Comparison of Ranks Prior to 2020 State Regional Round

This page shows the rankings from various sources of the state qualified teams prior to the regional round of the state tournaments.

WRPI is the official WIAA RPI used for seeding. TBRPI is an RPI calculated by me. Different from WRPI in that it doesn't use OOWP. There are also differences in what games are in scope and the games in the database (a few, not many). WL is just won-lost percentage. CapWL is captured wins and losses. LMpts is the teambrunnhilde.com points ranking: three time defending ranking champion. LMwl is similar to LMpts, except that the dependent variable for the linear model is 1 rather than the margin. So it just looks at wins and losses with the linear model rather than margin of the games. LMwl uses the same X matrix as LMpts. It is not an optimal technique for binary data, but yields a reasonable result. MP is the MaxPreps ranking taken from the MaxPreps site February 25. DP is a seeding based on district tournament placings. District champions seeded highest, with sorting based on the number of teams within the district(s) comprising the tournament. DP is an roughly similar to how the old WIAA draw worked, without the random veneer of drawing names from a hat that simulated fairness. The top eight correspond to the old WIAA #1 seeds (true #1 and whatever the other #1's were called). Inclusion in this comparison may remind you why you hated the old WIAA draw. Comp is a composite of the above ratings: an average of the other rankings for a team and then list those 1-16. All other rankings are the rank over the entire set of teams within a classification.

LMpts is based on margin. LMwl, RPI, WL, and CapWL are based on wins and losses. They are for the entire season. There is no "what have you done for me lately" component of the ranking. The MaxPreps rating uses wins and losses with greater weighting for playoff games. So there is some form of recent performance adjustment. It is unknown if scores (other than win/loss) affect the rating.

Rankings (except for the composite and district placement) are those for the team within the entire set of teams for a classification. They are not 1 to 16 for the 16 state qualifiers. This results in some holes in the rankings. These may be due to upsets at district tournaments, but more often reflect the uneven spread of good teams between the districts. It may also reflect district bracketing that pitted good teams against eachother.

Keen observers will note that the WRPI rank does not completely conform to the tournament seedings. WIAA alters seeds 13-16 to avoid loser out regional matchups between teams in the same league.

A note on King's boys. LMpts has them #1 and LMwl also has them ranked highly. King's lost all four games at a holiday tournament in Arizona, two to just mediocre teams. For the linear model rankings, games involving teams that only appear in one game in the dataset are excluded, so these Arizona games were not considered by the algorithm. This makes the LMpts and LMwl rankings for King's suspect.

For purposes of considering a #1 ranking unanimous, if a team won its district tournament, I consider DP to be a 1, ranking the district champions based arbitrarily on the size of the district. So there are five unanimous #1's: Woodinville (4A girls), Liberty (Spangle) (2B girls), Pomeroy (1B girls), Union (4A boys), and Odessa (1B boys). Last year there were two unanimous #1 picks. One did become champion, the other did not.

Last season LMpts was best, predicting under 75% of tournament games correctly. MaxPreps was second. Several rankings bunched between third and sixth, including WIAA RPI. The results for this year will be shown at the bottom of this page following the tournament.

Tables are sortable, just click on the column header to sort. The native sorting is by seed.

Girls 4A Ranks

Girls 3A Ranks

Girls 2A Ranks

Girls 1A Ranks

Girls 2B Ranks

Girls 1B Ranks

Boys 4A Ranks

Boys 3A Ranks

Boys 2A Ranks

Boys 1A Ranks

Boys 2B Ranks

Boys 1B Ranks

Home