UT-Bama Model Prediction

#1
checkerboard03

checkerboard03

Well-Known Member
Joined
Dec 28, 2008
Messages
206
Likes
0
#1
I made a model to predict the results of UT - Bama, and I thought some posters here might be interested in that kind of thing...

It is based on each team's scoring offense and scoring defense, adjusted based on the quality of opposition.

There's good news and bad news...

The bad news is that the prediction is Alabama 28, Tennessee 16.

The good news is that Tennessee wins 26% of the time, which is a little bit higher than I would have expected, actually.

The model predicts Alabama will score between 22 and 35 points, and Tennessee will score between 4 and 32 points. The huge range for Tennessee is due to our schizophrenic offense - bad performances against UCLA and Auburn, and good showings against WKY and UGA.
 
#4
#4
Sorry I should have explained better. The model gives a predicted score for each team, but there is error around the prediction. That's how I came up with a 26% chance for UT... Bama's score ends up in the error beneath the prediction, and UT's comes in above the prediction.
 
#5
#5
Your model is based on subjective imput, did your model predict the GA game? no...If there was a way to predict the outcome of games you would live in Vegas making money hand over fist.
 
#6
#6
Yeah I know its not perfect. Honestly, this is the first time I've tried it, but you're probably right, it wouldn't have predicted the Georgia game.

I just had fun putting it together. Of course, anything could happen during the game.
 
#7
#7
Easy Frances!!!! He just made a prediction by using Stats and models, which for your 411 is how the folks that do live in Vegas formulate betting lines!
 
#8
#8
Yeah, but they also adjust the line to try to get a lot of action = more money.

And I'm sure their models are 100x better than this simple thing I did.
 
#10
#10
It's a fair projection if you account for statistics only. The problem with statistical analyses is that they don't account for intangibles (such how well-prepared each team is for the game and how determined UT is to keep winning in the wake of the big defeat of GA). After the close loss to the number one team in the land earlier in the season, I think Kiffin and company are going to have the Vols well-prepared to go into Tuscaloosa and knock off the new number 1 team. If South Carolina can hold Bama to only 20 points at home, Monte Kiffin's Berry-flavored defense should be able to do even better. If Crompton can keep up his winning ways and keep the Vol offense clicking on all 11 cylinders, I give the Vols a great chance at coming home with a W. Besides, it doesn't hurt to have confidence in our troops in what may be their biggest road game of the season. I think the boys would like knowing that the Vol Nation has faith in them. Go Vols!
:td:
 
Last edited:
#11
#11
Trying not to be a homer, if we hit a couple of big plays against them we could have a chance. I don't think Bama will score 22 points. In fact I think if we score 20, we could win. The tough thing about Bama is they can dominate you and only win by 13. Saban is very content letting his defense win games and kicking field goals. I say Bama will win 16-10, but I hope I'm wrong.
 
#12
#12
Special Ed said:
It's a fair projection if you account for statistics only. The problem with statistical analyses is that they don't account for intangibles (such how well-prepared each team is for the game and how determined UT is to keep winning in the wake of the big defeat of GA). After the close loss to the number one team in the land earlier in the season, I think Kiffin and company are going to have the Vols well-prepared to go into Tuscaloosa and knock off the number 2 team. If South Carolina can hold Bama to only 20 points at home, Monte Kiffin's Berry-flavored defense should be able to do even better. If Crompton can keep up his winning ways and keep the Vol offense clicking on all 11 cylinders, I give the Vols a great chance at coming home with a W. Besides, it doesn't hurt to have confidence in our troops in what may be their biggest road game of the season. I think the boys would like knowing that the Vol Nation has faith in them. Go Vols!
:td:
Click to expand...

That's a good assessment, I think. Bama's offensive performance against USC was its worst of the year (20 points against a D that gave up 18 ppg against its other opponents). If Bama's O continues to struggle, and our O plays the way we did against UGA, we should have a really good chance of pulling it out.
 
#13
#13
Game summary whoever can line up and run the ball will win the game. Both teams have to take all the responsibility of the QB's
 
#14
#14
Your model makes sense to me. It's hard to see how we're going to score more than 20 against bama--and i'd give us no more than a 30 percent chance of winning this game.
 
#15
#15
checkerboard03 said:
I made a model to predict the results of UT - Bama, and I thought some posters here might be interested in that kind of thing...

It is based on each team's scoring offense and scoring defense, adjusted based on the quality of opposition.

There's good news and bad news...

The bad news is that the prediction is Alabama 28, Tennessee 16.

The good news is that Tennessee wins 26% of the time, which is a little bit higher than I would have expected, actually.

The model predicts Alabama will score between 22 and 35 points, and Tennessee will score between 4 and 32 points. The huge range for Tennessee is due to our schizophrenic offense - bad performances against UCLA and Auburn, and good showings against WKY and UGA.
Click to expand...

Nostredamus would be ashamed......UT by 8
 
#16
#16
checkerboard03 said:
I made a model to predict the results of UT - Bama, and I thought some posters here might be interested in that kind of thing...

It is based on each team's scoring offense and scoring defense, adjusted based on the quality of opposition.

There's good news and bad news...

The bad news is that the prediction is Alabama 28, Tennessee 16.

The good news is that Tennessee wins 26% of the time, which is a little bit higher than I would have expected, actually.

The model predicts Alabama will score between 22 and 35 points, and Tennessee will score between 4 and 32 points. The huge range for Tennessee is due to our schizophrenic offense - bad performances against UCLA and Auburn, and good showings against WKY and UGA.
Click to expand...


You have a clear outlier in your data set...toss out data for WKY and see what the Bama game says.
 
Last edited:
#17
#17
NEVolFan said:
You have a clear outlier in your data set...toss out data for WKY and see what the Bama game says.
Click to expand...

That's true, and it is bad news for Tennessee if you take it out. Even if you take out Bama's game against N Texas, which is their most comparable win, the prediction still changes to Alabama 27, Tennessee 12, and Tennessee's win chance drops to 14%.

However, there are some other changes you could make to the model, such as increasing the weight of the most recent game, or only using Tennessee's defensive performance against pro-style offenses, which would help UT's chances.
 
#19
#19
checkerboard03 said:
I made a model to predict the results of UT - Bama, and I thought some posters here might be interested in that kind of thing...

It is based on each team's scoring offense and scoring defense, adjusted based on the quality of opposition.

There's good news and bad news...

The bad news is that the prediction is Alabama 28, Tennessee 16.

The good news is that Tennessee wins 26% of the time, which is a little bit higher than I would have expected, actually.

The model predicts Alabama will score between 22 and 35 points, and Tennessee will score between 4 and 32 points. The huge range for Tennessee is due to our schizophrenic offense - bad performances against UCLA and Auburn, and good showings against WKY and UGA.
Click to expand...

Good stuff. I appreciate you work. I once was a civil engineer and we did models for designing earthen dams based on unlikely events, such a 100 year floods.

The gist of the matter for Vol fans is there is a definite possibility(25%) to see a win despite the odds.
 
#20
#20
Other intangibles also need to include how well the UA team prepares for the game. Don't assume that only UT thinks the game has some bearing. The team is also coming to play with winning on their mind. So when intangibles make sure you add all them.
 
#21
#21
checkerboard03 said:
I made a model to predict the results of UT - Bama, and I thought some posters here might be interested in that kind of thing...

It is based on each team's scoring offense and scoring defense, adjusted based on the quality of opposition.

There's good news and bad news...

The bad news is that the prediction is Alabama 28, Tennessee 16.

The good news is that Tennessee wins 26% of the time, which is a little bit higher than I would have expected, actually.

The model predicts Alabama will score between 22 and 35 points, and Tennessee will score between 4 and 32 points. The huge range for Tennessee is due to our schizophrenic offense - bad performances against UCLA and Auburn, and good showings against WKY and UGA.
Click to expand...

so we play well and score 32, Bama plays poorly and scores 22 we win by 10...

Go Vols
 
#23
#23
TideWarrior said:
Other intangibles also need to include how well the UA team prepares for the game. Don't assume that only UT thinks the game has some bearing. The team is also coming to play with winning on their mind. So when intangibles make sure you add all them.
Click to expand...


How can you add an "intangible", which by definition cannot be explicitly constructed in mathematical terms. For example, you might say that since Alabama is undefeated, the game means more to them, but I might say since Tennessee has already lost three games that they will not feel any pressure and play loose and reckless.
 
#24
#24
interesting model.. but yeah, the huge range for TN kinda skewes it.. we just have to see which TN shows up.
 
Last edited:
#25
#25
rockydoc said:
[/B]

How can you add an "intangible", which by definition cannot be explicitly constructed in mathematical terms. For example, you might say that since Alabama is undefeated, the game means more to them, but I might say since Tennessee has already lost three games that they will not feel any pressure and play loose and reckless.
Click to expand...

I was actually quoting someone else in the thread that said was saying the same things in regards to UT. I was being sarcastic. Since intangibles can not be measured.
 
New posts
You must log in or register to reply here.

VN Store



Back
Top