- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: Week 12 CFB Bet Thread
Posted on 11/19/15 at 5:34 pm to accnodefense
Posted on 11/19/15 at 5:34 pm to accnodefense
Hey guys, I have a question that I think you guys might be helpful with.
I'm working on a project for a programming class. The project must deal with data analysis and file input. My plan is to have the user input two CFB teams and the script would then spit out the predicted winner after analyzing the two different spreadsheets. The hard part here is trying to nail down the prediction model. That's where I need your help.
I've been playing around with cfbstats.com and I have my code importing the Team Stats into a spreadsheet pretty nicely. In theory I would go through the criteria and assign points to each team based on which one checked the criteria. Team with the most points is the predicted winner. Or something like that.
So any advice for a model (based on cfbstats.com, example LINK , or offer me a better stats site)is welcome. This won't be fool proof by any means, but maybe 60-70%?
*Don't mean to derail the thread, just thought you guys might know a thing or two.
I'm working on a project for a programming class. The project must deal with data analysis and file input. My plan is to have the user input two CFB teams and the script would then spit out the predicted winner after analyzing the two different spreadsheets. The hard part here is trying to nail down the prediction model. That's where I need your help.
I've been playing around with cfbstats.com and I have my code importing the Team Stats into a spreadsheet pretty nicely. In theory I would go through the criteria and assign points to each team based on which one checked the criteria. Team with the most points is the predicted winner. Or something like that.
So any advice for a model (based on cfbstats.com, example LINK , or offer me a better stats site)is welcome. This won't be fool proof by any means, but maybe 60-70%?
*Don't mean to derail the thread, just thought you guys might know a thing or two.
Posted on 11/19/15 at 5:50 pm to Ssubba
quote:
Hey guys, I have a question that I think you guys might be helpful with.
The stats you are going to find on cfbstats will not be predicative enough to come up with a good predicative model. They are very basic level stats and in football you need more context, i.e. what yard line the drive started on, how much time is on the clock, whether or not the drive is a "junk" drive.
At minimum you will need drive level data. Every single drive, offense and defense, every FBS vs. FBS game, for all 128 FBS teams. It's hard to come by for free. Ideally, you would also have play-by-play data for all teams, but that is near impossible to come by without shelling out major $$$. I had someone supply me with drive data, but every now and then a game would be missing and I would have to manually enter in the drive data myself, which took about 45 minutes a game. So if 2-3 games were missing I'd have to spend 2 hours doing the tedius task of entering in drive data. Yuck.
I ran a model this season up until the past two weeks(just don't have the time to update it and my stat supplier was becoming shaky as I mentioned in the above paragrah) and was hitting about 52.9%. Barely enough to be profitable.
I am a professional statistician(its what I do for a living) and this model I have used hit 52.9% this year and 53.2% last year. Modeling football is probably a fool's errand and this is probably my last year to do it. I have more success in basketball(College and WNBA).
60%-70% is impossible for football even if you had the best predicative model in the world. Too much variance and situational noise to make any football model that hits above 55% in the long run.
If you are going to model do basketball. All you need is a general box score to create a good system.
Posted on 11/19/15 at 6:10 pm to PeteRose
quote:
BG -4 1H
Ohio -.5 1Q
O BG 71
O BG 35 1H
Cmich -.5 1Q
U ULM 14 1Q
U ULM 30.5 1H
Posted on 11/19/15 at 6:13 pm to accnodefense
Sounds like he's only trying to predict the winner, not the winner ATS. And if its just for a class, it doesn't need to be near as complex or complete as yours, ACC. BTW, thanks for the work you put in during the season. Several of us were using your model, as you already know.
Posted on 11/19/15 at 7:02 pm to accnodefense
quote:
accnodefense
Fantastic post. Well said
Posted on 11/19/15 at 7:19 pm to maclauer
Heavy: Virginia Tech +6, Ohio State -13, Baylor pk
Light: Iowa State +6, Georgia Southern +14
Light: Iowa State +6, Georgia Southern +14
Posted on 11/19/15 at 10:53 pm to accnodefense
quote:
accnodefense
that avi
Posted on 11/20/15 at 7:30 am to accnodefense
ACC,
I had no idea how much time and effort you put into your models. Thanks for all your hardwork and most of all sharing your models with us!
I had no idea how much time and effort you put into your models. Thanks for all your hardwork and most of all sharing your models with us!
Posted on 11/20/15 at 8:43 am to ChemE in the OP
Updated Week 12 Card ~
Ole Miss -4
Ohio State -13
Utah -2
Ok State PK
Noticed Ole Miss and Ok State are moving my way...
Ole Miss -4
Ohio State -13
Utah -2
Ok State PK
Noticed Ole Miss and Ok State are moving my way...
Posted on 11/20/15 at 9:24 am to ChemE in the OP
Surprised Utah hasn't moved with Booker out.
Posted on 11/20/15 at 9:29 am to JG77056
Yeah I like watching Booker. Bruising runner for his size. Utes will go with Joe Williams. No many touches this year. Bruins Run D not all that good.
Posted on 11/20/15 at 11:03 am to PeteRose
Anyone like air force and the points?
Posted on 11/20/15 at 11:10 am to bloupe2
quote:
Anyone like air force and the points
I'm on AF
Posted on 11/20/15 at 11:17 am to Chuckiee
Yeah I'm about to throw down on AF (half unit lol)
Posted on 11/20/15 at 11:24 am to bamaatlsu
quote:
Sounds like he's only trying to predict the winner, not the winner ATS
If its just the winner, the stats will be fine. Just go with yards per play (offense and defense) and turnovers.
Posted on 11/20/15 at 11:54 am to accnodefense
ACC anything you like in cincy game tonight
Posted on 11/20/15 at 12:45 pm to bloupe2
quote:
ACC anything you like in cincy game tonight
No I do not
Posted on 11/20/15 at 2:54 pm to accnodefense
Oklahoma line finally posted... opened at -10, now at -11.
I planned on playing TCU but I think that line is actually favorable to Oklahoma. Will wait to see how it moves. Would love to grab TCU at +14.
I planned on playing TCU but I think that line is actually favorable to Oklahoma. Will wait to see how it moves. Would love to grab TCU at +14.
Posted on 11/20/15 at 3:00 pm to ChemE in the OP
quote:
Noticed Ole Miss and Ok State are moving my way...
Nah, Baylor is moving my way.
Edit: not that pk to 1 means a damn thing
This post was edited on 11/20/15 at 3:08 pm
Popular
Back to top
Follow TigerDroppings for LSU Football News