Page 2 of 3

Re: HSQBRank 2016-2017

Posted: Sun Mar 26, 2017 4:13 pm
by 1.82
Max Schindler provided this helpful form several years ago:
Dear Fred,

While I admire the much effort you have put into your rankings, you seem to be chronically underrating [insert team you are affiliated with here]. However, you seem to be favoring [your rival], despite the fact that we have beaten them [X] out of [Y] times we have played them, and of those [Y-X] times we lost, [some integer less than or equal to Y-X] were because [your best player if a coach/the word "I" if a player] was suffering from [make up a word that sounds like a tropical disease]. You also fail to notice that in [insert tournament here] our PPB was within [insert a number here] despite missing [insert player name here]. Also, we once won [insert crappy local tournament] by a score of [insert a number which probably are not divisible by 5, then a -, then another such number]. I realize that none of these rankings actually matter, but I am really disappointed you are being so blatantly unfair to [your team]. Hopefully we will show our true strength at [HSNCT or NSC] once [player name] has gotten over his [Ebola or Trypanosomiasis].

-[your name here]

Re: HSQBRank 2016-2017

Posted: Sun Mar 26, 2017 5:59 pm
by AKKOLADE
Shomik Ghose wrote:You can try looking at the "bigger package" (powers per game, points per game, negs per game, G/N, P/N, etc...), not just ppb. You could also try getting regional coordinators in each area who could notify you if some team on the list has played without their best player or split (for example, on your most recent rankings, St. Johns moved to 17th place but the ppb used was from when Andy Huff and I, who are the two highest scorers on the team, split. Likewise, Taylor moved to 34th place but the ppb used was from when William wasn't playing). I know you're trying your best, and I'm not trying to criticize or demean you in any way; I'm just offering suggestions as to how you could improve the rankings.
Actually, the best aPPB St. John's had is from October with a team that included yourself and Andy Huff.

The reason Taylor fell down from the rankings is because their performance hadn't matched their preseason projections to date. Their best performance in the last set of rankings is here. William Golden played for them.

At some point, though, the onus is on a team to play its full lineup. The purpose of the rankings is to provide as unbiased of a ranking as possible, which is why I do things the way I do. I'm not going to include preseason projections in a set of rankings published in February.

It is March 26. There's essentially one month left in the season. These are quiz bowl rankings. If you want accurately ranked in the quiz bowl rankings, then playing quiz bowl is a very good way to accomplish that.

To address additional points raised:
  • If your primary issue with the rankings is that I didn't properly account for the fact that a team did not have its best players compete, then no extant statistic will properly account for that if you base it completely off of actual performances.
  • My rankings are not plugged in as the seeding for NSC. They are looked at. Various other factors are also looked at as well. Our seeding process is similar to that for the HSNCT.
  • By the end of the season, I do look at powers per game as a factor in the rankings. It is a minor factor. Additionally, it's not a factor that will make a short-handed team look better.

Re: HSQBRank 2016-2017

Posted: Sun Mar 26, 2017 8:10 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Mon Apr 03, 2017 9:29 pm
by AKKOLADE
New rankings later this week.

Re: HSQBRank 2016-2017

Posted: Tue Apr 04, 2017 8:38 am
by Hyrdofluoric_Acid
I don't know if this is asking too much but can I see the stat adjustments for the sets this season? I am extremely curious about this.

Re: HSQBRank 2016-2017

Posted: Tue Apr 04, 2017 8:52 am
by the return of AHAN
Hyrdofluoric_Acid wrote:I don't know if this is asking too much but can I see the stat adjustments for the sets this season? I am extremely curious about this.
I'm extremely curious about how many pieces of luggage you intend to pack this weekend. (Check your e-mail, pls)

Re: HSQBRank 2016-2017

Posted: Tue Apr 04, 2017 11:39 am
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Thu Apr 06, 2017 10:50 am
by dhumphreys17
What, out of curiosity, is the current stat adjustment for POMMSS, if I may?

Re: HSQBRank 2016-2017

Posted: Thu Apr 06, 2017 11:10 am
by Ciorwrong
dhumphreys17 wrote:
What, out of curiosity, is the current stat adjustment for POMMSS, if I may?
Hey I was going to ask that, Devin! It seems from my cursory glance that teams average like 3 ppb less than they would on an IS set but I don't have the adjustment formula.

Re: HSQBRank 2016-2017

Posted: Sun Apr 09, 2017 8:45 pm
by AKKOLADE
I've been playing around with 2016 data and that year's HSNCT results. What I've found is - unsurprisingly - that bigger of a difference there is between the two teams' best aPPB from the year, the less likely there is an upset as defined by aPPB.

Here's what I found:

0-1 aPPB difference: 46.3% chance of an upset (130 upsets out of 281 total games)
1.01-2: 36.4% chance (96/264)
2.01-3: 28.2% (60/213)
3.01-4: 16.4% (31/189)
4.01-5: 13.8% (16/116)
5.01-6: 20.5% (16/78)
6.01-7: 12.2% (5/41)
7.01-8: 6.1% (2/33)
8.01-9: 0% (0/18)
9.01-10: 5.9% (1/17)
10+: 5.7% (2/35)

I excluded teams that didn't have an aPPB.

Re: HSQBRank 2016-2017

Posted: Mon Apr 10, 2017 8:44 am
by Monstruos de Bolsillo
AKKOLADE wrote: 6.01-7: 12.2% (5/41)
7.01-8: 6.1% (2/33)
8.01-9: 0% (0/18)
9.01-10: 5.9% (1/17)
10+: 5.7% (2/35)

I excluded teams that didn't have an aPPB.
Out of curiosity, which games were the upsets mentioned here?

Re: HSQBRank 2016-2017

Posted: Mon Apr 10, 2017 1:19 pm
by scottkim
Monstruos de Bolsillo wrote:
AKKOLADE wrote: 6.01-7: 12.2% (5/41)
7.01-8: 6.1% (2/33)
8.01-9: 0% (0/18)
9.01-10: 5.9% (1/17)
10+: 5.7% (2/35)

I excluded teams that didn't have an aPPB.
Out of curiosity, which games were the upsets mentioned here?
I'm willing to bet one (or more) of them involve Collins Hill B. Most of the team that earned the pre-nationals ranking couldn't attend due to the fact that our B team was on the standby list and our graduation was the day before HSNCT.

Re: HSQBRank 2016-2017

Posted: Mon Apr 10, 2017 4:18 pm
by AKKOLADE
Monstruos de Bolsillo wrote:
AKKOLADE wrote: 6.01-7: 12.2% (5/41)
7.01-8: 6.1% (2/33)
8.01-9: 0% (0/18)
9.01-10: 5.9% (1/17)
10+: 5.7% (2/35)

I excluded teams that didn't have an aPPB.
Out of curiosity, which games were the upsets mentioned here?
Jenks 195, Leland 175 (aPPB diff 6.16)
American Heritage Boca-Delray 275, Irvine 225 (6.19)
Norman B 90, Bishop Heelan 85 (6.60)
Western 190, Washington B 105 (6.67)
Cloverleaf A 100, Collins Hill 85 (6.87)
Belvidere North B 95, J. M. Tate 85 (7.19)
Cave Spring B 290, Davis 180 (7.52)
Cloverleaf B 85, Olentangy Liberty B 80 (9.9)
North Hagerstown 165, Bishop Heelan 120 (11.91)
North Hagerstown 145, Washington B 105 (14.75)

There's a ton of caveats with the above, but that's just straight up, context-free data.

Re: HSQBRank 2016-2017

Posted: Mon Apr 10, 2017 8:24 pm
by scottkim
AKKOLADE wrote: Cloverleaf A 100, Collins Hill 85 (6.87)
This was indeed our B team.

Re: HSQBRank 2016-2017

Posted: Mon Apr 10, 2017 9:55 pm
by AKKOLADE
Yeah, there's several short comings with the data as is. Additional work into clarifying if these were the "real" teams would reduce the number of upsets, but that's a low priority for me.

Re: HSQBRank 2016-2017

Posted: Sun Apr 16, 2017 12:57 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Wed Apr 26, 2017 8:32 pm
by AKKOLADE
I take a look at the Charter & Private Schools division of the NAQT SSNCT. https://hsqbrank.com/2017/04/26/naqt-sm ... e-schools/

Re: HSQBRank 2016-2017

Posted: Thu Apr 27, 2017 7:09 pm
by AKKOLADE
AKKOLADE wrote:I take a look at the Charter & Private Schools division of the NAQT SSNCT. https://hsqbrank.com/2017/04/26/naqt-sm ... e-schools/
I've updated the above to include stats for AFSA, Lionsgate, and Toledo Arts.

Re: HSQBRank 2016-2017

Posted: Thu Apr 27, 2017 8:22 pm
by AKKOLADE
Also, there's a Facebook page now!

Re: HSQBRank 2016-2017

Posted: Fri Apr 28, 2017 1:43 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 6:44 am
by tmcavoy
Can you do an update of small schools ranking now that SSNCT is over?

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 12:50 pm
by Antrobus63
I don't believe that Fred can or should have to account for who's playing at a given tournament. If you declare that your "A" team is playing (or don't put a letter on it) then that's your team. He's dealing with hundreds of teams, so it's just not practical for him to comb through rosters. Even if he did, how exactly would he quantify the relative contribution of each player given the competition, the particular set... ? The mind boggles.

On the other hand since, by now, most everyone knows that Fred's rankings don't always reflect the true strength of teams that don't always play their entire roster, or simply don't play as much as other teams, I am certainly in favor of coaches' and players' polls. If these polls are conducted on a regular basis, like Fred's, then the QB world will have a more nuanced, rounded picture of the actual playing strength of teams.

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 1:29 pm
by Father Comstock
Antrobus63 wrote:I don't believe that Fred can or should have to account for who's playing at a given tournament. If you declare that your "A" team is playing (or don't put a letter on it) then that's your team. He's dealing with hundreds of teams, so it's just not practical for him to comb through rosters. Even if he did, how exactly would he quantify the relative contribution of each player given the competition, the particular set... ? The mind boggles.

On the other hand since, by now, most everyone knows that Fred's rankings don't always reflect the true strength of teams that don't always play their entire roster, or simply don't play as much as other teams, I am certainly in favor of coaches' and players' polls. If these polls are conducted on a regular basis, like Fred's, then the QB world will have a more nuanced, rounded picture of the actual playing strength of teams.
I feel this. Mad respect to Fred and the time he's spent on this, but I think player polls would be an awesome addition to the rankings.

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 2:35 pm
by AKKOLADE
tmcavoy wrote:Can you do an update of small schools ranking now that SSNCT is over?
SSNCT will be incorporated into the pre-nationals rankings, but I feel no need to do a ranking for a tournament that happened.

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 2:38 pm
by tmcavoy
That's what I meant, just to include it in the next ranking.
Thanks, you've done a great job with these rankings.
I wish I had the time to dig through the stats like that.

Re: HSQBRank 2016-2017

Posted: Fri May 12, 2017 3:29 pm
by AKKOLADE
Father Comstock wrote:
Antrobus63 wrote:I don't believe that Fred can or should have to account for who's playing at a given tournament. If you declare that your "A" team is playing (or don't put a letter on it) then that's your team. He's dealing with hundreds of teams, so it's just not practical for him to comb through rosters. Even if he did, how exactly would he quantify the relative contribution of each player given the competition, the particular set... ? The mind boggles.

On the other hand since, by now, most everyone knows that Fred's rankings don't always reflect the true strength of teams that don't always play their entire roster, or simply don't play as much as other teams, I am certainly in favor of coaches' and players' polls. If these polls are conducted on a regular basis, like Fred's, then the QB world will have a more nuanced, rounded picture of the actual playing strength of teams.
I feel this. Mad respect to Fred and the time he's spent on this, but I think player polls would be an awesome addition to the rankings.
I do the All-World Teams each year. I also consider teams with two full-strength performances to have enough data to be sufficient to rank a team very confidently. If a team has an aberrant performance, I do look into if that was its normal roster.

Re: HSQBRank 2016-2017

Posted: Thu May 18, 2017 10:00 pm
by AKKOLADE
I have calculated all but 5-6 adjustments, 2 because I've done enough for the day and 4 because the sets are being used this weekend and I'm trying to incorporate as much data as possible.

Re: HSQBRank 2016-2017

Posted: Sat May 20, 2017 7:21 am
by luke1865
Just a note regarding SCT Div II adjustments. Our ppb at Texas Invitational is not listed correctly. I have written to them to change it, but haven't gotten a response. It was actually more like 18.2 to 18.3 (they listed the halftime score instead of the final score for one of our games but had the correct tossup info, so we were listed as having a pb of 4 for that match). I don't know if this affects our possible ranking or just your overall stat adjustment, but thought it might be useful?

Re: HSQBRank 2016-2017

Posted: Sat May 20, 2017 1:49 pm
by AKKOLADE
luke1865 wrote:Just a note regarding SCT Div II adjustments. Our ppb at Texas Invitational is not listed correctly. I have written to them to change it, but haven't gotten a response. It was actually more like 18.2 to 18.3 (they listed the halftime score instead of the final score for one of our games but had the correct tossup info, so we were listed as having a pb of 4 for that match). I don't know if this affects our possible ranking or just your overall stat adjustment, but thought it might be useful?
Appreciate it. In general, I don't accept student reported stat errors for the obvious reasons. If someone from TI emails me, I'll make the necessary change. Specific to you, it won't matter.

Re: HSQBRank 2016-2017

Posted: Sat May 20, 2017 11:46 pm
by AKKOLADE
250 teams will be ranked this time.

Re: HSQBRank 2016-2017

Posted: Mon May 22, 2017 2:14 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Mon May 22, 2017 6:16 pm
by nitzuga
Thank you for another great set of rankings! Can you share the PPB adjustments with us? Also, out of curiosity, how many school/ teams do you consider in total every time you make these?

Re: HSQBRank 2016-2017

Posted: Tue May 23, 2017 2:33 pm
by Important Bird Area
luke1865 wrote:Just a note regarding SCT Div II adjustments. Our ppb at Texas Invitational is not listed correctly. I have written to them to change it, but haven't gotten a response. It was actually more like 18.2 to 18.3 (they listed the halftime score instead of the final score for one of our games but had the correct tossup info, so we were listed as having a pb of 4 for that match). I don't know if this affects our possible ranking or just your overall stat adjustment, but thought it might be useful?
This has been fixed.

Re: HSQBRank 2016-2017

Posted: Wed May 24, 2017 10:12 am
by Scottietodd
I also want to chime in and add my thanks to Fred for these great rankings. It is a great to have access to this data and a lot of fun to analyze. Thanks for continuing to do this! Eighty-eight hours! Wow!

Re: HSQBRank 2016-2017

Posted: Wed May 24, 2017 10:25 am
by the return of AHAN
It's also an outlet for community pride when a team can point to a 'national' ranking of teams and show their principal, teachers, family, etc. where they rank nationally. Because when people step back and consider that their team is ranked, say, 114th in the entire nation, they realize that's pretty impressive when you consider the thousands of high schools that exist in the US.

Re: HSQBRank 2016-2017

Posted: Wed May 24, 2017 3:44 pm
by Father Comstock
the return of AHAN wrote:It's also an outlet for community pride when a team can point to a 'national' ranking of teams and show their principal, teachers, family, etc. where they rank nationally. Because when people step back and consider that their team is ranked, say, 114th in the entire nation, they realize that's pretty impressive when you consider the thousands of high schools that exist in the US.
It's difficult when you've won NAC to tell your faculty that you're 91st in the nation.. even though 91st is a clearly more legit ranking than "1st".

Re: HSQBRank 2016-2017

Posted: Thu May 25, 2017 10:03 am
by AKKOLADE
Scottietodd wrote:I also want to chime in and add my thanks to Fred for these great rankings. It is a great to have access to this data and a lot of fun to analyze. Thanks for continuing to do this! Eighty-eight hours! Wow!
88 & counting.

Re: HSQBRank 2016-2017

Posted: Thu May 25, 2017 12:17 pm
by AKKOLADE
Here are this year's All-World Honorable Mention recipients: https://hsqbrank.com/2017/05/25/2017-hs ... le-mention

Re: HSQBRank 2016-2017

Posted: Fri May 26, 2017 4:07 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Sat May 27, 2017 3:56 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Sat May 27, 2017 8:53 pm
by Stained Diviner
He had a pretty good day.

Re: HSQBRank 2016-2017

Posted: Tue May 30, 2017 12:32 pm
by A Dim-Witted Saboteur
Thanks for Player of the Year and also producing an excellent ranking system that's motivated myself and my team, gave us a sense of perspective as to where we stood in a national context, and given us something somewhat concrete that made sense to outsiders to show for our efforts.

Re: HSQBRank 2016-2017

Posted: Tue May 30, 2017 1:08 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Tue May 30, 2017 1:13 pm
by AKKOLADE
Sit Room Guy wrote:Thanks for Player of the Year and also producing an excellent ranking system that's motivated myself and my team, gave us a sense of perspective as to where we stood in a national context, and given us something somewhat concrete that made sense to outsiders to show for our efforts.
Thanks! I'm glad to be of service.

This concludes this year's HSQBRank work. My final tally is 98 hours of work, which is slightly less than expected but I also did fewer regular season updates than expected.

Re: HSQBRank 2016-2017

Posted: Wed May 31, 2017 12:30 pm
by etchdulac
Just to give Jakob's achievement at HSNCT some context, I dug through the stats of all Top 4-finishing HSNCT teams since 2006 (when playoff individual stats became available). In terms of percentage of points scored by one player for a team finishing that high, there's only one comparable performance in those 12 years.

2012 Adam Silverman, Centennial (GA): 2250 of his team's 2550 points (88.24%)
2017 Jakob Myers, Naperville North (IL): 2495 of his team's 2880 points (86.63%)

The next on the list would be Morgan Venkus, who played solo for 10 of his 19 games at HSNCT 2013, and scored 1990 of Loyola Academy's 2660 points for 74.81%.

Re: HSQBRank 2016-2017

Posted: Thu Jun 29, 2017 3:01 pm
by ryanrosenberg
I'm helping Fred out with some HSQBRank things this summer, which includes a new design and home page.

If you have particular things you'd like me to analyze or visualize with the HSQBRank data, feel free to post here or shoot me a PM.

Re: HSQBRank 2016-2017

Posted: Thu Jun 29, 2017 3:02 pm
by AKKOLADE
For the record, the NSC adjustment is not finalized (and is actually just the HSNCT adjustment for now).

Re: HSQBRank 2016-2017

Posted: Thu Jun 29, 2017 5:39 pm
by Chromica
AKKOLADE wrote:For the record, the NSC adjustment is not finalized (and is actually just the HSNCT adjustment for now).
What was the adjustment for this year's SSNCT?

Re: HSQBRank 2016-2017

Posted: Thu Jun 29, 2017 7:56 pm
by AKKOLADE

Re: HSQBRank 2016-2017

Posted: Mon Jul 17, 2017 8:28 am
by troyharris
Fred,
Did you happen to analyze the adjustment on the NASAT for 2017?