Page 1 of 1

Postseason Poll

Posted: Mon Apr 29, 2013 11:08 am
by Cheynem
In a little while, I will put up on the front page and in this thread the guidelines and deadline for the postseason collegiate poll. Now, some of you may ask "what's the point of this poll? didn't we just have two nationals to shake out who's best?" I would say that the nationals showed the complexity involved in ranking the teams, even after the national tournaments. We had two very, very, very good, very deserving winners (Yale and Illinois). We had very strong second and third place teams at ICT and Nats (UVA, Penn, Michigan). We had teams that played one nationals and did very well that didn't play the other, or who played one nationals shorthanded, or who played ACF and then ICT DII. Trying to make sense of these results is what the postseason poll is about. It is effectively the community's judgment now that everyone has played the national tournaments, who the top 25 teams are.

The rules from the last poll are still in effect--you must vote for 25 teams, no incomplete ballots, and you cannot vote for HS teams. You are comparing "best possible teams within reason," not necessarily the results of shorthanded teams (yet without just making up players who didn't play anything all year). You can vote for B, C, D teams, etc.

Now, I have a few questions I'd like to get feedback on regarding the criteria. The first is, "should you have had to play either ICT or ACF Nationals in order to get considered here?" The most glaring team this would affect is Brown, a consistent Top 15 pick. I only ask this because it is extremely difficult to compare teams as it is, so teams that did not play either of the nationals seem even tougher to compare. Secondly, I'd generally like to ask if there are any other criteria or rule restrictions that would improve such a poll.

Thanks.

Re: Postseason Poll

Posted: Mon Apr 29, 2013 2:10 pm
by Frater Taciturnus
I'm a big fan of the "in order to be considered a quizbowl team, you have to actually play quizbowl" rule, and I would like to hopefully use whatever hilariously minuscule clout I have here to urge voters to vote accordingly.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 1:43 pm
by Windows ME
Is it taboo if I make my ballot public? I was going to originally but realized nobody did it in the last thread.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 2:43 pm
by Cheynem
I would prefer not to just because I think it unduly influences voting. For one, the first person who posts a public ballot intentionally or not tends to set a tone. People who follow that person work through that person's teams, either including them because they don't know any better or specifically including or excluding teams in response (I don't think any of this is even consciously so, for the most part).

If you have particular points you wish to make (like I think Team Blank is underrate or overrated), I think that's fine to talk about.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 4:49 pm
by gyre and gimble
Frater Taciturnus wrote:I'm a big fan of the "in order to be considered a quizbowl team, you have to actually play quizbowl" rule, and I would like to hopefully use whatever hilariously minuscule clout I have here to urge voters to vote accordingly.
I mean, if you take that rule literally then no one could argue, but I think the wider application of this principle should be qualified. For example, I think Brown is better than a good number of the teams who finished in the Top 25 at each nationals, and they did "actually play quizbowl" at least a few times during the year, so they are an actual quizbowl team and should be considered so. I plan to include them on my ballot. On the other hand, if there's some doubt between how you should rank Team A relative to Team B, where the two are similar in skill but only Team B played nationals (and I guess performed at around the level you expected Team A to preform), you should probably rank Team B ahead of Team A since they actually proved themselves.

This isn't based on any real data but I get the sense that in the past, people have tended to do this anyway.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 5:14 pm
by theMoMA
I'm not sure if this will sway anyone's opinion, since people tend to just vote the way they've voted all year instead of actually incorporating ICT/Nats finishes into their thinking, but I thought that the end-of-season tournaments really showed that the Midwest was underrated as a quizbowl circuit this year. I think that Illinois, Michigan, and Minnesota all showed that they were higher caliber teams than they were given credit for throughout the season (and Chicago's two teams lived up to their rankings, for the most part).

While ICT had the expected Yale-UVA-Penn triumvirate atop the standings, NAQT is not the format we play most of the year. (And Illinois/Minnesota/Michigan managed a very respectable 4-6.) The fact that Midwest teams managed to snag two of the top three spots (and the championship) at Nationals is, to me, indicative of the quality of these teams throughout the year.

I think Yale was the most consistently excellent team all season, but I would be disappointed to see Illinois on the outside of the top two, and I think Michigan deserves a lot of consideration for the #3 spot as well. I understand that we at UMN didn't play our full lineup at most tournaments, but I think we consistently hung in against (and occasionally beat) Michigan and Illinois during the regular season and showed what our full lineup could do at ICT.

The top eight this year is pretty clear-cut for me (you have Yale, UVA, Penn, and Maryland in the East and Illinois, Michigan, Minnesota, and Chicago in the Midwest), but I get the feeling that the way in which the voters will blend these teams won't jive with how they actually managed to play each other at the end-of-year tournaments.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 5:15 pm
by Cheynem
Since I amusingly am already starting to get ballots, I will rule this:

A team that has played at least one regular difficulty or above collegiate tournament (anything above ACF Fall) qualifies for the purpose of the poll. Voters may decide on their own how to deal with teams that did not play the national tournaments or only played one.

I will post full rules/explications on the frontpage of the site and here shortly, with a set deadline.

Re: Postseason Poll

Posted: Tue Apr 30, 2013 6:07 pm
by Frater Taciturnus
Class ended early today, so here is every team that played DI ICT or ACF Nationals ordered by an average of their finishes. Perhaps not even a good metric, but it is a thing.

https://docs.google.com/spreadsheet/ccc ... FbkE#gid=0

Re: Postseason Poll

Posted: Tue Apr 30, 2013 7:32 pm
by Auks Ran Ova
theMoMA wrote:I think we consistently hung in against (and occasionally beat) Michigan
Psh, that doesn't mean anything; they lost to a community college!

Re: Postseason Poll

Posted: Tue Apr 30, 2013 8:39 pm
by itsthatoneguy
Ukonvasara wrote:
theMoMA wrote:I think we consistently hung in against (and occasionally beat) Michigan
Psh, that doesn't mean anything; they lost to a community college!
:-(

Re: Postseason Poll

Posted: Wed May 01, 2013 10:56 am
by Cheynem
[To be crossposted soon]

Ballots for the postseason collegiate poll are now open. They will be due Friday, May 10, at noon.

Please rank 25 collegiate teams (not high school). The criteria is something like the midseason poll--comparing the teams on nationals-level difficulty questions, with the logical realization that now we have more data, including many rounds in which teams did play each other on nationals-level difficulty. I don't expect or require people to just slavishly copy or average teams' finishes at ICT/Nats, but I would also obviously encourage some logical analysis of those results and not just sending in a midseason balllot again.

You can vote for any team that has played a regular season collegiate tournament all year, with the caveat that if you want to ignore (for lack of data) or penalize teams that did not play either nationals, I cannot stop you from doing so.

Please look over your ballots before sending it in to make sure you are ranking all the teams you wish to rank.

Re: Postseason Poll

Posted: Wed May 01, 2013 11:19 am
by gyre and gimble
If there's going to be a player poll--I'd like to see one happen, if just to find out how people think of players who led their teams to championships (Ike and Matt J) relative to those who swept the 1st place votes last year (Eric and Matt B)--I think it would be a good idea to run it concurrently with this one just so people don't have to look through a bunch of stats multiple times.

Re: Postseason Poll

Posted: Wed May 01, 2013 11:21 am
by Cheynem
Yeah, I can start that too.

Re: Postseason Poll

Posted: Wed May 01, 2013 11:26 pm
by Fond du lac operon
Frater Taciturnus wrote:Class ended early today, so here is every team that played DI ICT or ACF Nationals ordered by an average of their finishes. Perhaps not even a good metric, but it is a thing.

https://docs.google.com/spreadsheet/ccc ... FbkE#gid=0
I think this is a decent starting point for a ballot, but worth noting that at ICT, Ohio State beat Chicago A in a tiebreaker to make the top bracket, and Harvard beat Berkeley in a tiebreaker to make the second bracket. And at ACF, Minnesota beat Alabama in a tiebreaker for the top bracket. So those six teams in particular could have ended up varying in their final standings quite a bit.

Re: Postseason Poll

Posted: Wed May 01, 2013 11:58 pm
by Cheynem
Bear in mind, of course, that "Minnesota" was just Andrew Hart.

Re: Postseason Poll

Posted: Thu May 02, 2013 12:46 am
by theMoMA
Yeah, I'm not really sure that we should be considered as playing ACF Nationals at all for the purposes of averaging the finishes, considering it doesn't really reflect on the quality of our team the whole year.

Re: Postseason Poll

Posted: Thu May 02, 2013 1:53 am
by The Ununtiable Twine
theMoMA wrote:Yeah, I'm not really sure that we should be considered as playing ACF Nationals at all for the purposes of averaging the finishes, considering it doesn't really reflect on the quality of our team the whole year.
If only you weren't there... :party:

Re: Postseason Poll

Posted: Thu May 02, 2013 10:31 am
by Periplus of the Erythraean Sea
theMoMA wrote:Yeah, I'm not really sure that we should be considered as playing ACF Nationals at all for the purposes of averaging the finishes, considering it doesn't really reflect on the quality of our team the whole year.
I'd say the same for our team, except our situation was sort of the reverse of yours.

Re: Postseason Poll

Posted: Thu May 02, 2013 10:49 am
by gyre and gimble
Fond du lac operon wrote:
Frater Taciturnus wrote:Class ended early today, so here is every team that played DI ICT or ACF Nationals ordered by an average of their finishes. Perhaps not even a good metric, but it is a thing.

https://docs.google.com/spreadsheet/ccc ... FbkE#gid=0
I think this is a decent starting point for a ballot, but worth noting that at ICT, Ohio State beat Chicago A in a tiebreaker to make the top bracket, and Harvard beat Berkeley in a tiebreaker to make the second bracket. And at ACF, Minnesota beat Alabama in a tiebreaker for the top bracket. So those six teams in particular could have ended up varying in their final standings quite a bit.
I'd say this was a result of poor bracketing that severely underestimated Ohio State. With a team that fit the 17th seed description better in our bracket, I don't think Chicago, us, or Berkeley would have had trouble at least making the second bracket. I think Berkeley won the third bracket easily and we only lost once in ours.

What people should really take into account is margins of victory. Ohio State lost to Berkeley by 5, forcing the tiebreaker in the first place. Likewise, we lost to Chicago and Ohio State by a combined 80, forcing the tiebreaker with Berkeley. Both tiebreakers, I think, ended with pretty big margins.

I can't speak to the Alabama and Minnesota thing but I think each team's performance in the playoffs gives a good predictor of what might have happened had Alabama won the tiebreaker.

Unrelatedly, how did Virginia, Harvard, Ohio State, and Berkeley end up in the same prelim bracket at ACF? Come on. I'll also note we were down one Graham Moyer and up one Mark Arildsen at ACF. Not sure how that affected things in the long run for us.

Re: Postseason Poll

Posted: Thu May 02, 2013 11:24 am
by Cheynem
Ten ballots are already in.

Re: Postseason Poll

Posted: Thu May 02, 2013 11:28 am
by Sima Guang Hater
gyre and gimble wrote:What people should really take into account is margins of victory.
Yeah I want to emphasize this point. I'd also like to emphasize that a full body of stats should be considered when available, rather than just performances at nationals. The poll shouldn't just recapitulate the results of the two nationals; people should ideally compare all the pairwise matchups between two teams and look at the margins, the effects of bad questions, etc.

Re: Postseason Poll

Posted: Thu May 02, 2013 11:30 am
by Tanay
The Quest for the Historical Mukherjesus wrote:The poll shouldn't just recapitulate the results of the two nationals; people should ideally compare all the pairwise matchups between two teams and look at...the effects of bad questions, etc.
How exactly is this possible, and how ought this be calculated and taken into account?

Re: Postseason Poll

Posted: Thu May 02, 2013 11:52 am
by Sima Guang Hater
Tanay wrote:
The Quest for the Historical Mukherjesus wrote:The poll shouldn't just recapitulate the results of the two nationals; people should ideally compare all the pairwise matchups between two teams and look at...the effects of bad questions, etc.
How exactly is this possible, and how ought this be calculated and taken into account?
Say you want to compare UVA and Yale. Since these are two teams that don't clash much, you have only a handful of sources of information:

1. Playing against both teams/general feel.
2. Statistics
3. Head to head matchups

I can't tell you how to do 1. For 2, looking at Jacob's stats compilation is enough (you'll see that UVA almost universally has better ppb and power numbers). For 3, look at ICT and ACF nationals (which would tell you that the series is Yale 2-1 UVA, I believe). Then weigh these as you see fit.

Keep in mind that single results and single games are only a single draw from a bell curve of potential outcomes. Taking into account other factors will help mitigate potential sampling error.

Re: Postseason Poll

Posted: Thu May 02, 2013 11:54 am
by Tanay
The Quest for the Historical Mukherjesus wrote:
Tanay wrote:
The Quest for the Historical Mukherjesus wrote:The poll shouldn't just recapitulate the results of the two nationals; people should ideally compare all the pairwise matchups between two teams and look at...the effects of bad questions, etc.
How exactly is this possible, and how ought this be calculated and taken into account?
Say you want to compare UVA and Yale. Since these are two teams that don't clash much, you have only a handful of sources of information:

1. Playing against both teams/general feel.
2. Statistics
3. Head to head matchups

I can't tell you how to do 1. For 2, looking at Jacob's stats compilation is enough (you'll see that UVA almost universally has better ppb and power numbers). For 3, look at ICT and ACF nationals (which would tell you that the series is Yale 2-1 UVA, I believe). Then weigh these as you see fit.

Keep in mind that single results and single games are only a single draw from a bell curve of potential outcomes. Taking into account other factors will help mitigate potential sampling error.
Well, sure. This all makes sense. I thought you meant that there was a specific way of approaching the presence of bad questions that could have come up in specific rounds.

Re: Postseason Poll

Posted: Thu May 02, 2013 11:57 am
by Sima Guang Hater
Tanay wrote:Well, sure. This all makes sense. I thought you meant that there was a specific way of approaching the presence of bad questions that could have come up in specific rounds.
Oh I didn't realize that's what you were asking me. This is hard to do unless you were there, I suppose.

Re: Postseason Poll

Posted: Thu May 02, 2013 12:16 pm
by Cheynem
I thought the Yale/UVA series was 2-1, Yale. Yale with a win at ICT and a win at Nats, UVA with a win at ICT.

Re: Postseason Poll

Posted: Thu May 02, 2013 12:20 pm
by The King's Flight to the Scots
Cheynem wrote:I thought the Yale/UVA series was 2-1, Yale. Yale with a win at ICT and a win at Nats, UVA with a win at ICT.
Right, that's what Yale 2-1 UVA means.

Re: Postseason Poll

Posted: Thu May 02, 2013 12:21 pm
by Cheynem
Oh I'm sorry, I can't read.

Re: Postseason Poll

Posted: Thu May 02, 2013 12:45 pm
by The Ununtiable Twine
gyre and gimble wrote:I can't speak to the Alabama and Minnesota thing but I think each team's performance in the playoffs gives a good predictor of what might have happened had Alabama won the tiebreaker.
This is why I think using the entire season as an indicator is a fine idea. We had a bad opening round, lost by 10, and finished 16th instead of posting what would have inevitably been anywhere from 3-8 to 1-11 in the top bracket playoffs and earning some place between 10th and 12th, most likely 12th because that's what we do. It's just an example how one buzzer race can have a large impact if you vote directly from George's metric, which I expect exactly no one to do, of course. It isn't such a bad way to start your ballot off, moving teams up and down as you see fit. I will say that we spent all of February barely losing to Georgia Tech, a team that had some impressive wins at Nationals (some by equally impressive margins) and were impressive in defeat the majority of the time, so looking back at the season might not be such a bad idea especially in our case.

Re: Postseason Poll

Posted: Thu May 02, 2013 1:07 pm
by marnold
gyre and gimble wrote: I think Berkeley won the third bracket easily.
Um, no? What are you talking about? It was a circle of death in the playoffs with them, us and Rice all with one loss, with our game against them decided in extra tossups.

Re: Postseason Poll

Posted: Thu May 02, 2013 1:10 pm
by gyre and gimble
marnold wrote:
gyre and gimble wrote: I think Berkeley won the third bracket easily.
Um, no? What are you talking about? It was a circle of death in the playoffs with them, us and Rice all with one loss, with our game against them decided in extra tossups.
My bad, I didn't have the results open and for some reason I confused Berkeley at ICT and Stanford at ACF.

Re: Postseason Poll

Posted: Mon May 06, 2013 1:30 pm
by Cheynem
There are currently 20 ballots ranking a total of 35 teams.

Re: Postseason Poll

Posted: Fri May 10, 2013 12:31 pm
by Cheynem
The results. There were 26 ballots.

Voters: Will Alston, George Berry, Billy Busse, Bryce Durgin, Zach Foster, Jarret Greene, Saul Hankin, Andrew Hart, Ike Jose, Jasper Lee, Patrick Liao, Austin Listerud, Stephen Liu, Saajid Moyen, Eric Mukherjee, Will Nediger, Joe Nutter, Aaron Rosenberg, Jacob O'Rourke, Adam Sperber, Marshall Steinbaum, Jake Sundberg, Sinan Ulusoy, Jerry Vinokurov, Matt Weiner, and Richard Yu.

I will post overall point totals and then the highest and lowest rankings they received on a ballot. If there is no lowest, it means that they were unranked on at least one ballot. I will also mark how this placement differed than the last poll.

1. Yale A (636, highest: #1, lowest: #4), +2 (16 first place votes)
2. Illinois (621, highest: #1, lowest: #4), +3 (6 first place votes)
3. Virginia A (607, highest: #1, lowest: #5), -2 (4 first place votes)
4. Penn (572, highest: #2, lowest: #5), -2
5. Michigan (554, highest: #3, lowest: #5), -1
6. Maryland (494, highest: #6, lowest: #9) and Minnesota (494, highest: #6, lowest: #10). Maryland went up 1, Minnesota went up 2.
8. Chicago A (482, highest: #6, lowest: #9), -2
9. Georgia Tech (443, highest: #6, lowest: #12) (the same)
10. Harvard A (391, highest: #8, lowest: #21) (the same)
11. Ohio State (350, highest: #8, lowest: #19), +4
12. Chicago B (343, highest: #10, lowest: #17) (the same)
13. Alberta (335, highest: #10), +7
14. Stanford (297, highest: #11), +10
15. MCTC (245, highest: #9), +7
16. NYU (237, highest: #11, lowest: #24), -2
17. Alabama (221, highest: #14, lowest: #22), -1
18. Rice (185, highest: #13) (the same)
19. MIT (178, highest: #13), -6
20. Kings College London (157, highest: #10), wasn't ranked
21. Brown A (142, highest: #10), -10
22. Berkeley (125, highest: #17, lowest: #25), -1
23. Columbia (116, highest: #16), -6
24. Michigan State (102, highest: #18), -5
25. WUSTL (36, highest: #22), wasn't ranked

Also Receiving Votes: Dartmouth (24), VCU (20), Virginia B (19), Northwestern (9), Harvard B (6), Brown B (5), Yale B (5), George Mason (5), Cornell (1), Ottawa (1), and Rutgers (1).

Thanks everyone for voting.