Analytic models give way too much credit for losses to good teams. Its a loss. Not all losses are equal, but still an L

Report problem with this ad

Jan 3, 2025
2:04:07pm

Crazy Joe Davola Truly Addicted User

This message has been modified

Originally posted on Jan 3, 2025 at 2:04:07pm

Message modified by Crazy Joe Davola on Jan 3, 2025 at 2:04:57pm

Message modified by Crazy Joe Davola on Jan 3, 2025 at 2:05:40pm

Report problem with this ad

Crazy Joe Davola

Previous username: thebdawg
Bio page: Crazy Joe Davola
Joined: Jul 24, 2005
Last login: Jan 5, 2025
Total posts: 8,587 (26 FO)

Report problem with this ad

Messages

Author

Time

4 loss Bama's ranking, per the analytics

Indy Coug Loquacious Lummox

Jan 3, 2:02pm

Analytic models give way too much credit for losses to good teams. Its a loss. Not all losses are equal, but still an L

Crazy Joe Davola

Jan 3, 2:04pm

Analytics are primarily driven by scoring margins, adjusted for the quality of the opponent

Indy Coug

Jan 3, 2:05pm

Sounds like you prefer resume metrics, such as SOR, SP+ resume, or K Ford Most Deserving

TCuz

Jan 3, 2:10pm

Losses just need to be weighted heavily against a team. A 4 loss Bama team is not a top 10 team (top 5 in some of those)

Crazy Joe Davola

Jan 3, 2:15pm

The purpose of FPI is as a predictive tool. Maybe that’s the disconnect

TCuz

Jan 3, 2:20pm

There should be a difference, then, between ranking a team and predicting

Carnac

Jan 3, 2:26pm

But nobody is voting Alabama as a top 5 team, so I’m confused about what we are even mad about

TCuz

Jan 3, 2:31pm

Just realized they played Michigan in their final game 2 years in a row.

JAGA97

Jan 3, 2:04pm

That dang eye test

GFCougar

Jan 3, 2:05pm

This is the exact opposite of the eye test.

TCuz

Jan 3, 2:06pm

Right. This shows that the eye test does have some merit as opposed to the

GFCougar

Jan 3, 2:41pm

Hey, good losses matter!

TrueBlue67

Jan 3, 2:09pm

How does that make sense with losses and getting blown out in some games?

Tuskaloosa Cougar

Jan 3, 2:11pm

3 of the 4 losses were by a combined 18 points so only one was a blowout

Indy Coug

Jan 3, 2:14pm

Garbage data in... Garbage data out. The core metrics are bad.

BYUtka

Jan 3, 2:14pm

If the garbage beats the Vegas spread more often than not, how garbage-y is it

Indy Coug

Jan 3, 2:16pm

The SEC is getting filthy rich using it.

BYUtka

Jan 3, 2:41pm

And you could too. It's up to you if you want to be rich or be wrong.

Indy Coug

Jan 3, 2:43pm

They cheat with recruiting rankings. When someone gets an offer from an SEC school their star ranking goes up. These

HarlemCoug

Jan 3, 2:15pm

We should just stop playing games and rank based on measures.

Carnac

Jan 3, 2:12pm

How do you do that when everyone plays a vastly different schedule?

PHXSUNS31

Jan 3, 2:14pm

The measures come from playing the games.

Indy Coug

Jan 3, 2:14pm

Weird then. Four losses. Where do those play into the analytics?

Carnac

Jan 3, 2:16pm

It turns out when you play a weak schedule you don’t lose as much

Odysseus

Jan 3, 2:34pm

Explain Oklahoma.

Carnac

Jan 3, 2:37pm

Teams lose. Do you question all data and metrics when the Dodgers drop a series to the Rockies?

TCuz

Jan 3, 2:39pm

Should the games be played?

Carnac

Jan 3, 2:46pm

Of course. Games matter. That’s why Alabama was excluded despite these metrics.

TCuz

Jan 3, 2:48pm

A talented team in a tough environment played in the middle of an SEC schedule

Odysseus

Jan 3, 2:53pm

The system is so rigged that even when they lose they win. This is data that needs national attention. When you post it

HarlemCoug

Jan 3, 2:12pm

People should focus more on who teams beat than who they lost to

Odysseus

Jan 3, 2:31pm

Why? Both are relevant.

LifeLongLakerFan

Jan 3, 2:42pm

Of course they’re relevant. But we have to make judgments about who is better

Odysseus

Jan 3, 2:54pm

What about teams that you made score a lot of points to beat you? Think 🌹

GFCougar

Jan 3, 2:53pm

Do any of these models publish their assumptions and limitations?

kcca

Jan 3, 2:47pm

We should just do away with all ranking systems that don’t have Bama number 1.

JohnnyComeLately

Jan 3, 3:06pm

What did the analytics say about an Alabama-Michigan matchup?

Doogal McCougar

Jan 3, 3:32pm

RE: What did the analytics say about an Alabama-Michigan matchup?

Indy Coug

Jan 3, 3:42pm

I’m not saying they should be top 5… but come bball season we all say good

Briguy

Jan 3, 4:32pm

something seems very circular about all this

Bob Vance

Jan 3, 4:54pm

yup - and they'll start off next year in a good place too because they had another good recruiting class and because

Posting on CougarBoard