Started By
Message
re: A statistics nerd on Reddit plugged all 128 teams into a computer model
Posted on 8/25/15 at 3:52 pm to Tornado Alley
Posted on 8/25/15 at 3:52 pm to Tornado Alley
This is me and my work. I post the sheets on the MSB in the betting thread.
I understand its not a perfect system but its not really meant to predict an entire season. These are just my sheets I use for gambling purposes. I might use the whole season predictions for win totals and futures but thats it. Creating preseason ratings that make sense is so difficult that you can't please everyone, and its a crap shoot anyway. The MAIN REASON I made preseason ratings was so my system could have something to work with the ifrst 7-8 weeks of the year while I wait for the sample size of games played to build up. In the past I had to wait until Week 9 to use my system as the sample size was too low and therefore my system's output was unstable.
Hell I created the sheets in the Reddit post on Saturday, and I have already made a few tweaks and changes to my system so those sheets don't reflect where my system is at right now. For example my numbers now have Alabama losing to Georgia last time I checked.
The only thing I really care about is Week 1. My sheets go into much more detail on a week by week basis. Here you go South Carolina fans, here is what my model does on a per-game basis.
I understand its not a perfect system but its not really meant to predict an entire season. These are just my sheets I use for gambling purposes. I might use the whole season predictions for win totals and futures but thats it. Creating preseason ratings that make sense is so difficult that you can't please everyone, and its a crap shoot anyway. The MAIN REASON I made preseason ratings was so my system could have something to work with the ifrst 7-8 weeks of the year while I wait for the sample size of games played to build up. In the past I had to wait until Week 9 to use my system as the sample size was too low and therefore my system's output was unstable.
Hell I created the sheets in the Reddit post on Saturday, and I have already made a few tweaks and changes to my system so those sheets don't reflect where my system is at right now. For example my numbers now have Alabama losing to Georgia last time I checked.
The only thing I really care about is Week 1. My sheets go into much more detail on a week by week basis. Here you go South Carolina fans, here is what my model does on a per-game basis.
1
Posted on 8/25/15 at 3:56 pm to accnodefense
How cute, you use last year's data.
News flash, players graduate and new players come in. Sometimes new players can be studs and player replacing the graduated players can be worse or better.
Gotta love computer geeks thinking they have the game of football predictions mastered.
Those computer geeks always miss the biggest things with their calculations - the HUMAN factor.
News flash, players graduate and new players come in. Sometimes new players can be studs and player replacing the graduated players can be worse or better.
Gotta love computer geeks thinking they have the game of football predictions mastered.
Those computer geeks always miss the biggest things with their calculations - the HUMAN factor.
Page 1 of 1
Popular
Back to top
Follow SECRant for SEC Football News