He's keeping last year's results as 15% of the rating all year long.

Sign up, and you'll be able to vote in polls. Sign up

Report problem with this ad

Sep 16, 2018
9:51:22am

RdF3 All-American

He's keeping last year's results as 15% of the rating all year long.

Report problem with this ad

RdF3

Previous username: swishh_15
Bio page: RdF3
Joined: Nov 4, 2011
Last login: May 4, 2024
Total posts: 39,957 (2,657 FO)

Report problem with this ad

Messages

Author

Time

Sagarin's new rating system with last year always weighting so heavily is dumb

RdF3

9/16/18 9:47am

Wait till week 5 or 6 when last year’s results no longer factor in

NorCal_Coug

9/16/18 9:48am

Answered my question

9/16/18 9:49am

He's keeping last year's results as 15% of the rating all year long.

RdF3

9/16/18 9:51am

Wow. Idiotic

9/16/18 9:52am

Especially since last year BYU underperformed to talent substantially

9/16/18 9:52am

It's not idiotic. It makes the model more accurate overall. There will always

molodyets

9/16/18 10:00am

I’ll disagree with you on that slightly

9/16/18 10:25am

You make large assumptions that he didn't consider and test that as well.

molodyets

9/16/18 10:41am

I guess - his process is evolving - he’s said so in interviews this fall

9/16/18 10:45am

It's never perfect. These things are shooting for 75% ish straight up and 53%+

molodyets

9/16/18 10:51am

Sure. no disagreement. I think we can agree on that statement fully

9/16/18 10:59am

When does last year drop off completey

9/16/18 9:49am

2019

BYUinTX

9/16/18 9:59am

<< Deleted >>

Mr. Scoobs

9/16/18 9:49am

As long as it is consistent what is the problem?

TNT

9/16/18 9:50am

I think it's dumb that even in December, 15% of a team's rating will be based

RdF3

9/16/18 9:52am

What do you propose?

TNT

9/16/18 9:53am

Not using it?

9/16/18 9:54am

Dropping off last year's results after week 4 last he used to.

RdF3

9/16/18 9:54am

It does. They back test these things. How do you expect it to perform

molodyets

9/16/18 9:59am

And what's he using a base accuracy off of? The AP top 25? His opinion of how

RdF3

9/16/18 10:10am

On field results of what it predicted vs the outcome.

molodyets

9/16/18 10:12am

It would make sense to do a 5 year average of final ratings.

specsauce

9/16/18 10:08am

<< Deleted >>

Abe Froman

9/16/18 10:19am

Report problem with this ad

Posting on CougarBoard

In order to post, you will need to either sign up or log in.

Report problem with this ad