Started By
Message

SOG Prediction Model (1st real week) - #1 Alabama (-17.5) @ Texas A&M

Posted on 10/4/21 at 3:11 pm
Posted by SummerOfGeorge
Member since Jul 2013
102699 posts
Posted on 10/4/21 at 3:11 pm
Alright, we've got 3 games worth of data on teams (and their opponents vs other teams) now, so we're starting to get enough that we can actually see some stuff. Let's roll with it and see what we get.



VEGAS
Line : Alabama by 17.5
O/U : 51.5
Vegas Implied Score : Alabama 35-17


PREDICTION
ALABAMA - 35
TEXAS A&M - 11

Alabama Rush - 137 yards on 37 carries (3.7 YPA)
Alabama Pass - 318 yards on 33 attempts (9.7 YPA)
Alabama Total - 455 yards on 70 plays (6.5 YPP)


- Alabama RUSH OFFENSE is averaging 88% of opponent averages
- Alabama PASS OFFENSE is averaging 126% of opponent averages
- Alabama has averaged 140% of opponents Points Per Yard
- Alabama RUSH DEFENSE is allowing 82% of opponent averages
- Alabama PASS DEFENSE is allowing 78% of opponent averages
- Alabama has allowed 101% of opponents offenses Points Per Yard



Texas A&M Rush - 119 yards on 29 carries (4.1 YPA)
Texas A&M Pass - 147 yards on 34 attempts (4.3 YPA)
Texas A&M Total - 265 yards on 63 plays (4.2 YPP)


- Texas A&M RUSH OFFENSE is averaging 117% of opponent averages
- Texas A&M PASS OFFENSE is averaging 69% of opponent averages
- Texas A&M has averaged 63% of opponents Point Per Yard
- Texas A&M RUSH DEFENSE is allowing 104% of opponent averages
- Texas A&M PASS DEFENSE is allowing 122% of opponent averages
- Texas A&M has allowed 81% of opponents offenses Point Per Yard average
This post was edited on 10/4/21 at 3:14 pm
Posted by biclops
Member since Oct 2011
6149 posts
Posted on 10/4/21 at 3:13 pm to
SoG I just had an idea to possibly improve model accuracy. Instead of looking at per-game total rushing/passing yardage as a percent of opponent averages, look at *per play* averages (split running vs passing plays) to account for teams that like to "take what the defense gives". Does that make sense?

EDIT: Nevermind, it looks like that's what you already do?
This post was edited on 10/4/21 at 3:15 pm
Posted by SummerOfGeorge
Member since Jul 2013
102699 posts
Posted on 10/4/21 at 3:14 pm to
quote:

SoG I just had an idea to possibly improve model accuracy. Instead of looking at per-game total rushing/passing yardage as a percent of opponent averages, look at *per play* averages (split running vs passing plays) to account for teams that like to "take what the defense gives". Does that make sense?



That would be a very good change except that's exactly what is already done!

Your brain is in the right place
This post was edited on 10/4/21 at 3:15 pm
Posted by Robot Santa
Member since Oct 2009
44316 posts
Posted on 10/4/21 at 3:20 pm to
It'll be interesting to see if we have one of our usual big game hangover performances, or if Aggy just rolls over and dies. I will be extremely disappointed if we give up more than 14 though. Calzada is probably the worst QB we will see all season.
Posted by SummerOfGeorge
Member since Jul 2013
102699 posts
Posted on 10/4/21 at 3:23 pm to
quote:

It'll be interesting to see if we have one of our usual big game hangover performances, or if Aggy just rolls over and dies. I will be extremely disappointed if we give up more than 14 though. Calzada is probably the worst QB we will see all season.


They've been dreadfully terrible on offense consistently and their defense isn't THAT good either (good, but not elite or shut down worthy against good offenses).

That Colorado team they almost lost to is TERRIBLE, Arkansas is better but not great (and they got thoroughly beat) and State is average. I could see them slowing us down on offense, but like you said, if we give up more than 14-17 pts then that's a problem. They can't run block, their QB is a disaster and they have no scary receivers.
This post was edited on 10/4/21 at 3:24 pm
Posted by Robot Santa
Member since Oct 2009
44316 posts
Posted on 10/4/21 at 3:35 pm to
It's like a repeat of post-Jameis FSU, only they skipped over Golson and Francois. I still cannot believe they gave Jimbo that contract. And LSU fans try to act like Woodward is some wizard of an AD.
Posted by JoseyWalesTheOutlaw
In The Ham
Member since Nov 2017
11654 posts
Posted on 10/4/21 at 3:36 pm to
If Bryce plays as he has then Bama should be in pretty good shape.
Posted by SummerOfGeorge
Member since Jul 2013
102699 posts
Posted on 10/4/21 at 3:47 pm to
quote:

It's like a repeat of post-Jameis FSU, only they skipped over Golson and Francois


And they skipped the whole winning titles part of the Jameis era
Posted by 14&Counting
Eugene, OR
Member since Jul 2012
37559 posts
Posted on 10/4/21 at 4:02 pm to
quote:

nd they skipped the whole winning titles part of the Jameis era


Posted by Brostache48
Member since Dec 2019
554 posts
Posted on 10/4/21 at 5:17 pm to
Just for shits and giggles, how hard would it be to run the model on Bama and UGA?
Posted by TideSaint
Hill Country
Member since Sep 2008
75823 posts
Posted on 10/4/21 at 6:49 pm to
quote:

Just for shits and giggles, how hard would it be to run the model on Bama and UGA?




Easy, fella.

One game at a time.
Posted by Brostache48
Member since Dec 2019
554 posts
Posted on 10/4/21 at 7:22 pm to
Just curious more than anything else. They really haven't played a offense with a pulse yet and I'm curious how the prediction model would look
Posted by biclops
Member since Oct 2011
6149 posts
Posted on 10/4/21 at 7:30 pm to
quote:

Just for shits and giggles, how hard would it be to run the model on Bama and UGA?




The model would show UGA crushing bama. Their defense is fun to watch. I think we can beat them but it will take our A game.
This post was edited on 10/4/21 at 7:33 pm
Posted by Brostache48
Member since Dec 2019
554 posts
Posted on 10/4/21 at 7:33 pm to
The model might show that, true. Just curious though. I'm not sold on their offense even with Daniel's healthy. I really feel like our offense, if it keeps progressing like it has, can score at least 20-30 points on them. I think our defense can keep us in a game like that
This post was edited on 10/4/21 at 7:37 pm
Posted by Brostache48
Member since Dec 2019
554 posts
Posted on 10/4/21 at 7:41 pm to
Guess I should worry about the games coming up like tidesaint says, never know what the season has in store for us or anybody else. You know what they say about curiosity
Posted by BFANLC
The Beach
Member since Oct 2007
18116 posts
Posted on 10/4/21 at 8:58 pm to
Your pred. model made a mess of my nerves. Not even looking this week lol
Posted by Panthers4life
Huntsville
Member since Nov 2017
4350 posts
Posted on 10/4/21 at 9:59 pm to
Any model predicting how many deep passes Young will make or how many high snaps?
Posted by Bamadiver
Member since Jun 2014
3222 posts
Posted on 10/5/21 at 7:20 am to
A&M center and two CBs went ahead and had surgery. I'm thinking that'll wreck the model. Bama covers.
Posted by SummerOfGeorge
Member since Jul 2013
102699 posts
Posted on 10/5/21 at 7:41 am to
quote:

Your pred. model made a mess of my nerves. Not even looking this week lol



LOL I told you guys not to take it seriously.

This one should be taken more seriously (unless it's way off then we need more data ). Kidding, this week everybody has 2-3 games against other teams and those teams have 2-3 games against hose teams. It all kind of exponentially builds out once you get to early/mid October.
Posted by elposter
Member since Dec 2010
24817 posts
Posted on 10/5/21 at 8:04 am to
quote:

Just for shits and giggles, how hard would it be to run the model on Bama and UGA


Nobody is trying to be depressed in the middle of the week bro.
first pageprev pagePage 1 of 2Next pagelast page

Back to top
logoFollow SECRant for SEC Football News
Follow us on Twitter and Facebook to get the latest updates on SEC Football and Recruiting.

FacebookTwitter