http://insider.espn.go.com/college-football/story/_/id/9341117/usc-trojans-teams-were-better-their-records-showed-last-season-college-football

Trying to determine which college football teams are better than others has long been a fruitless exercise, with teams that are seemingly equal having such a small sample of games and sometimes no common opponents. But computers are allowing us to connect those dots like never before, and advanced metrics have been developed that can dig into game data and better analyze team performance -- which can be a useful tool as we evaluate teams for next season.

Expected Points Added is a metric that surveys the game situation to measure how much a given play or drive affects a team's likelihood of either scoring or allowing points. ESPN's Production Analytics team has used EPA with the NFL for a couple of seasons, and this is how it works. This season, we'll be unveiling college football EPA.

Our college EPA operates in the same framework as the NFL, except for one significant addition -- an adjustment must be made for opponent strength, since it varies so greatly across college schedules. Offensive performance is graded against the quality of the opposing defenses, defensive performance is graded against the quality of the opposing offenses, and even the special teams are graded against the quality of the opposing units (kick return versus kick coverage, etc.).

The metric was created for the purpose of evaluating those three phases of the game, but you can get a fairly accurate representation of team strength by adding the adjusted EPA for all phases. Looking at those numbers from last season, we see that Alabama was the best team and Oregon was second best. No surprises there.

But as you're looking for this season's sleeper teams, knowing which ones might have been better than public perception last season is a good place to start. Here are five teams that played better than their records showed in 2012 -- and might not need to improve all that much to be a factor in their conference races in 2013.



USC Trojans
The Trojans didn't come close to meeting expectations a year ago, and it felt even worse because of a preseason No. 1 ranking that had set the bar as high as it could go. The star-studded offensive cast of Matt Barkley, Robert Woods, Marqise Lee and Silas Redd rarely dominated the competition, but the numbers suggest that USC's overall performance was that of a top-25 team, even though the Trojans received no votes in the final polls because of their 7-6 record.

To look at USC in this light, you must focus on the schedule. While the average fan might scoff at the suggestion that Pac-12 teams play defense, the truth is that the Trojans played several well-above-average defensive teams in 2012, and all but one of those games came in conference play. Three USC opponents -- Stanford, Notre Dame and Oregon -- ranked among the nation's top 10 in adjusted defensive EPA, and the Trojans' offense performed extremely well against the Ducks and moved the ball relatively well against the Irish.

Therefore, the opponent adjustment makes the 2012 USC offense look a lot better than it seemed on the field, although it still wasn't close to matching the preseason hype. When you consider that Barkley missed the final two games due to injury, including the offense's dismal showing in the Sun Bowl, there was at least an excuse for the late-season swoon.

There's no trophy for being the best six-loss team in college football, but maybe USC can take consolation in being better than many four- and five-loss teams as well. And it's a positive sign that the Trojans can bounce back in 2013.



Oklahoma State Cowboys
The Cowboys weren't ranked in the final polls last year, but they had the adjusted EPA of a top-10 team. When OSU won, it usually did so in dominant fashion (seven of eight wins were by 21 points or more), and all its losses were against quality teams (three by seven points or fewer). While the 84-0 win over Savannah State did skew the numbers a little, it was barely the team's most impressive win of the season after the opponent adjustment (just ahead of a 65-24 win over a Louisiana-Lafayette team that finished 9-4).

On offense, the Cowboys were their typically dynamic selves, not only producing high national rankings in the traditional statistical categories but also having EPA numbers that backed it up. The defense, however, is what made Oklahoma State a much better team than most people recognized. It's easy to look at the scores put up by several of OSU's opponents last season and not take the defense seriously, but EPA is looking at more than the scoreboard.

Consider these Football Bowl Subdivision rankings for the 2012 Cowboys: 99th in points allowed off turnovers, 111th in kickoff coverage, 114th in punt coverage and 119th in non-offensive TDs allowed. That's out of 120 teams.

The EPA calculation reduces the responsibility of the defense for allowing points after it's forced to protect a short field because of a turnover or a long kick return. And when the opponent scores a touchdown against the offense or special teams, defensive EPA is not affected at all. Add to this the obligatory adjustment for the quality of most Big 12 offenses and EPA suggests that the Cowboys had a fringe top-20 defense last year.

When you pair the Oklahoma State offense with a defense of that caliber, it doesn't seem far-fetched to think of OSU as a team that often played like one of the best in the country. If the Cowboys become even average in some of the categories mentioned above this season, they could significantly improve their record.



Wisconsin Badgers
You don't have to convince anyone in Nebraska that Wisconsin was much better than its 8-6 record might indicate. While there's no question that the Badgers struggled to find their identity in the first half of the season, they never lost a game by more than seven points, and half of their losses came in overtime. That, in itself, suggests that Wisconsin was a good team that had some bad luck.

Adjusted EPA shows Wisconsin to be one of the top 25 teams of 2012, and anyone who watched the Badgers' final five games would have difficulty disputing that. What is a little surprising is to see Wisconsin with an adjusted offensive EPA that ranked in the top 30 nationally. The Badgers, using three starting quarterbacks, posed little threat in the passing game and, as a result, were largely ineffective at converting on third down. But what they did extremely well -- a staple of their offenses in recent years -- was not turn the ball over.

EPA recognizes that a turnover does two things. It reduces the percentage chance of the offense scoring on that particular drive to zero, and it usually increases the odds that the opponent will soon score. A conservative offense that protects the ball may not be fun to watch, but it typically gives the team a chance to win. Wisconsin had that opportunity in the fourth quarter of all 14 games last season. It will be interesting to see what happens this season if the Badgers have a few more bounces go their way.



Baylor Bears
Losing a talent like Robert Griffin III figured to be a huge blow to a program without a history of competing at a high level, so when the Bears opened conference play last season with a 1-5 record, it may have seemed like Baylor's 15 minutes were up. Then the team got hot and won its final four games -- three against ranked teams. It wasn't quite enough to get the Bears into the final polls with an 8-5 record, but adjusted EPA recognized Baylor as a legitimate top-20 team.

Even without Griffin, the offense was the reason why. The Bears weren't merely good on that side of the ball; they were arguably the best offense in the nation, according to EPA. (It was close between Baylor and Texas A&M.) And Baylor accomplished it with balance, ranking top five in the country in adjusted EPA on designed running plays and adjusted EPA on called pass plays.

Even though Baylor had, by most measurements, one of the weaker defenses in the AQ conferences, it closed strong on that side of the ball. The Bears' three best performances of the season in adjusted defensive EPA came in the final four games -- wins against UCLA, Kansas State and Oklahoma State. While the Bears didn't hold any of those opponents under 24 points, they improved their EPA by contributing to the team effort with a surge of sacks and turnovers forced.

With an unstoppable offense and a defense that was making momentum-swinging plays, Baylor had the feel of the 2011 Oklahoma State team as the season ended. From Nov. 1 through the bowls, only Texas A&M had a higher adjusted net EPA than Baylor.



Oregon State Beavers
The only member of this list that was actually ranked in the final polls of 2012, Oregon State was considered by the voters to be about the 20th-best team in the country. Adjusted EPA, however, says the Beavers had an argument for being a top-10 team.

There was nothing flashy about this group from an EPA standpoint. After opponent adjustment, the OSU offense and defense both ranked top 20 in the nation for the season, and the special-teams unit was top 30. The Beavers had the best single-game offensive EPA performance of the year against the BYU defense and the best single-game defensive EPA performance of the year against the Wisconsin offense.

The one thing that stood out in the EPA numbers for Oregon State is that the team was consistently solid but rarely dominant. Typically, when one side of the ball played well, the other side was fairly average. That explains why five of the Beavers' 12 games against FBS competition were decided by four points or fewer. If OSU had found a way to win more than half of those games, it might have had a final poll ranking in line with its EPA numbers.

On Wednesday, I'll switch gears and look at five teams from 2012 that adjusted EPA says were not as good as public perception.

RECOMMEND3TWEET39COMMENTS19EMAILPRINTSUBSCRIBE

Brad Edwards
ESPN Insider
FollowArchive
• Analyzes college football and the BCS as part of ESPN's Stats & Information Group
• Analyst for both College GameDay on ESPN Radio and the ESPN College Football app