A good, honest (and a little bit long) look at the Score Czar’s computer football rankings
Hey, it’s Thanksgiving week, and all of the football teams in Clark County are done for the season.
So I had some time on my hands. And if you have some time on yours, here’s a blog post about The Score Czar and his computer rankings.
This blog post was inspired by a podcast I recently listened to that featured the Score Czar and Ryland Spencer.
The Score Czar is Vancouver’s Scott Odiorne who has spent the past seven years using a mathematical algorithm to rank high school sports teams. Spencer is a prep football aficionado who writes for Northwest Elite Index, a website that promotes high school football players in the state.
The two met recently to discuss perceived flaws in the manner in which the WIAA draws its state tournament brackets for football, and how the Score Czar’s rankings could be used to help rectify these flaws.
It’s an interesting dialogue and you can listen to it here.
The Score Czar throws out a lot of impressive statistics that support the validity of his rankings. But there was one number he threw out that bugged me when I first heard it, and it still bugs me now.
And it’s this stat that we will discuss here.
The Score Czar said that in all the sports he ranks (he does football, soccer, basketball, baseball and softball) in the 13 states where he ranks teams, the average Score Czar ranking of all the state champions in all of those sports in all of those states was 1.7.
That’s an impressive number. I know I was impressed when I first heard it.
But here’s the thing. That ranking he refers is how his computer ranks those state champions AFTER they’ve already won the state championship.
You see, every week in the season from Week 3, the Score Czar takes scores from games around the state, puts them into his algorithm and the computer spits out rankings.
But he doesn’t stop ranking teams once the state playoffs begin. He takes the results from the first round of state, and the computer spits rankings. He does it again after the quarterfinals, and after the semifinals and after the state championship game. And that’s the ranking he’s talking about.
The Score Czar claims that the final ranking shows that his system works. But to me, it seems a little like 20-20 hindsight.
I’m sure if the Associated Press voters were asked to rank teams AFTER the state championship game had been played, I’m pretty sure the average ranking of the state champions would be 1.0.
The best way, and really only way, to look at the Score Czar rankings honestly is to do it after the Week 10 playoffs, which are used to determine the 16-team state tournament. That’s because it is these rankings that the Score Czar is suggesting the WIAA use to help construct a more equitable state tournament bracket.
So that is what I’ve done here. Below are the four semifinal teams in each classification along with their Score Czar ranking after Week 10 (and for fun I threw in their Associated Press ranking from the same time frame).
In theory, if the Score Czar rankings are accurate, the top 4 ranked teams from each classification should still be alive, unless of course two top-4 ranked teams met in the first round or quarterfinals. So I’ve listed the four semifinalists, plus I listed how any top 4 team not in the semifinal was eliminated.
CLASS 4A
- No. 2 Lake Stevens (AP 2)
- No. 3 Gonzaga Prep (AP 4)
- No. 5 Richland (AP 6)
- No. 7 Skyline (AP 7)
How other top 4 teams were eliminated:
- No. 1 Gig Harbor lost to No. 7 Skyline 17-15 in first round
- No. 4 Camas lost to No. 7 Skyline 35-10 in quarterfinals
CLASS 3A
- No. 1 Eastside Catholic (AP 1)
- No. 3 Bellevue (AP 2)
- No. 8 Bishop Blanchet (AP 6)
- No. 9 Lakes (AP 9)
How other top 4 teams were eliminated:
- No. 2 Kennedy Catholic lost 49-15 to No. 3 Bellevue in first round
- No. 4 Lincoln lost 35-28 to No. 23 Bonney Lake in first round
CLASS 2A
- No. 3 Archbishop Murphy (AP 4)
- No. 4 Prosser (AP 6)
- No. 6 Tumwater (AP 1)
- No. 13 Sedro-Woolley (AP not ranked)
How other top 4 teams were eliminated:
- No. 1 Hockinson lost 31-28 to No. 13 Sedro-Woolley in quarterfinals
- No. 2 Olympic lost 34-14 to No. 13 Sedro-Woolley in first round
CLASS 1A
- No. 1 Royal (AP 1)
- No. 3 Connell (AP 2)
- No. 4 Montesano (AP 10)
- No. 12 King’s (AP 3)
How other top 4 team was eliminated:
- No. 2 Port Townsend lost 24-7 to No. 12 King’s in first round
CLASS 2B
- No. 1 Pe Ell-Willapa Valley (AP 1)
- No. 4 Napavine (AP 6)
- No. 5 Okanogan (AP 4)
- No. 6 Northwest Christian-Colbert (AP 7)
How other top 4 teams were eliminated:
- No. 2 North Beach lost 20-14 to No. 4 Napavine in the first round
- No. 3 Lind-Ritzville/Sprague lost 35-28 to No. 5 Okanogan in quarterfinals
CLASS 1B
- No. 1 Neah Bay (AP 2)
- No. 4 Almira-Coulee-Hartline (AP 3)
- No. 5 Liberty Christian (AP 1)
- No. 9 Lummi Nation (AP 7)
How other top 4 teams were eliminated:
- No. 2 Touchet lost 30-22 to No. 4 Almira-Coulee-Hartline in quarterfinals
- No. 3 Evergreen Lutheran lost 50-0 to No. 9 Lummi Nation in first round
So again, if we look at the top-4 ranked teams, the Score Czar correctly predicted 13 of the 24 semifinal teams in the six classifications. Coincidentally, that’s the same number the AP votes corrected projected.
BUT, there were three instances in which two top-4 teams met in the early rounds of state. In two of the three instances, the Score Czar’s No. 5 advanced to the semifinals. So we can actually give him credit for 15 of the 24 semifinalists.
That’s pretty good. Not great. Not really even “Scary Good” as the Score Czar likes to tout his rankings on his site scoreczar.com. Just pretty good.
If we look at top-2 ranked teams, the Score Czar only has five of 12 teams still alive. And in no classification does he have both the No. 1- and No. 2-ranked teams still alive.
The AP voters have 9 of 12 top-2 teams still alive. And one of those teams eliminated was eliminated by a No. 1 – when No. 1 Tumwater beat No. 2 Squalicum in last week’s 2A quarterfinals.
So when you look at it that, the Score Czar doesn’t fare as well.
And when you look closer at his misses, an inherent flaw is revealed.
Of the Score Czar’s nine misses on top-4 teams, four lost in very close games. But five lost by a significant margin, and in most cases, to a team ranked much lower.
So either the team that lost was over-ranked, the team that won was under-ranked, or a combination of the two.
The most egregious example of this was Kennedy Catholic in 3A.
The Score Czar predicted that the Lancers would beat Bellevue in the first round, even though anyone with a pulse – and that would include the Score Czar himself – knew that Bellevue was going to win that game and by a good margin.
The Score Czar explains that Kennedy plays in a 3A/2A league that is heavy with 2A teams, and 2A teams that aren’t all that strong.
The Score Czar’s rankings take three components into account: the points a team scores, the points a team allows and a strength-of-schedule component.
The Score Czar said Kennedy played a very weak schedule. But the strength of schedule component could not mitigate the fact that the Lancers scored almost 400 points on the season while allowing just 68.
Then the Score Czar dismisses Kennedy’s errant ranking by saying: “There are 400 teams in the state, and I got one wrong.”
This remark really bothered me. Because I think if you look hard at the Score Czar rankings, a discerning observer will find several other teams that are misranked similar to how Kennedy was. It is a by-product of a significant hole is the Score Czar’s computations.
If the strength-of-schedule component cannot bring some of these teams back to reality, then what you have is a system the rewards teams for playing soft schedules while penalizing teams that play tough ones.
Take a look at King’s at 1A. The AP votes had King’s at No. 3. The Score Czar had them at No. 12.
Why? Well King’s, which was 9-1 after Week 10, played a tough schedule that included some good 2A schools, including a 49-7 loss to Archbishop Murphy. Archbishop Murphy is one of the top 2A teams, if not THE best 2A team in the state.
But that one lopsided loss dropped King’s in the Score Czar’s rankings.
Meanwhile, Port Townsend (10-0) got fat off a weak schedule that included having the Redhawks playing the same four weak opponents twice each.
Because of that, Port Townsend was No. 2 in the Score Czar’s rankings.
So it should come as no surprise that King’s beat Port Townsend 24-7 in the first round at state.
Let’s make one thing perfectly clear: I am not anti-Score Czar. I think he is onto something good here with his rankings. I just don’t think his system is good enough to be employed in WIAA bracket seeding …. yet.
If the Score Czar is willing to take a good hard look at his system, and fix its flaws – and the one listed here is a significant – I would be more willing to get behind his system.
And that’s the thing. If you want the WIAA to consider changing the way it’s been doing things for a long, long time, you need to come with a proposal that is ironclad.
Because if flaws are clearly visible in your proposal, the WIAA will say “No thanks. We’ll stick with what we’ve been doing.”
