Tuesday, November 29, 2016

NXN Projections - a compilation

I'm not going to do my usual in-depth look at projecting NXN, because as I said in my other post I already have a rough idea of the groupings and I am not going to take the time to refine it as much as usual this year. However, what I will do is post a quick comparison of my (VERY ROUGH) projections along with some other notable projections: Bill Meylan's over at Tullyrunners, Paul Schmitz/Adam Kopet at DyeStat (Boys/Girls), and whoever is doing the rankings these days at Milesplit.

I'll also include a little math to show the general trends/consensus, or how varying the different projections have each team finishing.

My apologies in advance for for any poor formatting. I will put in bold the teams that we all seemed to roughly agree on (teams with sample variability of < 6).

tl;dr summary:

My girls projections appear to be a bit more favorable towards the Southwest and South while slightly less optimistic for the Midwest and Southeast; Milesplit seems to be significantly higher on the Southeast and slightly lower on the Northwest than the rest of us; Meylan and DyeStat typically were closer to the median rankings.

My boys projections appear to be a bit more favorable towards the South while less optimistic towards the Northeast/New York and the Southeast; Milesplit seems to be higher on the Heartland, while DyeStat seems to be a slightly lower on the Southwest. Meylan was all over the place compared to the rest of us (probably because he looked deeper into what the teams have done the rest of the year) with regards to California.

Full breakdown:

Boys =

(Variability) | Median | Average | Team | Watchout | Meylan | DyeStat | Milesplit

(0.25)  |  1.0 |  1.25 | American Fork UT         = 1 | 2 | 1 | 1
(1.58)  |  2.0 |  2.25 | Neuqua Valley IL         = 4 | 1 | 2 | 2
(1.58)  |  4.0 |  3.75 | Bozeman MT               = 2 | 5 | 4 | 4
(4.67)  |  4.5 |  5.00 | Brentwood TN             = 5 | 4 | 3 | 8
(2.92)  |  5.5 |  5.25 | Great Oak CA             = 6 | 3 | 5 | 7
(2.92)  |  5.5 |  5.25 | The Woodlands TX         = 3 | 6 | 7 | 5
(36.92) |  6.5 |  8.25 | Christian Brothers NJ    = 17 | 7 | 6 | 3
(2.00)  |  8.5 |  8.00 | Wayzata MN               = 8 | 9 | 9 | 6
(3.58)  |  8.5 |  9.25 | Downers Grove North IL   = 9 | 8 | 8 | 12
(4.00)  | 11.0 | 10.00 | Summit OR                = 7 | 11 | 11 | 11
(19.00) | 12.5 | 13.50 | Davis UT                 = 10 | 15 | 19 | 10
(7.58)  | 13.5 | 13.25 | Dana Hills CA            = 16 | 10 | 12 | 15
(17.67) | 14.0 | 14.50 | Liverpool NY             = 15 | 13 | 10 | 20
(21.67) | 14.5 | 14.50 | St. Xavier KY            = 20 | 16 | 13 | 9
(7.58)  | 14.5 | 14.75 | Mahomet-Seymour IL       = 18 | 12 | 16 | 13
(3.00)  | 15.5 | 15.50 | Springville UT           = 14 | 17 | 17 | 14
(4.92)  | 16.0 | 16.25 | Staples CT               = 19 | 14 | 15 | 17
(13.67) | 16.0 | 16.50 | Cathedral Catholic CA    = 13 | 21 | 14 | 18
(18.00) | 18.0 | 17.00 | Southlake Carroll TX     = 11 | 18 | 18 | 21
(0.92)  | 19.5 | 19.75 | Fayetteville-Manlius NY  = 21 | 19 | 20 | 19
(6.92)  | 20.5 | 19.75 | Edina MN                 = 22 | 20 | 21 | 16
(25.00) | 22.0 | 19.50 | College Park TX          = 12 | 22 | 22 | 22



Notes: My projections are only off of the NXR races (except The Woodlands' missing runner), while Meylan's uses a synthesis of ratings but I believe weighs most heavily on NXR and State, while the others I'm not sure. Milesplit is actual rankings rather than an NXN-only projection (I don't think they post actual NXN projections?).

CBA, Davis, Liverpool, St. Xavier, Cathedral Catholic, Southlake Carroll and College Park seem to be pretty variable depending on which projection you look at.

CBA is the most severe and is because I have them rated significantly worse than others, it seems, while Milesplit is a bit more favorable than the rest.
College Park and St. Xavier are the next most extreme; CP because I have them rated significantly higher, and St. Xavier because of a combination between Milesplit has them much higher and I have them a bit lower.

Davis, Southlake Carroll and Liverpool are also notable differences: Davis is no one seems to agree on them except Milesplit and my projections, which have them higher than the rest. Southlake Carroll looks like it is due to my rating them higher... I guess I valued the NXN South race more than others (or they are penalizing them, as well as College Park, for their poor races at State - which is very understandable). Liverpool is a slightly less extreme version of Davis, where this time Milesplit has them rated lower than everyone else and DyeStat only slightly higher than Meylan and I.

A modest difference on our views of Cathedral Catholic, with DyeStat and I having them around #13/14 while Meylan (21) and Milesplit (18) are a little lower.

Seems to be very good agreement in general beyond those teams, although Dana Hills (two around 10-12 and two at 15/16), Mahomet-Seymour (two around 12/13, two around 16-18), and Edina (three between 20-22, but Milesplit at 16) are slightly different.


General consensus notes:

The projections all pretty much agree on who is included in the top tier of (6-8) teams, with the only questions being whether or not CBA or Wayzata belong with that group. We also pretty much agree that MW#2 (DGN) and NW#2 (Summit) are mid-pack teams. It does get a bit muddy beyond those two points of agreement, though.



Girls =

(Variability) | Median | Average | Team | Watchout | Meylan | DyeStat | Milesplit

(0.00)  |  1.0 |  1.00 | Fayetteville-Manlius NY  = 1 | 1 | 1 | 1
(0.00)  |  2.0 |  2.00 | Great Oak CA             = 2 | 2 | 2 | 2
(0.25)  |  3.0 |  3.25 | Davis CA                 = 4 | 3 | 3 | 3
(8.25)  |  4.5 |  5.75 | Carmel IN                = 10 | 4 | 5 | 4
(1.67)  |  6.5 |  6.50 | Edina MN                 = 6 | 8 | 7 | 5
(14.92) |  7.0 |  7.75 | Bozeman MT               = 8 | 6 | 4 | 13
(3.33)  |  8.0 |  8.00 | Willmar MN               = 7 | 10 | 9 | 6
(8.25)  |  8.5 |  7.25 | Southlake Carroll TX     = 3 | 9 | 8 | 9
(8.67)  |  9.0 |  9.00 | Minooka IL               = 11 | 7 | 6 | 12
(8.25)  | 10.5 |  9.25 | Keller TX                = 5 | 11 | 11 | 10
(20.92) | 12.0 | 11.25 | Desert Vista AZ          = 12 | 5 | 12 | 16
(3.33)  | 13.0 | 13.00 | Naperville North IL      = 15 | 12 | 14 | 11
(3.58)  | 13.5 | 12.75 | Shenendehowa NY          = 14 | 13 | 10 | 14
(5.58)  | 14.0 | 14.75 | Sunset OR                = 13 | 15 | 13 | 18
(28.67) | 14.5 | 14.00 | Blacksburg VA            = 20 | 14 | 15 | 7
(23.33) | 16.5 | 15.00 | Lake Braddock VA         = 19 | 16 | 17 | 8
(18.00) | 16.5 | 15.00 | Mountain Vista CO        = 9 | 18 | 18 | 15
(2.00)  | 16.5 | 17.00 | Palos Verdes CA          = 16 | 17 | 16 | 19
(1.33)  | 18.0 | 18.00 | Summit OR                = 17 | 19 | 19 | 17
(1.58)  | 20.0 | 19.75 | North Hunterdon NJ       = 18 | 21 | 20 | 20
(0.67)  | 21.0 | 21.00 | Broomfield CO            = 21 | 20 | 21 | 22
(0.25)  | 22.0 | 21.75 | Souhegan NH              = 22 | 22 | 22 | 21



Notes: My projections are only off of the NXR races, while Meylan's uses a synthesis of ratings but I believe weighs most heavily on NXR and State, while the others I'm not sure. Milesplit is actual rankings rather than an NXN-only projection (I don't think they post actual NXN projections?).

Blacksburg, Lake Braddock, Desert Vista, and to a lesser extent Mountain Vista and Bozeman are pretty variable. Minooka, Southlake Carroll, Keller, Carmel, and to an extent Sunset are somewhat variable but relatively stable. Lots of agreement beyond that.

Blacksburg is due to vast differences between Milesplit (#7) and myself (20), with both Meylan and DyeStat about halfway between (#14 and 15).

Lake Braddock is due to Milesplit (#8) significantly ahead of the rest of us: myself at #19, though Meylan (16) and DyeStat (17) not much different.

Desert Vista is mostly due to Meylan being notably higher on them (#5), and to be honest if I was looking at more than just their NXR race I would probably agree with him. However, as it is I have them at #12, which is similar to DyeStat (12) and Milesplit (16).

Mountain Vista looks to be down to me being higher on them (#9) than the rest (#15, 18 and 18).

Bozeman got a pretty wide variety, although if the others have a similar tier of teams as I do (Meylan pretty much does: for example, I have Bozeman at roughly 255 points and he has them at 246, so pretty slight difference on them but a bigger difference on the Heartland and especially South teams) then the only "notable" outlier is Milesplit at #13.

The other somewhat notable differences is mainly the two South teams (I have them finishing higher than the rest) and Minooka (Milesplit and I have them around #11/12, while Meylan/DyeStat have them at #6/7). Carmel is also a big difference, though that is just my ranking - I'm assuming the difference is due to whether or not we include Sarah Leinheiser (who has been dealing with an apparent foot injury). I didn't include her in the projections, and Meylan included her as their #5


General consensus notes:

The projections all pretty much agree on who is included in the top tier of (3-4) teams, with the only questions being whether or not Southlake Carroll belong with that group (FWIW, I am the outlier as I have them scoring in the 180s to being neck and neck with Davis). There also seems to be a general agreement on the rest of the field as well: even most of the highly variable teams mentioned above look to be roughly in the same tiers, with the only exceptions beyond the Southeast (milesplit is higher), Southwest (mixed bag) and South (I'm higher) teams. Very good agreement on the lower tier of teams. Much better consensus overall this year than in the boys race.

No comments:

Post a Comment