BYU vs TCU 2011: The long-haired assassin goes to Dallas

[TheBYUGuy]

Or turning the ball over.

That's why I strategically used the word "typically".

[pelagius]

Money line is out now: BYU +380 and TCU -480. So assuming markets are approximately efficient (within transaction costs) the odds of BYU winning are about 17-20%.

[jay santos]

Money line is out now: BYU +380 and TCU -480. So assuming markets are approximately efficient (within transaction costs) the odds of BYU winning are about 17-20%.

I've never disagreed more with a BYU line. I just don't see it.

[pelagius]

I've never disagreed more with a BYU line. I just don't see it.

I think BYU is hard to value at the moment; it certainly is hard to value using a statistical model. How different is BYU now with Riley as the quarterback? In any MOV based statistical model, the Utah game is going to be a influential observation. So, for example, in a pure MOV model, I get the following for BYU and TCU:

Rank  Team                    Rating   W   L
13    TCU                     89.657   5   2
61    Brigham Young           73.532   6   2

Pure MOV predicts a 16 point victory for TCU on a neutral field. Pure MOV models love TCU (it is one of the highest rated 2 loss teams using pur move ... along with Georgia and Arizona St.).

[pelagius]

Also, if you split BYU into two teams: Pre Utah St (Brigham Young) and Utah St on (Brigham Young 2), this is how they would be ranked by a pure MOV model (pure MOV is almost identical to Sagarin's predictor model):

Rank  Team                    Rating   W   L
13    TCU                     89.828   5   2
38    Brigham Young2          82.484   4   0
92    Brigham Young           65.098   2   2

Rating differential is still 7 points for TCU on a neutral field for Riley only BYU.

[jay santos]

Also, if you split BYU into two teams: Pre Utah St (Brigham Young) and Utah St on (Brigham Young 2), this is how they would be ranked by a pure MOV model (pure MOV is almost identical to Sagarin's predictor model):

Rank  Team                    Rating   W   L
13    TCU                     89.828   5   2
38    Brigham Young2          82.484   4   0
92    Brigham Young           65.098   2   2

Rating differential is still 7 points for TCU on a neutral field for Riley only BYU.

TCU's #33 here.

http://masseyratings.com/cf/compare.htm

I think that sounds about right. Very few models have them in the top 25.

[pelagius]

TCU's #33 here.

http://masseyratings.com/cf/compare.htm

I think that sounds about right. Very few models have them in the top 25.

That may be right; here is what happening though. Pure MOV models love TCU because their loses have been very close and their wins have been by large margins. Models that are W/L based or concave functions of MOV rate TCU lower because they have two loses and a relatively weak schedule. It just depends on what you believe is more likely to be correct. That link is just average across different types of functional forms from pure W/L to pure MOV (but probably giving more weight to W/L because of the requirements of the BCS).

At the very least BYU and TCU are both teams that are most likely to have noisy ratings from statistical models. They're both among the set of teams were the modelling assumptions have a huge effect on the estimated ratings.

[jay santos]

That may be right; here is what happening though. Pure MOV models love TCU because their loses have been very close and their wins have been by large margins. Models that are W/L based or concave functions of MOV rate TCU lower because they have two loses and a relatively weak schedule. It just depends on what you believe is more likely to be correct. That link is just average across different types of functional forms from pure W/L to pure MOV (but probably giving more weight to W/L because of the requirements of the BCS).

At the very least BYU and TCU are both teams that are most likely to have noisy ratings from statistical models. They're both among the set of teams were the modelling assumptions have a huge effect on the estimated ratings.

Makes sense Vegas lines would be tied to an MOV model, but I personally think the computer average at masseyratings is more accurate. Or at least equally accurate. I think this is a case of TCU being overvalued and BYU being undervalued (but that might be the homer in me coming out).

[pelagius]

Makes sense Vegas lines would be tied to an MOV model, but I personally think the computer average at masseyratings is more accurate. Or at least equally accurate. I think this is a case of TCU being overvalued and BYU being undervalued (but that might be the homer in me coming out).

That may very well right, or at the very least I think BYU and TCU are the types of teams that are the most likely to mispriced. First, they are hard to statistically model. Two, a cognitive bias is that people tend to overreact or overweight dramatic information. In a football setting, losing 54-10 to a team ranked in the 40-50s is dramatic news.

[Blueintheface]

I have a non-nerd question:

Any CUF get togethers on this one?

I guess if I don't know of any already I wasn't invited and therefore my question is nerdy afterall.

[TheBYUGuy]

I would love to go to this one, but I will be in training until late-ish on Friday afternoon and I just wouldn't have enough time to go through rush-hour traffic in Houston and Dallas to get there before sometime in the second quarter.

[UtahJazzFan88]

The BYU-TCU game sounds like it's going to be a barnburner IMO based off how both have been playing lately, but I'm going to a Grizzlies game on Friday night as I got free seats right along the glass, when is the rebroadcast on BYUtv?

[YOhio]

Prestige worldwide.

[jay santos]

Prestige worldwide.

I wonder if they show Fox Rocky Mountain in China?

[pellegrino]

I'm surprised at how many BYU fans are predicting a BYU loss in the prediction thread. Who has TCU played? Baylor? Air Force? SDSU? I mean seriously, we're not that great of a team, but we're playing better each week. TCU has put up big points against bad teams. I think we have a good shot at this one.