clock menu more-arrow no yes

Filed under:

Forecasting The Big 12 South - 2009

New, comments

Anybody have a problem with speculative guessing, supported by a framework of decent analysis techniques but sketchy inputs? I didn't think so.

In trying to predict the Big 12 South this year, what information do we have to work with? Only the schedules. Any power ratings of teams are just guesses and predictions right now. Once the games start, and the results come in, we'll have a much better idea of relative quality and chances to win. I tried a little exercise where I used Phil Steele's predicted power ratings, projected a resulting point spread and odds of each team winning. Here's a sample:

Texas Pts Odds
ULM -35 0.99
@Wyo -32 0.98
TT -17 0.84
UTEP -28 0.98
CU -23 0.96
OU -1 0.51
@Mizzou -19 0.88
@OSU -6 0.59
UCF -30 0.98
@BU -18 0.86
KU -25 0.97
@TAMU -20 0.9

I added the odds of each game up to come up with an expected win total, and then multiplieded the odds of each game to predict the chances of an undefeated season. I took the 12th root of that, to create a rough even chance of winning each game (because I'm too lazy to do Monte Carlo modeling macros), and used that in the binomial distribution function to figure chances of winning 10 or 11 games. Here's what that looked like...

Expected (sum) 10.44
Exp. B12 rec 6.51
root odds 0.85
Bowl Eligible Odds 99.94%
10 win odds 29.06%
11 win odds 30.77%
12 win odds 14.94%

This says Texas is expected to go 10.44 - 1.56 (ha-ha), go 6.51 - 1.49 in the Big 12, and has almost a 15% chance of going undefeated (through the regular season. Assume the B12S champ has a 60 - 65% chance of winning the B12 CG and 50% chance of winning the bowl game). Here it is for the other B12 south teams:

OU
BYU -10 0.65 Expected (sum) 9.375
Idaho St -40 0.99 Exp. B12 rec 6.135
Tulsa -29 0.98 root odds 0.76
@Mia -8 0.62
BU -23 0.93 Bowl Eligible Odds 98.87%
Texas 1 0.49 10 win odds 24.45%
@KU -17 0.84 11 win odds 14.08%
KSU -29 0.98 12 win odds 3.72%
@NU -8 0.62
TAMU -30 0.99
@tt -9 0.635
OSU -10 0.65
OSU
UGA -1.5 0.515 Expected (sum) 9.345
Hou -19 0.88 Exp. B12 rec 5.97
Rice -35 0.99 root odds 0.74
Gram -40 0.99
@TAMU -24 0.94 Bowl Eligible Odds 98.13%
Mizzou -24 0.94 10 win odds 21.64%
@BU -11 0.72 11 win odds 11.05%
Texas 6 0.41 12 win odds 2.59%
@ISU -32 0.98
TT -10 0.7
CU -23 0.93
@OU 10 0.35
TT
ND -40 0.99 Expected (sum) 7.81
Rice -35 0.99 Exp. B12 rec 4.52
@Texas 17 0.16 root odds 0.57
@Hou -6 0.59
UNM -35 0.99 Bowl Eligible Odds 78.59%
KSU -17 0.84 10 win odds 4.48%
@NU 7 0.395 11 win odds 1.08%
TAMU -23 0.93 12 win odds 0.12%
KU -11 0.72
@OSU 10 0.3
OU 9 0.365
@BU -4 0.54
TAMU
UNM -25 0.95 Expected (sum) 4.625
Utah St -18 0.86 Exp. B12 rec 1.955
UAB -14 0.78 root odds 0.21
Ark 22 0.08
OSU 24 0.06 Bowl Eligible Odds 2.18%
@KSU 9 0.365 10 win odds 0.00%
@TT 23 0.07 11 win odds 0.00%
ISU -11 0.72 12 win odds 0.00%
@CU 11 0.28
@OU 30 0.01
BU 10 0.35
Texas 20 0.1
BU
@WF -7 0.605 Expected (sum) 6.825
Conn -18 0.86 Exp. B12 rec 3.38
NWSt -35 0.99 root odds 0.45
Kent St -35 0.99
@OU 23 0.07 Bowl Eligible Odds 48.04%
@ISU -18 0.86 10 win odds 0.71%
OSU 11 0.28 11 win odds 0.11%
NU 8 0.38 12 win odds 0.01%
@Mizzou -4 0.54
Texas 18 0.14
@TAMU -10 0.65
TT 4 0.46

You can quibble with the spreads and odds if you like. Disagree with power ratings? Revise them and the odds to update to your liking. I like Phil Steele's ratings because:

1. He puts thought into them

2. I didn't develop them, so they are free of my personal bias.

Once the season starts, I can update with real results and Sagarin predictions. Observations:

Phil Steele really likes Baylor this year. He has them rated notably higher than Wake Forest, Connecticut, and TAMU. I was skeptical, but we noted in the earlier post about defenses that the Baylor DC knows what he's doing, and he only loses three starters. That team is not all Robert Griffin.
Texas' two biggest games are OU and @OSU.
OU's biggest game is Texas of course, but @Miami, @NU, and @TT are threats.
OSU's biggest games are UGA, Texas, and @OU. Note that two are at home.
Tech looks to have a big dropoff this year.
The Baylor @ TAMU game could be a lot of fun. I kind of expect TAMU to be better than predicted here. One of the forecast tools I use is "How are a team's fortunes affected if their most important player is hurt?" TAMU is relatively immune to that- does it really change them much if Johnson or Fuller is out? Baylor, on the other hand, may be near-Rice crappy if Griffin were hurt...and he is a running QB who has to play UT and OU. Obviously, an injury to McCoy or Bradford will quickly revise UT and OU's hopes in 2009.
Thoughts?