|
I'll be doing my poll this year; the first few weeks are always too close to last year's results to be interesting though.Captain von Trapp posted:After a really embarrassingly long time of coding, I wrote a Python script to convert the cfbstats.com files to the NCAA csv format. Please spot-check it to make sure I haven't borked the coding somewhere. Sadly it's just FBS. If someone finds this kind of data for FCS let me know and I'll include it as well. This is the result from cfbstats.com's data on their website now, which does not yet include this weekend's games. When it does I'll post that as well. I don't use this data but thanks for doing that! I hope everyone can get their polls up and running again this year.
|
# ¿ Sep 8, 2013 22:31 |
|
|
# ¿ May 22, 2024 15:26 |
|
Every week I keep thinking I'll finally get motivated and run my poll, and every week I don't bother. Maybe this will be the week.
|
# ¿ Nov 11, 2013 19:55 |
|
Alright, I finally got around to running my poll for the past 12 weeks.quote:Description The story of the season so far is Oregon took #1 from Bama in week 4 and has held it since then. They stayed a heavy favorite against Utah so they haven't dropped since the Stanford loss. The top four teams have been pretty far ahead of the rest until this week, when Ohio State jumped up closer to that group. My poll is really starting to show it's bias towards high-scoring teams this year thanks to Baylor, Oregon, and FSU. Tidbits: -This week the poll predicted the spreads of each game to within an average difference of 3.64 points. Not spectacular because it was averaging around 3.00 in the later weeks of 2012. I didn't make any changes to the formula this year. -FSU will probably lose points next week. My poll is predicting the spread to be a whopping -62.73, while the actual spread is "only" -56.5. This is an extreme example of how lopsided games have too much of an influence on my rankings. They'll only lose a point or two, but I should probably change the poll next year to have the influence variable scale down as the spread goes up. code:
|
# ¿ Nov 19, 2013 02:37 |
|
Hey so I finally ran the last six weeks of my poll. congrats to Oregon! FSU was actually in first going into bowls, but the huge spread for Oregon over Texas was enough to put them on top. Also my poll has been down on Auburn since the start. code:
wa27 fucked around with this message at 02:00 on May 6, 2014 |
# ¿ May 6, 2014 01:06 |
|
Woah! I just found a bug in my poll. Apparently Vegasinsider uses "pk" instead of 0 for an even matchup. I missed this for the past three years, so games that should have been even got practically random values assigned for their point spreads. I'm not sure how I didn't notice this before, especially in previous years when I spent a lot of time looking at the results. Thankfully Vegasinsider still has data for the past few years on their site. I've updated my last post with the actual final poll. Although I based my initial values off last year's final values, so changes I make to 2011's data might have a minor impact on 2013's results.
|
# ¿ May 6, 2014 02:07 |
|
|
# ¿ May 22, 2024 15:26 |
|
kayakyakr posted:Your poll still really, really likes TCU. It looks like, for some reason, bettors really liked TCU more than they should have. TCU was favored against Texas and Texas Tech, and heavily favored over West Virginia (which my poll also liked for a while). They went 4-8 against the spread. On the other end of the spectrum, Tulane didn't get any respect from bettors. They went 9-4 against the spread and were the lowest ranked team to go to a bowl.
|
# ¿ May 6, 2014 17:01 |