]

I'd like to see it done, but only if people realize it's not definite in any way. There is no way it can reasonably define the top 25 teams in order; the only places that can really do that are NAQT and PACE's national tourneys. If it's kept in mind that it's for entertainment purposes only, and that if, for example, Maggie Walker should end up higher than TJ, MW is not neccessarily the better team.

I think that with regards to whom to poll, it's important to hit various people - some hs coaches, some university folk and some reps of the two (three?) significant nationals. Also it'd be important that they receive feedback from people who, while may not be exposed to enough top teams to cast a ballot, have seen a few top regional teams and can compare them with others they have seen.

Just my thoughts.

[Captain Sinico]

The one problem with this is that most university teams don't care about and aren't involved very deeply in high school quizbowl. For example, nobody other than myself will likely see even a single high school match outside of our tournament all year, and I will see only perhaps 10 (assuming I help out at state finals again, and perhaps another tournament.) While your proposal has merit, I don't think it would work very well. A poll of coaches seems to be the only thing like a fair system (and, as has been noted, they are notoriously non-responsive.)

MaS

[Tegan]

The one problem with this is that most university teams don't care about and aren't involved very deeply in high school quizbowl. For example, nobody other than myself will likely see even a single high school match outside of our tournament all year, and I will see only perhaps 10 (assuming I help out at state finals again, and perhaps another tournament.) While your proposal has merit, I don't think it would work very well. A poll of coaches seems to be the only thing like a fair system (and, as has been noted, they are notoriously non-responsive.)
MaS

The impression I get is that at most university run tournaments, there is a bigger cross section of the state (or greater area) represented. Illinois, as an example, did not see anywhere near a large cross section of the best teams in the state...if I recall, they did see Auburn and DMac (certainly two of the very best)....but New Trier, Fenwick, Loyola, Fremd, Hoffman Estates, Marist, Lincoln-Way East, Latin, Evergreen Park, Timothy Christian, Maine South, Stevenson, etc, etc, etc weren't there. If it is similar across the coutnry, a poll of this type will miss a lot of good teams that are also involved teams (and just don't make it to the university run events).

BTW...if Illinois is looking to increase attendance, there may be a few dates opening up in the "traditionall cannonical calendar of Illinois tournaments"...you should contact someone at the Coaches Association for more info....if you moved the date to a little later in the year, you would get a turnout.

The East and Northeast might be able to do a regional poll....but outside of that, it likely has to go on a state-by-state basis (Illinois already has one in place)

[gosaints]

I think the best way to do it would be to poll coaches who have teams that compete a great deal in a given region (or that compete all over the place). They would have a good idea of the relative strengths of teams in their geographic area. Mr. Barry from Brookwood would probably be an excellent choice from the southeast.

There could also be a posting area where players/coaches make their case for being in the poll. This would help the voters rank some teams that are outside of their geographic area.

For example, I believe that every voter would consider putting TJ at the top, even if they haven't seen them play in person. A TJ person could list all of their accomplishments for the year and other people who have seen them play could pitch in their two cents.

There will also be some teams who have traveled out of their area who can give some idea about the relative strengths of different geographic regions. Our team has played in TN, IL, GA, and VA this year. We won the IL tournament but never made it past the final 8 in any of the other tournaments. Although this doesn't tell the whole story, it would give voters something to work with when comparing teams from the IL region with teams from the southeast or mid-atlantic region.

[dtaylor4]

There will also be some teams who have traveled out of their area who can give some idea about the relative strengths of different geographic regions. Our team has played in TN, IL, GA, and VA this year. We won the IL tournament but never made it past the final 8 in any of the other tournaments. Although this doesn't tell the whole story, it would give voters something to work with when comparing teams from the IL region with teams from the southeast or mid-atlantic region.

Here in Illinois, for some strange reason that can kiss my a**, there is a restriction on the number of dates teams can play during a season outside of state series (18), and after state series, schools themselves are only allowed to attend one tournament. In my league, we have four triangulars which count as four "dates" and a v and jv tourney, other than that we have been to tournaments almost every weekend since october, with the exception of thanksgiving, christmas, one in january which was cancelled, and last weekend (some decided not to come to practice, so we simply didn't go)

[Matt Weiner]

It's often been stated that, since Illinois defines their Scholastic Bowl activity as a five on five game, you can attend as many four on four quizbowl tournaments or out of state events as you want without falling under the IHSA jurisdiction. I don't know the inner workings of things in Illinois, but I know that teams have gotten around the restriction in the past with the above reasoning, so you may want to talk to a team that's done so if that's what's stopping you.

On the poll in general: Until I stopped the original poll, I was trying to get people to put their ballots together based mostly/entirely on the results posted on this site, since even the most active coaches don't see every contender. So, if it were to resume, I'd be more interested in people who were willing to take it seriously and look at those results than in people who are particularly well-traveled. This usually means active coaches and active collegiate tournament directors. and especially the ASCN people don't particularly know or care what goes on at tournaments they don't run; NAQT and PACE people have their ears to the ground enough that they could probably use the information on this site properly. If people really want to see a poll again, the first step is to make sure that there are enough results on this site to draw conclusions. For that end or just to help fulfill the main purpose of the site, please make sure every tournament held in your state this year is listed in the database with results.

[solonqb]

So does that mean you feel there aren't really enough results to go on yet? Dumb question, perhaps, but maybe if people get their act together by say, May you could pull off a pre-Nationals poll or something.

The requisite data exists scattered all over the Internet. It just needs people to input it here. I'll volunteer to do so if nothing gets done, but I'm sure people, given a goal like the above to work towards, could overcome their inertia in that regard.

[pblessman]

One of the main problems with national polls is that

1. There is too much data for most people and
2. There are not enough teams who travel to play out of state and out of region teams.

This makes polling very difficult. So what if quiz bowl tried a system similar to FIDE's (the International Chess Federation) rating system, based on the ELO system? They rank thousands and thousands of chess players across the world, and most of these chess players don't travel large distances either... and their ratings have real predictive power, i.e. one can use the ratings to predict the percentage probability that one player will beat another and this has been shown to work rather well. Magic and other trading card games also use a similar system... Input?

[Matt Weiner]

A chess game or a card game is always played under the same rules and conditions, though. We have to take into account the varying formats and writers in this poll. If someone wins a tournament on questions it means almost nothing; the results of even a smaller tournament on questions by a good high school or college program or NAQT are much more worthwile.

[pblessman]

To answer Matt's point: Yes, it would be ideal if each game was played under the same rules, but this wouldn't be addressed in a Poll either... If Bumble Bee High beats TJ in 's tournament, then pollsters would have to deal with it, and couldn't ignore it... the computer would also have to adjust the ELO rating appropriately. The good thing about the ELO ratings is that at least we would be trying to keep track and take into account as many tournaments as possible, and this would be done in a scientific fashion. If TJ's only loss is the one I described above, I'm sure the computer would still have them #1. (I hope there is no Bumblee High out there...)

P.S.: I might have to be more specific on how the ELO system works: Each team receivesa certain rating to start out with, e.g. 1000. After each match, you gain points for a victory and lose points for a loss. The amount of points gained or lost depends on how many points differential there are between the two teams. This should ensure that the tope teams are ones that play a lot of matches and win many of them.

[Matt Weiner]

I would be fairly confident in assigning a field strength and question/format reliability rating to each tournament based on my own subjective but informed experience, if we really wanted to do this mathematically, but it still requires knowing certain things about each tournament. If people can make sure the data is there I'm willing to try out both a traditional and computerized poll to see what happens.

[bigmac]

The problems of data availability, to say nothing of reliability and completeness, and the difficulty of ranking teams that have never faced each other are real stumbling blocks for a national poll with any claim to validity (after the #1 spot, of course, though only for this year). Formulas and weighted formats do not seem to me a convincing basis for generating a top 20 list when no one has seen even a quarter of the teams. How well would this system, even if the data were complete, have predicted the two nationals last year? Would it have predicted that so many teams from one circuit would have done well at PACE? (I am not convinced that the Mid-Atlantic is quite as strong this year, btw).

Perhaps a human poll within each region coupled with a ranking of regions would make more sense. Coaches and/or top players could look at summaries of recent data (say, top eight finishes at the last 8-10 tourneys) and could check the data against their own sense of teams they actually have seen a few times. Then others (college players with a sense of hsqb, teams that travel a good bit or teams that are in between two regions) could rank the regions, which would also have to be defined a bit more clearly.

This system would avoid the impossible discussions about whether or not the #2 team from Michigan could beat the #5 team from the South. "We should be 4 not 5" arguments about a particular team could stay within a forum that could look to head-to-head performance. The regional comparisons would still be inexact, but would give teams a sense of what they are really looking for in a national poll: how might they fare against the best team from a few states over. Hey, if an entire region felt underrated, then its schools might get fired up enough to go out of state a bit more.

[Matt Weiner]

I certainly understand where you're coming from and some of what you said is why the poll was discontinued in the first place. However, to respond to a few points:

How well would this system, even if the data were complete, have predicted the two nationals last year? Would it have predicted that so many teams from one circuit would have done well at PACE?

This would be the test of any poll, and in fact I would like to take the available pre-nationals data and see what sort of formula produces something close to the nationals results, then weight that information heavily in creating the final formula.

Perhaps a human poll within each region coupled with a ranking of regions would make more sense.

I've never seen the point of regional/state polls. Don't teams from the same area play each other in tournaments often? Those are the ultimately correct rankings within the region.

This system would avoid the impossible discussions about whether or not the #2 team from Michigan could beat the #5 team from the South.

Well, that's the point of doing the poll at all, and at least given the proper parameters, the question is answerable. For example, if Team A has 30 powers and 20 PPB bonus conversion on the same NAQT set as Team B, which had 10 powers and 15 PPB, then I think we can confidently say that Team A is superior at the NAQT format and would beat Team B most of the time in NAQT matches, because those stats are largely independent of field strength and other non-common factors. The ideal of a computerized poll is to correct for field strength, question quality, etc, and make that sort of comparison for everyone. I'm sympathetic to the view that this is just not possible, and I'm not entirely convinced that it is, but that's the idea.

[STPickrell]

Then, of course, you have the problem of teams who only appear occasionally (often in a state series), but when you do see them, they are highly impressive.

Let's look at Jefferson Forest High School, located in scenic Forest, VA, which is near Lynchburg (although as their coach joked, "We don't claim Lynchburg.") They have been absent from tournaments, so we had no way of judging them before they appeared at VHSL States.

If I could get results from District and Region people, we could judge them a bit better, but I've had a devil of a time trying to get results from people (this year, I got most of the Regions, and I'm going to put in financial incentives to get Districts to send me their results.)

At States, they defeated eventual winner Charlottesville 205-155 and lost to Spotswood in a 165-155 overtime thriller. They also destroyed Abingdon 250-90 and got blown out by Charlottesville 260-150 (after taking two rounds off) to exit the tournament.

I know our questions aren't perfect, but I don't think anyone could accuse them of being Beall Bowl. I'd like to think they are legit enough that good performances in the VHSL series mean something. But what value, exactly, do we give them?

[First Chairman]

Or as I like to point out, it's how dancing is done. :)

I'll admit part of my own analysis consists of a "proficiency points" system similar to how rankings in amateur ballroom dance is performed. I don't completely co-op it but it's similar.

Basically what is most important to me is the final placement of teams in a given tournament. Then I take into account other secondary factors such as game format, number of teams, competitiveness of playoff field.

What the points system gives is a merit-based record over the history of a period of time. The disadvantage is that you do reward people who go to a ton of tournaments and leave many really good teams that cannot afford to travel so much behind.

Granted, in the ballroom world, you don't qualify for nationals in any event until you have gotten enough "proficiency points" at various events throughout the year. Points in that case are awarded based on the number of competitors and the stringency of selection to placements.

[Zip Zap Rap Pants]

you know a big despute people would have over the national rankings would be whether or not to include b teams. a team that would normally be 23rd might get mad at being knocked off by b teams, but on the other hand tj b is probably around 12th in the country so you cant really overlook them. i mean just look at the duke results: tj a, mlw a, tj b, tj c, mlw b, raleigh charter a. and then i heard tj b beat tj a at gonzaga's tournament, though neither had their regular lineups. that's another thing, if any team doesnt have their regular lineup for whatever reason, then they'll drop down in rankings and is that really fair?

well maybe the solution is this: have a coach's national poll, then have a tournament director's national poll which includes college students who are really informed, as well as national tournament staffs. then combine them, and factor in strength of tournament fields and quality wins into a computer to generate the rankings in a system called the Tournament Championship Series (tcs). then we can put the top 25 in naqt nationals and throw in automatic bids for state champions (even in the worst states like south dakota), thus rendering regular invitational tournaments almost meaningless.....

or we could just stick with what we have, just a thought