2019 Ranking Comparison

The results are in!  No I’m not talking Eastlake, O’Dea, or Washougal.  It is the 2019 ranking comparison.  Once again the Team Brunnhilde Linear Model Points rating was the champion.  Going 40-8 on Thursday had me thinking I could hit 75%, which is a gold standard for picking tournament winners.  Going into the twelve championship games a 9-3 record would suffice, but losing all three 7PM games scuttled that.

In summary, the final standings are: TeamBrunnhilde LM Points 205, MaxPreps 199, TeamBrunnhilde RPI 193, Composite at 190 (but a higher percentage due to unpicked games), WIAA RPI and LM won-loss tied at 191, Evans at 188, Captured WL at 187 and Won-Loss at 173 (with 7 unpicked games).  Unanimous picks hit 81.2% correct.

The details are here.

WIAA RPI is the ranking of interest since that is how the tournaments are seeded.  I don’t suggest using the TeamBrunnhilde LM Points or MaxPreps ratings.  For different reasons neither is suitable for seeding the tournament.  However the comparison with TeamBrunnhilde RPI is germane.  TB RPI, albeit with a slightly different database (I think more accurate) and different set of games in-scope, does not use OOWP.  WIAA RPI uses OOWP.  WIAA RPI also puts in fake data for out-of-state teams for OOWP.  Maybe OOWP should just be dropped.  It doesn’t seem to make more accurate rankings.

2019 Rankings Comparison

This year’s version of the TeamBrunnhilde state tournament rankings comparison has eight rankings.  The same seven as last year and a return of a TeamBrunnhilde RPI.  There is again a composite ranking included.  The TeamBrunnhilde linear model for points has always won this comparison, but I sometimes wonder.

This year for the third year in a row Cashmere is my 1A girls #1 team.  The previous years that looked good until Lynden Christian came around.  So I wonder again.  This year Lynden Christian girls lost three games in four tries in the last two weeks of January.  Before that, it was three losses in 87 games.  So maybe it is Cashmere’s year.

In Boys 2B, Odessa is the consensus choice for second place.  Sorry.  Not what they want.

For the 96 regional games, 59 are picked unanimously.  Last year it was 57.  The Girls 3A tournament regionals would be entirely unanimous but for Hudson’s Bay being ranked higher than Roosevelt in Win-Loss.

The full comparison is here.  To be updated as the tournaments progress.

What was that score?

Every day it occurs. Several times. The score of a game is reported differently in various sources. On Thursday, out of 94 games, 12 were reported with a variety of scores.

An example. Thursday’s Ridgefield v Columbia River boys game. WPAN has 67-41 (just a bare score); MaxPreps had 65-39 (with a line score and a play-by-play reporter who ends with a 67-39 score); Vancouver Columbian has 65-41 with a consistent boxscore and a detailed story written by a journalist actually at the game. I’m going with Vancouver Columbian. But they’re unofficial. It’s also two votes for 65 and two votes for 41.

It could be just typos. I do that too. That’s why I try to get multiple sources for a score. I discover a lot of my mistakes that way. But finding multiple scores for the same game occurs too often for just fat fingers. I’m thinking that most of the score differences stem from the score reporters. Sometimes I see a score revised from score A to score B. One team scorebook keeper is clobbering the report by other one. How hard can it be to report the score that was on the scoreboard when the game ended? Maybe the scoreboard failed in the last second and missed a three-pointer to change from 65-42 to 65-45. That actually happened this season. I saw it. But that is a first for me, after ~1500 high school basketball games seen.

HEY! Scorebook keepers! Do a little checking. Check with your opposite keeping book for the other team. Agree on the score to report. And then make sure the quarter scores, if you’re reporting those, add up to the final score. They should.

Let it Snow

It snowed nine inches at my house on February 1, 1990. That was before I was going to high school basketball games. But with the research I can see just how disruptive that was to local basketball tournaments. In 1990, the league Liberty was a part of, the Seamount, had three full weeks from the end of the season until the WIAA ‘draw date’. So tiebreakers, the league tournament, and the district tournament could spread across three weeks. Even as late at 2010, Kingco 3A likewise had three full weeks, 19 scheduleable dates (not counting Sundays) for tiebreakers, league, and district tournaments.  Snow was not a problem.

In 2019, Kingco 3A (which is Liberty’s league but Liberty is not part of for tournaments), cut a full week away from the schedule decades ago. Instead of three weeks, the pre-state playoffs must complete in with 13 scheduleable dates.

In 2018 on the ‘bigweekend’ of district tournaments—Friday and Saturday—there were106 boys games and 111 girls games (a few league games in there). Some on Monday too. With the snow early last week, games were already being pushed toward the ‘big weekend’. Losing Friday, Saturday, and a lot of Monday, the deficit grows. Instead of 217 games only 37 games were played this year on the big weekend. Every district lost some district tournament games, even in districts where the tournament was able to continue. Big districts like Southwest,saw games in Ilwaco, where the weather was probably ‘nice enough’, canceled. Teams, and fans, couldn’t make the trip.

By last Friday, many district tournament had four days of games required to complete their tournament. Unless they can get in games today (Monday) there will be only one unused day left before the WIAA deadline. It isn’t a draw date anymore. No matchups to pull out of a hat in a roomful of anxious coaches. So no real need to complete by Sunday the 17th. The WIAA could relax their grip on the last two weeks of February and let district tournament games slide into the beginning of regional week.

There are more pre-state tournament games nowadays also. In 1990, with four classifications there were fewer district tournaments. Nearly 100 fewer playoff games before state for girls: 343 for 1990, 434 for 2018. So with only two weeks instead of three, the ‘density’ of tournament games is getting close to double what it was in 1990. Also in 1990 there were only two referees per game, now there are three, so the number of referees required is about three times what it was 29 years ago.

2016-2017 was an awful year for postponements. The southern third of the state had week after week of storms. The rest of the state had lots of postponements as well. The Northwest Conference season was curtailed. That happened to the Whitman County League back in 1988-89.

The State of Journalism, 2018

… As it relates to high school basketball.

Last year saw the Bellingham Herald drop most of its local high school sports coverage. Just the big stories are left. That is, the story that can be written without a particular deadline and run whenever available column inches allow. That followed similar exits from day-to-day coverage by the Kitsap Sun, Seattle Times, and others. It looks like the Tri-Cities Herald is going the same route this year.

I don’t understand. For the Seattle Times, the local sports teams ARE the Seahawks, Mariners (unfortunately), Sounders, and Huskies. But for smaller markets, what do people expect to see in their local papers? The local sports! There are zillions of places to check NFL statistics in order to make fantasy sports draft picks. Where do you get to see your kid’s name in agate type? In the Bellingham Herald (well, not anymore), the (Longview) Daily News, the Walla Walla Union Bulletin, the Quad-Cities Herald. Maybe it is cheaper just to pick up syndicated content from elsewhere, about sports elsewhere. But as a reader, I can just go elsewhere and get that content directly. Lots of it. Why should I look in the Abedeen Daily World for the latest stories on the NBA? There’s a daily show on ESPN with just that. If I want to see just how Montesano fared against Elma, well it’s ONLY in the Aberdeen Daily World. Protect your monopoly!

At least the Seattle Times is still offering on-line high school data. How’re they doing? Not well. Today’s (December 7, 2018) database is typical. It’s a really nice looking database front end. Probably looks great on an iPhone X. The content is mostly missing. And what content there is too often wrong. I take the opposite view. My website is clunky. But the html is tight, with high signal to noise ratio, and the data is right: not perfect, but I check and make corrections.

Here’s from the boys basketball part of ScoreBook Live, the Seattle Times’s outsourced data provider. The Seattle Times wishes that somebody live-update the game for them (gee thanks, what do I get paid? Nuttin’). The Lakeside v West Seattle game didn’t get very far: 2-0. But what is the game? Lakeside Girls v West Seattle Boys. Come on, man!

There is a spiffy map to show the location of the game. Here for Nathan Hale at Cleveland.

Is that right? Just go east on I-90. Way east.

Or this map. Really useful, huh?

Game is being played somewhere in the world. Don’t need to ask Elon Musk for a ride.

They’ve even got a choice for games from defunct leagues.

The screen caps are from the boys portion of the database. But that doesn’t mean the girls aren’t participating:

They do know that Edmonds-Woodway boys played Kamiak boys on Friday, although I bet it wasn’t at 11:15 during first lunch. Edmonds-Woodway apparently had a double header according to the Seattle Times. No, that’s the girls game. It wasn’t in the girls section of the database. Here’s another:

Yeah. The girls game. But at least there is content, even if it is wrong.

For girls games on Friday December 7, the Seattle Times has no games for Cascade league(right!); no games for Emerald City (there were four); five games for Kingco 3A/2A with no results–there were four games; six games for Kingco 4A with no results–there were six games, so there were at least the right number of empty spots; seven for Metro with two results, and that boys v girls game ended at 2-0 again; and finally for Wesco three games spread across the two choices with no results,there were actually nine games. (Phew, that sentence was an arduous journey).  So two results for 30 games. Impressive!

I’m sure the Seattle Times sports editors would say that if nobody reports the results it’s not their problem. Besides, they’re all too busy writing the next 1-10 rating of greatest Seahawk what-ever thingy that they can figure up a list for. They fired all the high school sports writers. The old old old (three iterations ago) Seattle Times database is still around, just not linked from the Seattle Times website. But if you want to compare, here it is (until they take it down out of sheer embarrassment at the comparison). Note that the Seattle Times actually covered high school sports from Pierce to Whatcom county.

Go, Go, Sammamish!

Once upon a time, Sammamish girls basketball was good. Regularly stately; four of six years between 1990 and 1995. Then the coach retired or was otherwise replaced (I really don’t know the circumstances).

I recall a brief conversation with Lorinda Flikkema, the new coach, before a Liberty @ Sammamish game in 2001. Teresa Eliott, a 6-2 pretty good senior center but also national-class rower had been dropped from the roster. I noted the absence and asked Lorinda. “Teresa is not on the team.” I persisted a bit until “TERESA is NOT on the TEAM!!” I understand. It is of what we do not speak.

By 2004, Sammamish was playing with six varsity players. A seventh was on the roster but only as an emergency promoted JV in case of dire need. Although only six players, they easily made the league tournament (not a terribly difficult task). For league, at Bellevue Community College, Lorinda dressed NINE players. ‘Good’, I thought. ‘Build for the future even from the crumbs you have.’ Sammamish held close to Newport through the first half, but in the third quarter the bottom fell out. Soon down by 30 points. A Sammamish senior fouled out. In goes girl #6. Lorinda commiserates with the DQ’d player. No, mostly it looked like a girl chat from across the floor–fhe three varsity wannabees ignored at the end of the bench. Somebody yells across the court, ‘Clear your bench!’ No. Lorinda just kept her back on the others. The next day she quit. Good riddance!

It was a hard row to hoe that she left. An absolutely empty roster. A losing streak that reached 110 league games, But in 2015, Sammamish made state and advanced into the quarter-finals. Now in 2018-2019, on December 3,2018; no Kingco 3A/2A girls team has won a game. No (0) wins. Sixteen (16) losses. It is up to Sammamish on Tuesday. Highline is a low bar, and Sammamish hasn’t carried the Kingco Banner for 30 years.  But at this point, Kingco 3A/2A girls need whatever they can get. “Just Win, Baby!”

New (Old) Data

I’ve posted additional years for 1986-1987 and 1987-1988 girls teams. With the addition of Walla Walla Union-Bulletin archives now available to me, via the King County Library System, I’ve also updated 1988-1989 and 1989-1990 for girls teams.

Three pointers

While researching I came across an article in the Walla Walla Union-Bulletin describing the Walla Walla girls’ game of January 21, 1989. Wa-Hi beat Eisenhower 52-38, not a surprise. The noteworthy thing about the game was that Angie Zagelow made the first three-point basket in Wa-Hi girls history. I don’t know when three pointers are added to the high school game.  The Spokesman Review added them to boxscores in 1987-88.  That might be it.  So in the 37th game, a Wa-Hi girl hit a three pointer.

That doesn’t appear to be that unusual for the times, though. Many GSL and Frontier boxscores in the Spokane newspapers list no three pointers made. A list of GSL girl high scorers (1987-1988) had 18 girls above 10 pts/game. Among those 18 girls, there were a total of 36 three-point baskets made; and one girl–Kristin Noe of Rogers–had 31 of them. I suspect there were GSL girls teams with nary a three pointer all season. Hard to imagine, nowadays, even a mediocre team without a couple of girls who can make threes.

RPI revisions

After some calculations to determine an ‘optimum’ RPI (whatever that is), I’m dropping OOWP from my RPI. The new weights will be 0.40 for WP and 0.60 for OWP. Previous seasons have been recalculated with this formulation. Further discussion.

Team Brunnhilde version 2.00

Mid July. 90 degrees under the shade on my deck. Pinot Noir and Puyallup loganberries.

But since March I’ve been busy. The data for this website (except for the games!) is now in a database. Ergo, version 2 of teambrunnhilde.com. Besides the games there is a lot of other data. Who played, what season, what league, what classification, who the out-of-state opponents were, what was their record, who changed names, who switched partners (for the co-op teams)? Games are almost an afterthought.

Anyway, you’ll notice that Crosspoint Academy school page now has King’s West and Bremerton Christian. Ritzville school page now has Lind-Ritzville and Lind-Ritzville/Sprague as well. No. Tekoa-Oakesdale doesn’t have Tekoa-Rosalia. But Rosalia does.  Getting the data into an actual database was the IT task for this off-season. The database isn’t accessible from teambrunnhilde.com. Maybe that will come. For now the pages are still static, but the perl scripts generate them from a database rather than from an ever increasing collection of files that only I could possibly figure out.

This was actually done in May or June. The out-of-state team records were added then. That addition allows good calculation of RPI back maybe 14 years. Data there which was obtained from OPSN, idahosports.com, and MaxPreps. Further back, I’ve plunged through scanned newspaper archives. It is a work in progress. The RPI calculations have been redone. Looking up the out-of-state school records also afforded an opportunity to regularize the school names and verify that the Washington school actually played that out-of-state team and not some name-alike school.

2018 ranking comparison

For the fourth year in a row, the TeamBrunnhilde linear model was the most accurate ranker in predicting the outcome of state tournament games.  The linear model for win-loss had an impressive debut, tying for second place with MaxPreps.  Fourth was a composite ranking.  WIAA RPI was fifth.  Everybody did a decent job of picking champions as their #1 ranked team.  Three rankers tied with eight #1’s that were champions: Evans Ranking, WIAA RPI, and TeamBrunnhilde linear model.  Lynden the top seed in Boys 2A won that championship, but was #2 in WIAA RPI.  Sorry, WIAA.  No mulligans.

I’ve read complaints that WIAA RPI doesn’t count Win-Loss enough.  Win-Loss as a ranking algorithm finished dead last.  I think there are better rankings than RPI, but RPI is an attempt to improve on Win-Loss with an understandable method for the less-mathematically inclined.  This year demonstrates that a properly done RPI does that.  Last year WIAA RPI had fatal flaws: not counting playoff games and using fake data for out of state teams.  With these corrected WIAA RPI finished 5th, well ahead of Won-Loss (which it trailed last year).

The TeamBrunnhilde linear model for win-loss, similar to TeamBrunnhilde linear model for points but with a different Y vector, did pretty good considering that the mathematics are not best suited for binary data (win-loss).  I’ll have to look at that before next year.

Everybody had successes.  Everybody had misses.  Not just at the top of the rankings.  We should all go back and work on slaying the variance monster next year.  But as Jesus might say, “The Random you will always have with you.”

Full results.

Regionals

Last year I saw six regional games and four were absolute barn burners. This year of six games seen, one went to the final seconds, one other was close in the fourth quarter, the other four ranged from clear wins to routs. In all, girls regional games reverted to their 2016 form: only 5 games decided by 5 or fewer points (16 last year), only 13 by less than 10. So I didn’t get unlucky in choosing my games.

One other thing.  I’ve probably seen between 1000 and 1500 games over the last 23 years.  A Garfield girls free throw shooter was called for exceeding 10 seconds before taking a shot.  That’s the first time I’d ever seen that called.  Justified?  I dunno.  The referee was likely the only one in the building counting.

Check out the WIAA Regional program cards. The picture: La Center girls (check: in 1A), White Swan girls (check: in 2B), Federal Way boys (check: in 4A), Anacortes boys. What happened to them? Program cards were free at Tumwater and Auburn Mountainview. They cost $1 at Bellevue College. Is that another case of getting rich people in blue areas to subsidize the rest of the state? Of the six program cards, three have good quality pictures of the teams. The ones that don’t were the ones that cost a buck. Horribly pixilated grainy photographs. And howcum a dollar anyway? For District 1/2 2A tournament, each team got a full page, often in color, and these were free. Game prices were lower also.

Bellevue College charged $5 for a hot dog and $2 for a very small bag of popcorn. When last I was there it was $3 and $1 for a bigger bag: same as the standard prices for booster club concessions at high school sites. Maybe price gouging for concessions is just to get us warmed up for the Tacoma Dome. At least the police at Bellevue College weren’t taking on the job of confiscating food coming into the venue. Seems that Tacoma’s Finest consider the duty of safeguarding exorbitant prices charged by T-Dome vendors for sub-standard food to be their primary job. Wouldn’t want to patrol the hilltop now, would we? OK, that was during the Goodwill games, snatching sandwiches from the mouths of four year olds in 1990. Nowadays Tacoma Dome food-Nazis work for an outsourced security firm. Today protecting $9 hot dogs in Tacoma—tomorrow getting billion dollar contracts to kill Iraqis waiting in gasoline lines.

Ranking comparison update

After 96 regional games, LM-pts (aka TeamBrunnhilde points, MY ranking) has 80 correct. Next is MaxPreps 73, Composite 70, LM-wl (see below) at 69, Evans Ranking at 68, WIAA RPI and Won-Loss at 67, Captured W-L at 64. Of the 57 games which all rankers agreed, 51 were picked correctly, nearly 90%. Overall, rankers did better this year than last at regionals, were the top ranker had only 73 right and WIAA RPI had 60 right. Do bad games make for better rankings?

Linear Model using win-loss, while coming up with reasonable answers (4th place so far), is upon further review not the using an optimum regression technique for binary win-loss kind of data. I’ll have to revisit my binary data texts from 35 years ago—a task for this coming off-season.

Contentious matchups for Wednesday, judging by the disagreements among the rankers, are Taholah v Almira Coulee Hartline (1B Boys), Prairie v Wilson (3A Boys), Lewis and Clark v Sunnyside (4A Boys), Entiat v Mount Rainier Lutheran (1B Girls), Seattle Christian v Lakeside (Nine Mile Falls) (1A Girls), East Valley (Yakima) v Port Angeles (2A Girls), Black Hills v White River (2A Girls). Thirty of 48 games on Wednesday are unanimous choices.