Rankings - Regular Season

[allhoc11]

I'd say its possible no Lake team will be in the state tournament, or should Bloomington just give up now? As for 6AA, I'd say BSM's chances are worse this year than last. Tonka may not be quite up to last year, but Hopkins and Wayzata are better. Getting through 6AA may require wins against 3 really strong teams.

Huh???

The assumption was that EP or Edina will cruise through section 2AA. I think Bloomington may disagree. Is Chaska/Chanhassen still in 2AA? They are no pushover either.

I'd say Chaska/Chan probably hasn't played the toughest schedule, and didn't look like a team that could compete with an EP or Edina based on their performance against a very good BSM team today. Bloomington, while I haven't seen them play this year, has a goalie that can keep them in games, but when you start comparing scores/shots against like opponents with the top teams in section 2, my guess is they would get badly outshot and need their goalie to play amazing in hopes to keep it close against EP or Edina. In general I'd say there aren't many nights off for teams and the girls game has become closer to any team could win on any night league, however if I was the betting type, my jelly beans would be a Edina/EP section final.

[MNHockeyFan]

Latest LPH rankings.

In Class AA, no change in the Top 3; BSM and Edina trade places at 4 & 5; EP moves back into the Top 10.

In Class A, TRF and Blake remain 1-2; EGF and RW switch places at 3 & 4; SSP moves ahead of Warroad at #5.

December 31, 2013

Class AA (First-place votes)
1 MINNETONKA ( 8 )
2 Wayzata ( 2 )
3 Hopkins
4 Benilde-St. Margaret's
5 Edina
6 Hill-Murray
7 Lakeville North
8 Blaine
9 Eden Prairie
10 Irondale
11 Lakeville South
12 Dodge County
13 Roseville Area
14 Stillwater Area
15 Mounds View
16 Buffalo
17 Jefferson/Kennedy
18 Burnsville
19 Andover
20 Eastview

Receiving votes: Armstrong/Cooper, Rochester Century, Maple Grove, Cretin-Derham Hall, Elk River/Zimmerman, Centennial, Chaska/Chanhassen, North Wright County


Class A (First-place votes)
1 THIEF RIVER FALLS ( 8 )
2 Blake ( 2 )
3 East Grand Forks
4 Red Wing
5 South St. Paul
6 Warroad
7 Achiever Academy
8 Orono
9 Totino-Grace
10 Hibbing/Chisholm
11 Mound Westonka
12 Mankato East/Loyola
13 Breck
14 Proctor/Hermantown
15 Alexandria
16 St. Paul United
17 New Prague
18 Chisago Lakes Area
19 Northfield
20 New Ulm

Receiving votes: Princeton, Crookston, Mahtomedi

[AAA Dad]

High School Girls AA


1 HOPKINS

2 Minnetonka
3 Benilde-St. Margaret's
4 Wayzata
5 Hill-Murray
6 Edina
7 Lakeville North
8 Eden Prairie
9 Blaine
10 Irondale
11 Lakeville South
12 Dodge County
13 Mounds View
14 Roseville Area
15 Stillwater Area
16 Buffalo
17 Jefferson/Kennedy
18 Eastview
19 Andover

20 Burnsville







High School Girls A


1 THIEF RIVER FALLS

2 Blake
3 East Grand Forks
4 Red Wing
5 Achiever Academy
6 South St. Paul
7 Warroad
8 Orono
9 Mound Westonka
10 Totino-Grace
11 Hibbing/Chisholm
12 Mankato East/Loyola
13 Breck
14 Proctor/Hermantown
15 Alexandria
16 St. Paul United
17 New Prague
18 Chisago Lakes Area
19 New Ulm

20 Northfield

[MNHockeyFan]

January 15, 2014

Class AA
1 HOPKINS
2 Minnetonka
3 Benilde-St. Margaret's
4 Wayzata
5 Edina
6 Lakeville North
7 Hill-Murray
8 Eden Prairie
9 Blaine
10 Irondale
11 Dodge County
12 Lakeville South
13 Stillwater Area
14 Mounds View
15 Jefferson/Kennedy
16 Roseville Area
17 Burnsville
18 Cretin-Derham Hall
19 Buffalo
20 Andover

Class A
1 THIEF RIVER FALLS
2 Blake
3 East Grand Forks
4 Red Wing
5 Achiever Academy
6 Orono
7 Warroad
8 South St. Paul
9 Mound Westonka
10 Hibbing/Chisholm
11 Totino-Grace
12 Mankato East/Loyola
13 Breck
14 St. Paul United
15 Proctor/Hermantown
16 Alexandria
17 Chisago Lakes Area
18 New Prague
19 New Ulm
20 Princeton

[MNHockeyFan]

January 22, 2014

In the Top 10, Tonka, BSM, Edina, Hill-Murray, Eden Prairie and Blaine all move up a notch or two; Hopkins, Wayzata and Lakeville North go in the other direction.

Class AA
1 MINNETONKA
2 Benilde-St. Margaret's
3 Edina
4 Hopkins
5 Hill-Murray
6 Eden Prairie
7 Wayzata
8 Blaine
9 Lakeville North
10 Irondale
11 Dodge County
12 Stillwater Area
13 Mounds View
14 Burnsville
15 Lakeville South
16 Bloomington Jefferson
17 Cretin-Derham Hall
18 Roseville Area
19 Buffalo
20 Maple Grove


In Class A, TRF holds at No. 1, with EGF, Red Wing and AA each moving up one spot. Mankato East/Loyola replaces Hibbing/Chisholm at No. 10, while Blake takes the biggest hit, falling from No. 2 to No. 5.

Class A
1 THIEF RIVER FALLS
2 East Grand Forks
3 Red Wing
4 Achiever Academy
5 Blake
6 Orono
7 Warroad
8 South St. Paul
9 Mound Westonka
10 Mankato East/Loyola
11 Breck
12 St. Paul United
13 Proctor/Hermantown
14 Hibbing/Chisholm
15 Totino-Grace
16 Alexandria
17 Chisago Lakes Area
18 New Prague
19 New Ulm
20 Princeton

[MNHockeyFan]

Latest LPH rankings.

January 29, 2014

Hopkins jumps from 4th to take over No. 1 and Eden Prairie goes from No. 6 to No. 2. Meanwhile, Tonka falls out of the No.1 spot to No. 3, and BSM falls from No. 2 to No. 5.

Class AA
1 HOPKINS
2 Eden Prairie
3 Minnetonka
4 Hill-Murray
5 Benilde-St. Margaret's
6 Wayzata
7 Edina
8 Blaine
9 Lakeville North
10 Irondale
11 Dodge County
12 Stillwater Area
13 Mounds View
14 Burnsville
15 Lakeville South
16 Bloomington Jefferson
17 Cretin-Derham Hall
18 Maple Grove
19 Roseville Area
20 Buffalo


In Class A, no movement whatsoever in the Top 10.

Class A
1 THIEF RIVER FALLS
2 East Grand Forks
3 Red Wing
4 Achiever Academy
5 Blake
6 Orono
7 Warroad
8 South St. Paul
9 Mound Westonka
10 Mankato East/Loyola
11 Breck
12 St. Paul United
13 Proctor/Hermantown
14 Hibbing/Chisholm
15 Totino-Grace
16 Alexandria
17 New Prague
18 Chisago Lakes Area
19 New Ulm
20 Princeton

[OsMetroDad]

What are the plus/minus values (or standard deviation) assigned to the "ranking values" listed in LPH? Are the rankings based on a human poll with inherent bias toward each pollers team? From each of the 3 computer ranking systems that I have experience with (QRF, LSQRANK and KRACH) the effective standard deviation associated with say the top six team is large enough that if one performs a test between means, i.e., Is one team really better than another, or is there insufficient separation between them, the data suggests that the top six teams or put it anoher way that most any sequential group of 5 or 6 teams, the teams within the group are more or less interchangable. For example, the LSQRANK results for the cumulative season up to and including 27-JAN-2014 games yeilds (top 25):

Code: Select all

LSQRANK Team SEC Score GD SOS PPc PKc ADJPTS (W - L-T) percent
-----------------------------------------------------------------------------------------------------
1 EdenPrairie 2 AA EP 0.16 0.0 0.88 25.5 85.7 1.2 (15- 6-0) 0.714
2 HillMurray 4 AA HM 25.24 0.3 0.78 26.7 96.3 1.3 (19- 3-0) 0.864
3 BSM 6 AA BSM 34.48 0.3 0.84 45.8 71.4 1.5 (18- 2-1) 0.881
4 Hopkins 6 AA HOP 43.20 0.4 0.86 28.9 85.9 1.5 (21- 3-0) 0.875
5 Edina 2 AA EDN 44.96 0.4 0.89 30.9 83.6 1.2 (15- 6-2) 0.696
6 Minnetonka 6 AA MTK 66.95 0.7 0.90 33.3 80.8 1.5 (16- 2-2) 0.850
7 Wayzata 6 AA WYZ 75.19 0.8 0.86 42.6 83.6 1.2 (16- 6-1) 0.717
8 DodgeCounty 1 AA DC 108.18 1.1 0.75 28.4 88.5 1.1 (19- 5-0) 0.792
9 Achiever 4 A ACH 128.87 1.3 0.78 23.9 87.8 1.2 (18- 4-1) 0.804
10 LakevilleS 1 AA LVS 149.30 1.5 0.84 29.1 87.9 1.1 (16- 7-0) 0.696
11 Burnsville 3 AA BNV 159.27 1.6 0.83 27.2 85.7 1.0 (14- 8-1) 0.630
12 MapleGrove 6 AA MG 159.93 1.6 0.82 27.5 84.3 1.1 (14- 5-2) 0.714
13 LakevilleN 1 AA LVN 160.65 1.6 0.85 15.4 90.9 1.2 (15- 5-2) 0.727
14 Stillwater 4 AA STW 161.83 1.6 0.83 24.1 71.4 1.1 (15- 6-1) 0.705
15 Blaine 5 AA BLA 175.06 1.8 0.80 22.1 88.6 1.3 (18- 3-2) 0.826
16 EastGrandForks 8 A EGF 175.28 1.8 0.72 27.5 93.3 1.3 (21- 1-1) 0.935
17 Irondale 5 AA IRO 185.49 1.9 0.77 41.2 73.3 1.2 (18- 5-0) 0.783
18 ThiefRiverFalls 8 A TRF 191.47 1.9 0.74 22.9 95.1 1.4 (21- 1-0) 0.955
19 MoundsView 5 AA MSV 223.77 2.2 0.83 40.8 81.5 0.9 (12- 8-3) 0.587
20 Andover 7 AA AND 232.13 2.3 0.81 34.0 83.3 0.9 (12- 7-2) 0.619
21 EastRidge 3 AA ERG 237.65 2.4 0.76 25.8 82.7 1.0 (15- 7-1) 0.674
22 Eagan 3 AA EAG 238.27 2.4 0.85 29.4 73.1 0.7 ( 9-12-1) 0.432
23 Warroad 8 A WAR 242.05 2.4 0.78 16.0 78.3 0.8 (12- 9-1) 0.568
24 Eastview 3 AA EVW 242.84 2.4 0.82 16.7 67.9 1.0 (14- 8-1) 0.630
25 CDH 4 AA CDH 245.82 2.5 0.82 41.2 53.1 1.0 (15- 8-0) 0.652

The plus/minus standard deviation on the GD ("Goal Differential") is pretty close to 1.0. For example, given that Eden Prairie is ranked at #1 having a GD value of 0.0 +/- 1.0 vs. #4 Hopkins at 0.4 +/- 1.0. The question in any of these rankings is "Is the perceived #1 any different than the percieved #4 or #3, etc?" Again, based on sound numerical (unbiased) data the answer is NO if one performs a simple statistical test between means. So my point is this, if you are going to report the solution to a ranking, please also report the other half of the data which is equally important, that qualifies the meaning of these ranking values. BTW, I am thrilled that Hopkins in your human poll in ranked at #1 since several of it players participates in our Spring/Summer/Fall program, but as a nerdy engineer/statistcian who has "crunched the numbers" I know that in reality it is really somewhat of a statistically tie amoung the top 5 or 6 teams no matter what type of human expert justifications are provided to ligitamize the human poll, and the feel-good emotionals that go along with it. So if you are going to provide the poll results, then go that next step and also provide the crucial statistic or measure that numerically qualifiies the result. Again, to give you a real life example, it is the same as the surveyor providing the aerial mapping company with a list of 3D ground control point coordinates (XYZ's) to produce a map, but are the XYZ's good to plus or minus 0.1 Feet or are they only good to +/- 50 Feet? Its makes a difference in how one interprets the data, and likewise with this polls you have a similar situation because like it or not, they are used in section seeding meetings. I hope that the points I bring up can help improve the process of polling, and help quantify the results.

[MNHockeyFan]

What are the plus/minus values (or standard deviation) assigned to the "ranking values" listed in LPH? Are the rankings based on a human poll with inherent bias toward each pollers team?

The LPH rankings have always been a human poll and as such - yes, there could be some bias that creeps in. However it is my understanding that the "pollers" are made up of independent observers and media types - as opposed to coaches who would be more likely to favor his or her own team. So bias SHOULD be kept to a minimum. And from my experience, having followed the game and kept an eye on these polls for a number of years, they seem to do a pretty good job, especially considering the fact that those doing the voting can't possibly get out to actually see all of the contending teams on a regular basis.

[allhoc11]

What are the plus/minus values (or standard deviation) assigned to the "ranking values" listed in LPH? Are the rankings based on a human poll with inherent bias toward each pollers team?

The LPH rankings have always been a human poll and as such - yes, there could be some bias that creeps in. However it is my understanding that the "pollers" are made up of independent observers and media types - as opposed to coaches who would be more likely to favor his or her own team. So bias SHOULD be kept to a minimum. And from my experience, having followed the game and kept an eye on these polls for a number of years, they seem to do a pretty good job, especially considering the fact that those doing the voting can't possibly get out to actually see all of the contending teams on a regular basis.

One correction coaches are included in the voting, I know some coaches who get an email weekly to submit rankings. Not sure how many do, and I do think it's a mix of coaches and media types.

[MNHockeyFan]

One correction coaches are included in the voting, I know some coaches who get an email weekly to submit rankings. Not sure how many do, and I do think it's a mix of coaches and media types.

Thanks for the additional info. When I've seen the results in the Strib paper edition they have occasionally shown the first place votes in parenthesis. As I recall there is typically only about a dozen that show up, so it must be a pretty select group of coaches who are included. Either that or many who could vote choose not to, for whatever reason.

[OsMetroDad]

While snow blowing the drive-way early this AM, I was thinking about how to exact a plus/minus value from the LPH human poll results. Correct me if I'm off base here, but the of collecting human polling data, is that not exactly the same as what product development focus groups do to figure out what features potential customers find importantin a product? Bussinesses like 3M do these things all the time, and extract the not only what the important attributes are (i.e. "rankings" in our case) but also qualify those results with variance value numbers. So the method is there to retrieve the plus/minus values, one just needs to get a hold of the polling data to compute the variance values, and then subsequently report those values along with the poll results.

[Bulldog3489]

Help me out, but is the only computer formula that has AA in first place in class A the formula from an AA parent?

[OsMetroDad]

You mean class AA as in "double A"?

[MNHockeyFan]

You mean class AA as in "double A"?

Class AA is commonly referred to as "2A"

Class A is commonly referred to as "1A"

[EPIC97]

What Bulldog is trying to say is "Help me out here, but is the only formula that has Achiever Academy in first place in class A the formula from an Achiever Acadaemy parent"

Just so we are all on the same page.
Single A are smaller schools, Double A are big schools, Triple A is summer hockey and the rest of the country in the winter also known as AAA.
1A, 1AA also known as 1 double A or One Double A, 2A and 2AA also known as 2 double A or two double A are sections and they continue all the way up to 8A and 8AA…
AA is known as Alcoholics Anonaymous like, (Honey I am off to my AA meeting, don't wait up for me.) AA is also Achiever Academy, AAA is Achiever Academy Aces. (Do you think this was intended or just a coincidence).
So just to be clear. AA or AAA play Single A in 4A.
The beauty of the Achiever name is its flexibility and can expand as they grow.
I propose we call them Achiever or A while they are in Single A, as they move up we can refer to them as Achiever Academy or AA when they play in Double A and once they leave MSHSL to play with Shattuck we can call them Achiever Academy Aces or AAA when they will be playing AAA.
Lastly, just to come full circle here.
Eh!!! is like a period at the end of a sentence in Canada Eh!

[Hard water fan]

Very good post Epic! Good way to start a Friday. Here's some music to go with your post...
http://www.youtube.com/watch?v=iEEyWDL2vLg

[OsMetroDad]

oops sorry I mis-read your post in my hast to scan your content. The answer to your original question is no, like any computerized rating program, LSQRANK technolgy and similarily I suppose what ever method QRF & KRACH use do not favor or bias any specific team in the data set. The data is processed and the post-adjustment statistics and residuals are reveiwed. BTW, in the last processed results Eden Prairie was ranked at the top, along with the group of Hill Murray & BSM, Hopkins, etc, not Achiever. Achiever moved up very modestly to I think to 9th from 13th or 12th where they were trending due to a couple recent data observation additions: (1) A close overtime win by Achiever over highly rank Lakeville North; (2) A close overtime loss to highly ranked Dodge County; (3) a TRFs loss to a better than average SSR team; and (4) a EGFs tie to an average Bemidji team according to their records and SOS.

These computerized ranking software should in principle solve for the best-fit of the bulk of the data over the entire data set (i.e., "entire season"), although it is possible to uniformly adjust the weighting system so that perhaps the games towards the end of the season are weighted more heavily than than the games played during early season.

From last nights games, Hopkins needs to get its first line back together, ouch 3-2 loss to MapleGrove; Centenial beat Minnetonks by 1 which should move them up a bit; BSM should flip-flop with EP with a 4-2 win; Hill Murray solidifies their high rank with a dominating 3-0 win (shots 42-14) over a very good South Saint Paul team; Achiever had a solid 7-1 win over SPU.

I'll add the new data from WED and THUR and post the new result for you. Enjoy

[MNHockeyFan]

Latest LPH rankings.

February 5, 2014

In the Top 10 in Class 2A, Hopkins stays at No. 1, while BSM is the big mover, jumping from No. 5 to No. 2. Meanwhile, no changes in the Top 9 in Class 1A.

Class AA
1 HOPKINS
2 Benilde-St. Margaret's
3 Eden Prairie
4 Minnetonka
5 Hill-Murray
6 Wayzata
7 Edina
8 Blaine
9 Lakeville North
10 Irondale
11 Dodge County
12 Burnsville
13 Stillwater Area
14 Mounds View
15 Maple Grove
16 Lakeville South
17 Bloomington Jefferson
18 Cretin-Derham Hall
19 Roseville Area
20 Buffalo


Class A
1 THIEF RIVER FALLS
2 East Grand Forks
3 Red Wing
4 Achiever Academy
5 Blake
6 Orono
7 Warroad
8 South St. Paul
9 Mound Westonka
10 Breck
11 Mankato East/Loyola
12 St. Paul United
13 Proctor/Hermantown
14 Totino-Grace
15 Hibbing/Chisholm
16 Alexandria
17 New Prague
18 Chisago Lakes Area
19 New Ulm
20 Moose Lake Area

[MNHockeyFan]

Latest LPH rankings.

February 12, 2014

(The poll obviously taken before tonight's 6AA games)

Class AA
1 HOPKINS
2 Benilde-St. Margaret's
3 Eden Prairie
4 Minnetonka
5 Hill-Murray
6 Wayzata
7 Edina
8 Blaine
9 Lakeville North
10 Dodge County
11 Burnsville
12 Irondale
13 Stillwater Area
14 Mounds View
15 Maple Grove
16 Lakeville South
17 Cretin-Derham Hall
18 Roseville Area
19 Buffalo
20 Bloomington Jefferson


Class A
1 THIEF RIVER FALLS
2 East Grand Forks
3 Red Wing
4 Achiever Academy
5 Blake
6 Warroad
7 South St. Paul
8 Orono
9 Breck
10 Mound Westonka
11 St. Paul United
12 Mankato East/Loyola
13 Proctor/Hermantown
14 Totino-Grace
15 Hibbing/Chisholm
16 Alexandria
17 New Prague
18 New Ulm
19 Princeton
20 Moose Lake Area

[OsMetroDad]

Section play-offs are almost at their conclusion, and for the most part the higher ranked (computer ranking or otherwise) team has prevailed, except in two general cases: (1) Where the two team were a statistical tie going into the game; and (2) the upset where a clearly lower ranked team manages to pull out a win, due to take your pick .... Examples of both case are: (a) Princeton win over Orono (the upset); and (b) Wayzata win over Hopkins (the statistical ranking tie). In case example (b) Wayzata numerically gets pushed again of Hopkins simply due to another close win, but in case(a) that doesn't happen, and is no where close to making the numerical flip-flop. Another subset of case (a) was tues nights game South St Paul (SSP) versus Saint Paul United (SPU). The game went into overtime, and it was not looking good for SPU, SSP pretty much had the momentum and appeared to be the better team based on the time it spent in the offensive zone , but to SSP chagrin, SPU managed to finally get the puck out of their zone and down the ice where they sneaked in the winning goal. SSP was significantly ranked ahead of SPU before the game, and afterwards closed the gap a lot by weighting this particular game more heavily, so now the two are more or less a statistical tie, with SSP still slightly ahead with regard to raw rank. Numerically what keep SPU slightly behind SSP is the bulk of season's results. So what is to be done? Are the late season play-off games deserve the higher weight?

[Nevertoomuchhockey]

Even though SPU also beat SSP 3-1 in regulation only 8 days before??

[OsMetroDad]

Yup, The observation equation used takes in account more than just winning the game, it uses shot differential, penality kills successes/attempts, accounts for major penalities, and open net goals. I have not implemented MOV corrections for short-hand goals yet. Anyway, getting back to the 3-1 win, I believe from the HHUb stats that it had an open net goal late in the 3rd period, plus SSP had out shot SPU and the PKs differential had some effect. So basically the composite MOV on that particular game was a net 0.0 +/- 0.2, which numerically is a wash for both teams.

[Nevertoomuchhockey]

God I hate math. I appreciate the additional explanation. Maybe by the time my daughter gets to high school I will comprehend this.

[MNHockeyFan]

Final LPH Rankings

February 26, 2014

Class AA
1 HILL-MURRAY
2 Benilde-St. Margaret's
3 Eden Prairie
4 Hopkins
5 Minnetonka
6 Lakeville North
7 Wayzata
8 Edina
9 Burnsville
10 Dodge County
11 Mounds View
12 Blaine
13 Irondale
14 Stillwater Area
15 Andover
16 Maple Grove
17 Lakeville South
18 Cretin-Derham Hall
19 Roseville Area
20 Roseau


Class A
1 BLAKE
2 East Grand Forks
3 Red Wing
4 Thief River Falls
5 Warroad
6 South St. Paul
7 Orono
8 Breck
9 Mound Westonka
10 St. Paul United
11 Proctor/Hermantown
12 Totino-Grace
13 Hibbing/Chisholm
14 Alexandria
15 Mankato East/Loyola
16 New Prague
17 New Ulm
18 Princeton
19 Moose Lake Area
20 Chisago Lakes Area

[Nimrod]

Interesting final rankings by LPH. I get why teams like Hopkins, Tonka and Wayzata who didn't get to state are ranked high. If they were in most other sections they go to State and have as good a chance as anyone at winning it. And no offense to EP but how does LN end up behind them (and not just 1 or 2 spots) when they spanked EP in the third place game at State?