HSQBRank 2013-2014: Witty Subtitle
Posted: Sat Jul 06, 2013 4:19 pm
Sponsored by the Partnership for Academic Competition Excellence (Twitter: @PACENSC)
https://www.hsquizbowl.org/forums/
Ladue's in a spot where I'd expect my method to fail to ID them as a top 25. YO SHADOW TOO BIG SON.Mewto55555 wrote:You appear to have left out Ladue (like, obviously this is pre-season and you're spitballing on little information so it doesn't matter, but with all due respect to only-senior-on-Ladue B David, plugging either of returning A-team members Ben/Jialin into his missing spot here would have turned a T-21 team EASILY into a t-8 returning all their members).
I find it quite ironic that LADUUUUUUUUE?! of all teams is complaining about their ranking.Mewto55555 wrote:You appear to have left out Ladue (like, obviously this is pre-season and you're spitballing on little information so it doesn't matter, but with all due respect to only-senior-on-Ladue B David, plugging either of returning A-team members Ben/Jialin into his missing spot here would have turned a T-21 team EASILY into a t-8 returning all their members).
Uh, we don't have a ranking.MorganV wrote:I find it quite ironic that LADUUUUUUUUE?! of all teams is complaining about their ranking.Mewto55555 wrote:You appear to have left out Ladue (like, obviously this is pre-season and you're spitballing on little information so it doesn't matter, but with all due respect to only-senior-on-Ladue B David, plugging either of returning A-team members Ben/Jialin into his missing spot here would have turned a T-21 team EASILY into a t-8 returning all their members).
Please tell us more about this super secret preseason ranking method.Grams's Go-Go Boots wrote:my method
Three Kentucky teams in the top 25?!Fred Ranks Things wrote:
9. DuPont Manual (KY)
10. Dunbar (KY)
24. Pikeville (KY)
It mostly consists of me bribing Fred Morlan.Christ, I Know wrote:
Please tell us more about this super secret preseason ranking method.
It's just PPB from HSNCT pro-rated by returning players' individual PPG, with performances from NSC (determined in the same manner) incorporated by comparison against performances there as appropriate and necessary.Christ, I Know wrote:Please tell us more about this super secret preseason ranking method.
Grams's Go-Go Boots wrote:It's just PPB from HSNCT pro-rated by returning players...Christ, I Know wrote:Please tell us more about this super secret preseason ranking method.
For some reason I had you guys losing Brandon. I'll go back and correct it later today. (It's the PPB that I care about, for the record).Mr. Scogan wrote:Grams's Go-Go Boots wrote:It's just PPB from HSNCT pro-rated by returning players...Christ, I Know wrote:Please tell us more about this super secret preseason ranking method.
Fisher Catholic returns everyone from a squad that had a higher HSNCT ppb then Carbondale, Menlo, State College, and Oak Ridge, and who also beat two teams you had in your top 25 for a while (Macomb and St. Joe's) while in Atlanta.
I'm inclined to agree. I'm pretty sure that Escobar will return 2/4 of their HSNCT team, as well as Jonathan Lam (who didn't attend HSNCT, but scared the hell out of us the one time we played them).DJCocoPuffs wrote:I don't know how to feel about these rankings. On one hand, it's a big step up from last year, where the Arcadia A I was on didn't get a pre-season ranking. On the other hand, Arcade Fire.
Anyway, I'm no expert our Northern counterparts, but I wouldn't have expected "Escobar" to rank so much lower than Menlo, given that Escobar returns a lot of decent players as far as I know, and they have a strong W-L track record against Menlo.
It's pretty inaccurate to assume that a team will lose a majority of its bonus points from one year to the next upon losing a player/players who get(s) the majority of the tossups for the team. At the really elite level, many players on the same individual team will all know the answers to the easy and middle parts of bonuses, so the hit to points per bonus will be pretty low. And that doesn't factor in the fact that many elite teams have players that overlap knowledge on tossups, too, so the remaining players can pick up a lot of what they've lost by buzzing one clue or two behind their top scorer on those same questions. (examples: State College from 2009 to 2010, GDS from 2009 to 2010, East Chapel Hill from 2012 to 2013, Loyola from 2012 to 2013). By design, this formula is definitely going to severely underrate teams like Ladue and Maggie Walker that return an excellent second-scorer and many solid B-teamers.Grams's Go-Go Boots wrote:It's just PPB from HSNCT pro-rated by returning players' individual PPG, with performances from NSC (determined in the same manner) incorporated by comparison against performances there as appropriate and necessary.Christ, I Know wrote:Please tell us more about this super secret preseason ranking method.
That's correct. The reason for that is that, presently, there's no way to adjust for those problems if you're stuck with the current stats that we produce. Not that SQBS is better, but look at this stat report for Ladue from NAQT. From this information, it's impossible to account for shadow effects. If you try prorating out the team's PPB by returning powers rather than just individual scoring, in case that was the key difference, it wouldn't work either; Ladue's returning PPB would actually be lower using only powers.RyuAqua wrote:At the really elite level, many players on the same individual team will all know the answers to the easy and middle parts of bonuses, so the hit to points per bonus will be pretty low. And that doesn't factor in the fact that many elite teams have players that overlap knowledge on tossups, too, so the remaining players can pick up a lot of what they've lost by buzzing one clue or two behind their top scorer on those same questions. (examples: State College from 2009 to 2010, GDS from 2009 to 2010, East Chapel Hill from 2012 to 2013, Loyola from 2012 to 2013). By design, this formula is definitely going to severely underrate teams like Ladue and Maggie Walker that return an excellent second-scorer and many solid B-teamers.
Code: Select all
Hunter 1 -> 2
Bellarmine 2 -> 1
Centennial 3 -> 3
Auburn 4 -> 5
St. Anselm's 5 -> 11
LASA 6 -> 19
duPont Manual 7 -> 29
Dunbar 8 -> 18
GDS 9 -> 15
IMSA 10 -> 4
University School of Nashville 11 -> 58 (produced exactly one stat line)
Dorman 12 -> 7
Carbondale 13 -> 14
Walt Whitman 14 -> 13
Loyola 15 -> 9
Ladue 16 -> 8
Seven Lakes 17 -> 6
Cistercian 18 -> 17
Mission San Jose 19 -> UR (I don't think I had results for a "Mission San Jose" team all year)
Thomas Jefferson 20 -> 12
Hunter B 21 -> 30
Richard Montgomery 22 -> 23
St. Paul Central 23 -> 92
Arcadia 24 -> 55
Olmsted Falls 25 -> 26
MSJ played as "Escobar" all of last year. You'll probably find many more stat lines for Escobar.Grams's Go-Go Boots wrote:The Mission San Jose situation was out of my control. USN produced one stat line, which isn't doing anyone any favors.
Yeah, at the time it wasn't clear if that was okay or not. I don't deal with undercover teams.Tanay wrote:MSJ played as "Escobar" all of last year. You'll probably find many more stat lines for Escobar.Grams's Go-Go Boots wrote:The Mission San Jose situation was out of my control. USN produced one stat line, which isn't doing anyone any favors.
At least in the case of St. Paul Central although the preseason rank of 23rd was way too high it wasn't as erroneous as suggested by the pre-nationals ranking which was based off three tournaments: at one we were a two-man team for almost half of the tournament, at another we were a two-man team with the second and third scorers missing and at the third tournament the top scorer and the third scorer were missing.Grams's Go-Go Boots wrote:So this leaves two teams in the last five of the rankings as ones I was way off of: St. Paul Central and Arcadia. I don't know of any excuses to be made here; both Central and Arcadia only played three tournaments that produced usable PPB, but Centennial, Dunbar and LASA A only played four, and unless that's the secret magic number of tournaments played that lets you know if a team's good, that excuse doesn't hold water. St. Paul Central finished tied for 49th at HSNCT and Arcadia finished tied for 33rd (neither played at NSC), both performances that were better than their pre-nationals ranking and closer to their pre-season ranking.
Unless the account gets suspended because I try to use Twitter for its apparent purpose, which is Tweeting at people. This is apparently the first step of many in which I lose touch with technology. If anybody needs me, I'll be playing N64.Grams's Go-Go Boots wrote:Tweeting, dude.
Maybe a picture of people playing quizbowl or something?Grams's Go-Go Boots wrote:And if someone wants to contribute a background image that isn't something that I made in 3 seconds with Paint to amuse myself, get in touch with me.
I think Mountain Lakes' performance at FAcT was factored in, that's all.Chunky Sean wrote:I'm just wondering how Mountain Lakes A got ranked when they were beaten by MAST by 400 points at PHSAT and then MAST was beaten by us (Wilmignton Charter B) by 85 points...additionally I couldn't help but notice that Torrey Pines was put down twice...
Torrey Pines issue is fixed.Chunky Sean wrote:I'm just wondering how Mountain Lakes A got ranked when they were beaten by MAST by 400 points at PHSAT and then MAST was beaten by us (Wilmignton Charter B) by 85 points...additionally I couldn't help but notice that Torrey Pines was put down twice...
Hmmmm...I see. Well, that explains it... #A-setLoveEBAcademicTeam wrote:I think Mountain Lakes' performance at FAcT was factored in, that's all.Chunky Sean wrote:I'm just wondering how Mountain Lakes A got ranked when they were beaten by MAST by 400 points at PHSAT and then MAST was beaten by us (Wilmignton Charter B) by 85 points...additionally I couldn't help but notice that Torrey Pines was put down twice...
This is a known bug on naqt.com (games without individual stats are lumped in with "bonus conversion"). We will at some point fix this, but it might not be soon.samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
Which teams would be affected by this?samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
When should I expect this problem to appear so I don't lose my mind by May?bt_green_warbler wrote:This is a known bug on naqt.com (games without individual stats are lumped in with "bonus conversion"). We will at some point fix this, but it might not be soon.samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
Just DCC/Charter A (you can see here that the final has no individual stats recorded).Grams's Go-Go Boots wrote:Which teams would be affected by this?samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
I mean in the more general sense: what kind of tournaments would be affected? What cases? Other statements ending in question marks?bt_green_warbler wrote:Just DCC/Charter A (you can see here that the final has no individual stats recorded).Grams's Go-Go Boots wrote:Which teams would be affected by this?samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
Tournaments in which some games have detailed stats available but some do not.Grams's Go-Go Boots wrote:I mean in the more general sense: what kind of tournaments would be affected? What cases? Other statements ending in question marks?bt_green_warbler wrote:Just DCC/Charter A (you can see here that the final has no individual stats recorded).Grams's Go-Go Boots wrote:Which teams would be affected by this?samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
I just sent you some details about this.Grams's Go-Go Boots wrote:When should I expect this problem to appear so I don't lose my mind by May?bt_green_warbler wrote:This is a known bug on naqt.com (games without individual stats are lumped in with "bonus conversion"). We will at some point fix this, but it might not be soon.samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
I know Tippecanoe in Ohio's A-set tournament has always had skewed bonus numbers due to the "no individual stats" in the playoff stage.Grams's Go-Go Boots wrote:I mean in the more general sense: what kind of tournaments would be affected? What cases? Other statements ending in question marks?bt_green_warbler wrote:Just DCC/Charter A (you can see here that the final has no individual stats recorded).Grams's Go-Go Boots wrote:Which teams would be affected by this?samus149 wrote:If Charter A and DCC A are higher than us due to PHSAT stats, it's because they got an extra 1.5 PPB or so on the NAQT website (NAQT messed up the data input for the final round). The entry on the database is more accurate, though not combined.
Thank you for letting me know about this issue. I've updated the rankings to reflect this corrected information.samus149 wrote:Regarding PHSAT: NAQT does a weird thing with the stats where the ppb comes out bad on their website. If that's what you used, the stats on the database are more accurate.
edit: tone