Wednesday, December 22, 2010

What's GUGS?

I wanted to take a quick break from our bowl previews and talk about the Games U Gotta See (GUGS) system.* This is our in-house measure of how interested a casual fan could or should be in a game based on the TFG and RBA projections. There are three elements of GUGS:

  1. Competitiveness (C). How close is the game expected to be? Blowouts are boring, so the closer it is to a coin toss, the higher the competitiveness components.
  2. Quality (Q). How good are the teams? The higher their expected winning percentages, the higher the quality factor.
  3. Scoring (S). Will this be a high-flying, high-scoring, who-needs-defense affair? Or a grind-em-out, one-inch-at-a-time defensive showcase? Let's put it this way: would you rather re-watch the 67-65 Michigan-Illinois game, or a 9-6 baseball Big XII game?

A perfect 100.0 -- which is actually not possible to get -- would involve two teams with 1.000 expected winning percentages projected to have a 60-59 final score. A perfect 0.0 -- also not possible -- would involve two teams with 0.000 expected winning percentages where one was juuuust good enough to blow the other out, 1-0, with a 100% chance of winning. In reality, the vast majority of GUGS scores fall between 20 and 60.

Let's look at three games from this past year and compare and contrast how GUGS looked at them versus how the media saw them.

Miami Hurricanes at Ohio State Buckeyes
September 11, 2010
GUGS Score: 47.4

The official game preview tried to make this sound competitive:
Second-ranked Ohio State will be a clear favorite when it faces No. 12 Miami for the first time since that double-overtime championship thriller, but the Hurricanes can pave their own path to BCS contention with a victory Saturday.
The problem is that this game wasn't going to be that close. Let's break down the GUGS components:
  • C: 33.0
  • Q: 86.1
  • S: 73.3
The teams were excellent, the scoring would be moderate (the Buckeyes boast an excellent defense), but this wasn't going to be that close. TFG was eerily accurate, predicting a final score of 35-24 and a 90.3% chance of Ohio State winning. RBA was close, too, predicting a 31-20 final and a 77.5% chance of an Ohio State victory.

Notre Dame Fighting Irish at Michigan State Spartans
September 18, 2010
GUGS Score: 57.7

The closing line for this one was the Spartans by 3, but it's unlikely Vegas thought it would end like this. There was some bitching and moaning on ESPN about whether or not the Spartans got the play off on time, but they've got no one to blame but themselves because the ESPN play clock wasn't actually synchronized to the stadium clock (compare the ESPN clock to the clock in the top-left). But let's look at what GUGS predicted:

  • C: 76.7
  • Q: 70.3
  • S: 78.7

Two good-but-not-great teams playing a close game with more scoring than the Ohio State/Miami game. Unless you're a Hurricane or Buckeye fan, it'd be hard to argue that you'd rather watch that game than what turned out to be a real nail-biter in East Lansing. The predictions said it would be close but disagreed about the outcome. TFG predicted Michigan State, 33-30 (60.1%) while RBA said Notre Dame, 28-27 (63.2%).

Oklahoma Sooners vs Nebraska Cornhuskers
December 4, 2010
GUGS Score: 71.1

Here we have a complete package. Two legitimately elite teams on a neutral field, one of whom plays fast and efficiently and will drag the other team along with it. Let's look at the details:

  • C: 80.3
  • Q: 88.5
  • S: 74.7

This game was projected to be closer than either of the above, involves two teams that are collectively just as good (if not better) than the Buckeye/Hurricane matchup, and will involve decent amounts of scoring. GUGS thought this was the 7th-best game of the year, and it lived up to the hype with the Sooners coming from behind to win the Big XII title in a 23-20 game. The predictions overestimated the scoring a bit, and we disagreed about the outcome, but knew it would be close. TFG said the Sooners, 32-31 (57.4%) while RBA said the Cornhuskers, 27-24 (62.3%).

Of the six games projected to be better, four have been played so far; all six involved either Oregon, Auburn, Stanford, or Virginia Tech. So keep watching our bowl previews to find out the best of the college postseason.

* We're open to ideas for new names or acronyms. Damn you, Pomeroy, for using FanMatch!