Oklahoma State No. 1 in the Computers? How Did Things Get so Messed Up?

Raise your hand if you think that Oklahoma State is the best team in the country. 

Put your hand down if you're an Oklahoma State fan. 

Now put your hand down if you're a computer.

Oh, wait, computers don't have hands. They don't have opinions, either, which is why some people think that their Oklahoma State No. 1 ranking in them is evidence that they are in fact the best team in the country. 

Stripped of the perceived "SEC bias," are the computers giving the most objective rankings? Or are they simply spitting out what they are programmed to? 

Most humans are not giving the Cowboys the same respect that the computers are. LSU and Alabama are ranked Nos. 1 and 2 in the human polls respectively. Not one human being in either the AP or the USA Today poll have Oklahoma State ranked first. 

Why the disparity between the human and computers? Who has it right? 

Let's take a look at Oklahoma State. It is a very good football team. It has a tremendous offense, arguably the most explosive in the nation. In fact, it's probably not generous enough to say "arguably" the best in the nation as there may not be anyone else in the argument. Perhaps Houston could be in the discussion, but Houston's numbers have come against a far weaker schedule. 

The Cowboys average 555 yards per game, ranking third in total offense. They score just shy of 50 points per game, at 49.9, which ranks second in the nation. Again, they are a very good team, but are they the best team in the country?

Why are the computers saying they are? The bulk of the reason is that according to the computers they have played the toughest schedule of all the undefeated teams, and that's why they are ranked first. Whether they are the "best" team is one discussion, but the toughest schedule? That's a problem.

In according to the computers, they have done so against the sixth-toughest schedule in college football. Oklahoma State's schedule is the sixth toughest in the nation? That's something worth taking a look at.

There are three reasons the computers are giving up a false impression of Oklahoma State's schedule. First, they have substituted yardage for scoring because they've been disallowed from using scoring. Second, they are giving the Big 12 too much credit. Third, there are some major issues with how they quantify strength of schedule.

The Yardage Problem 

Part of the problem begins with the BCS mandate to not include scoring in the computer rankings. Ostensibly, this has resulted in many of the computer rankings opening up a fallacious back door. 

The problem is that you can't equate a one-point win with a 30-point win if you're trying to do some sort of objective analysis. To get around the point boycott, most of the polls substitute yardage for points. This can have a very specious side effect. 

First of all, it can "reward" teams that have a tendency to turn over the ball and, at the same time, punish those teams that don't turnover the ball. Second, it rewards offensive-based teams more so than defensive-based teams. 

For example, take the LSU at West Virginia game. West Virginia out-gained LSU, 533 yards to 366. LSU won the game in a blowout 47-21. However, when the computers look at the game, it looks to them like a game that LSU "stole" more than a decisive win for the Tigers.

Oklahoma State, comparatively, out-gained the Texas A&M Aggies by 13 yards in their narrow, one-point win. The computers, therefore, give the Cowboys more credit for their one-point win than they give LSU for its 26-point win.

LSU is the most efficient offense in the country, averaging just 9.47 yards per point scored. Oklahoma State is 11th, needing 11.12 yards for every point it's scored. Certainly, that's still very impressive, but it is more favorable for OSU to compare yards than points. 

It works out the same way defensively. Oklahoma State's opponents need only 17.19 yards for every point scored compared to 19.37 for LSU's opponents and 26.15 for Alabama's. In terms of net yards per point, Alabama is first in the country at 14.54 (meaning its opponents need to average 14.54 more yards for every point they score), Temple is second with 12.95, and LSU is third with 12.40. OSU is 11th with 6.08.

This isn't an attempt to compare how "good" the teams are, but to represent that Oklahoma State benefits from the substitution of yards for points in the computer rankings.

It also needs to be taken into consideration that the computers aren't accurately representing the relative value of offense and defense. A team that gives up 400 yards but gains 500 yards is not as accurately represented as a team that gains 300 yards, but gives up only 200 yards. In some ways, this can be qualified as a logical problem just based on the adage "defense wins championships."

However, there's more to it than that. There is an underlying logic. If a team is limiting the other team's yards, generally speaking, it is going to play with a shorter field, which means it is going to need fewer yards to score.

Furthermore, a defensive-oriented team is going to generate more turnovers, which will have the effect of, again, generally shortening the field. If you look at it as a ratio, Alabama out-gains its opponents by 253 percent, best in the country; LSU out-gains its by 148 percent, 10th best. OSU is at only 128 percent, 38th best.

Again, this isn't an attempt to "rank" the opponents, but to show that yardage is a poor substitute for scoring. 

In fact, in the Sagarin Rankings, where there is an alternate ranking based on scoring, the Cowboys fall to sixth best. 

Overrating the Big 12?

Part of the problem is the computers inflate the Big 12 as a whole. They have the Big 12 ranked as the top conference. It isn't. It seems there is a collective push to make the Big 12 and its schedule better than it is. Looking more closely at the schedules, though, it's hard to fathom why. 

Simmer down, Big 12 fans. Read the reasoning before waging your shock-and-awe campaign against me.  

The computers primarily weigh three things—yardage, records and strength of schedule—to come up with the best teams. They need to take out-of-conference records into consideration, though. 

Obviously, any time there is an in-conference game, there is going to be a win and a loss measured against a loss. Therefore, a 5-5 record in the SEC isn't the same as a 5-5 record in Conference USA. When you put together five or six top teams in the same conference, it becomes automatic that at least one very good team might have as many as four or five losses. 

Therefore, the computers need to account for that logic by first ranking the conferences before ranking the teams individually. Pretty much they use the same criteria. As a conference, the Big 12 has the best record in college football.

As a result, all the teams within the conference are individually higher rated, and then, because those teams are higher rated, the individual schedules are considered harder, which in turn makes the individual teams ranked higher. 

According to the Sagarin Rankings, among top 25 teams, the four toughest schedules all belong to Big 12 teams. Missouri has the toughest schedule in the nation. This "toughest schedule" has come against Big 12 teams. 

Why is Missouri's schedule so tough? It's played Oklahoma, Oklahoma State, Texas A&M and Kansas State. All of its signature games, win or loss, come against Big 12 teams. Go through all the Big 12 schedules, and you'll see the same thing. The Big 12 schedules are "tough" because of their Big 12 schedule. 

The picture starts to get clear. Is it that the "blind" computer polls are ascribing too much credit to the Big 12 has a whole, or are the human polls not giving the Big 12 enough credit? How much stock should be put in the Big 12's out-of-conference schedule?

This is really where the crux of the entire debate comes down. If the computers are right about the Big 12, they are right about Oklahoma State. What is the basis for giving all the credit to the Big 12? How much stock can be up in their having the best OOC record in NCAAF?

If you average out the defenses they've faced, they have given up an average of 420 yards per game. Only one defense, Texas, gives up fewer than 300 yards per game, and only two others, Missouri and Louisiana Lafayette, give up less than 400—and both of those are over 350. 

Just as with the yards on the individual level, the same goes for the conference as a whole. It's a conference stacked with offenses and less-than-stellar defenses. 

While OSU has certainly had an explosive offense, it hasn't exactly been hammering the ball down the throats of the nation's elite defenses. The average Big 12 game sees 937 yards and 72 points. How much of that is good offense and how much is good defense? 

Is the Big 12 racking up massive offensive numbers as a whole, "tricking" the polls into making the computers believe that it's better than it really is? 

This is not a unique situation. The Big 12 teams, year after year, put up huge numbers offensively, and year after year, get beaten in their BCS bowl games.

Since the advent of the BCS, the Big 12 is 7-11 on college football's elite bowl stage. Only the ACC has a lower winning percentage among the BCS conferences. 

In fact, the Big 12, over the last five years, has only one win against a top-five team in the BCS and only one win against a higher ranked team in the BCS. Both of these were accomplished in the same game, Kansas' win over No. 3 Virginia Tech in the 2008 Orange Bowl. The only other win against a top 10 BCS conference team was the same year with Texas' win over No. 10 Ohio State. 

The Big 12 hasn't beaten the SEC in a BCS bowl game since Nebraska beat Tennessee in the 2000 Fiesta Bowl.

In other words, the Big 12 hasn't even had a signature out-of-conference win in three seasons. It hasn't had a signature win over an SEC school in a decade. It's collected its best wins against mid-major schools and the weaker BCS conferences.

There are going to be those who want to prop up Oklahoma's win over Florida State this year. If that's your best argument, you are proving my point. 

I'm not impressed with Oklahoma's 48-20 beat down of unranked Connecticut in last year's Fiesta Bowl as a signature win for an entire conference, either.

There are those who are going to want to argue you can't use last year to argue what is happening this year. They argued the same thing last year, and the year before that and the year before that. Then, they watched their teams either get whacked in their respective bowl games or play against weak non-BCS opponents or against the weakest BCS opponents. 

Misrepresenting Strength of Schedule

If there were something to this season to suggest that there's a reason to believe that this year is different, it would be a valid argument that you can't use history. However, there isn't anything different this year.

Where is the out-of-conference signature win? Oklahoma State's "best" out of conference win is against Tulsa at home. 

Furthermore, the Cowboys have not beaten a one-loss team this year, in or out of conference. The only team they have beaten that is presently in the BCS Top 25 is No. 22 Texas. 

Contrast that with LSU's schedule, which has beaten three presently ranked teams—Oregon, West Virginia and Auburn—all of which have beaten presently ranked opponents. Two of those teams played for the national championship last year. 

There are only two teams Oklahoma State has beaten that were ranked on game day, Texas and Texas A&M, and the latter has fallen from the rankings since.

LSU has played five, including potentially the winner of two BCS conferences in West Virginia and Oregon. 

Yet to the computers, Oklahoma State has played the tougher schedule. Really? Really, really? How on earth can this be?

LSU has beaten two of the presently ranked Top 25 teams in the BCS poll out of conference. The Tigers beat the Oregon Ducks, the third-highest ranked one-loss team, on a neutral field and they beat 24th-ranked West Virginia on the road. They won those two games by a total of 39 points. 

The Big 12, as an entire conference, has played one out of conference game against a team that is currently ranked in the Top 25 in the BCS and it lost! And, it was to an SEC team. Arkansas beat Texas A&M. 

The Big 12 has the best winning percentage against out of conference teams, but whom those wins come against matter. Effectively, what the conference has done is put together a number of wins against above-average mid-major schools, which has the effect of artificially inflating its strength of schedule.

The conference's "best" out-of-conference wins include wins over the likes of Louisiana-Lafayette, BYU, two against Tulsa, TCU and Nevada. The best wins they have against BCS conference teams are against the likes of now-unranked Florida State and Miami. 

While the Big 12 was feasting on marginal competition outside of the Big 12, SEC teams that were in contention were playing elite teams. LSU had the aforementioned games against Oregon and West Virginia. Alabama played Penn State. Auburn played Clemson. Georgia played Boise State. Tennessee played Cincinnati. Finally, as mentioned earlier, Arkansas beat Texas A&M. 

Did the SEC teams win all of those games? No, but they did win all but three of them. Three of the four wins came away from home, and by the top two teams in the conference.

As a conference, the SEC has a winning record against currently ranked, out-of-conference teams, and that's something the Big 12 cannot claim.

So, how does this translate into a "better" schedule for Oklahoma State?

In large part, it's because the computers just average out the rankings of the schedules, so it has the effect of making one weak team make the entire schedule look much worse than it is. For instance, compare LSU's out-of-conference schedule with OSU's. 

Based on the Massey Rankings, LSU has played No. 6 Oregon, No. 161 Northwestern Louisiana and No. 40 West Virginia. That's an average out-of-conference ranking of 69. OSU has played No. 85 Louisiana-Lafayette, No. 33 Tulsa and  No. 53 Arizona. 

However, it's a trick of the math with LSU's schedule completely thrown out of whack by the Championship Division team on its schedule. While the Tigers have played two teams that are ranked in the Top 25 away from home, and OSU hasn't played anyone, the computers are tricked into thinking OSU's schedule is tougher. 

When it comes to making a case for the best team in the country, it's not the "average" opponent that matters, but the best opponents. 

The Big 12 as a whole is similar, feasting on above-average mid-majors like TCU and BYU, while the SEC has been beating teams like West Virginia, Oregon and Penn State, but diluting those wins with wins over teams like Northwestern State. 

The Big 12's record overall is 27-3 (90 percent) out of conference. The SEC's is 31-4 (89 percent) out of conference. The SEC has played seven OOC games against presently ranked opponents to the Big 12's one. 

When you compare the two conferences record against OOC ranked teams, the SEC is 4-3, a .571 winning percentage, to the Big 12's 0-1, a .000 winning percentage. When you compare the schedule against ranked unranked opponents, the SEC is 27-1, a .964 winning percentage to the Big 12's .931 winning percentage.

Yes, the SEC has a better record against ranked opponents and a better record against unranked opponents but the Big 12 has a better record and therefore has a "tougher schedule." When OSU's win at home over Louisiana-Lafayette gets more credit than LSU's win over Mississippi State on the road, there's something wrong with the rankings. 

In a topsy-turvy logic, the Big 12 schedule is tougher than the SEC's. By not playing significant out-of-conference games, it props up the conference's overall record out of conference. That creates the illusion that the teams in the conference are better than they are, which in turn makes it look like their in-conference schedule looks tougher than it really is. And that's how a weak schedule makes the computers think they have a tougher schedule. 

The bottom line is that neither the Big 12 or Oklahoma State are as good as the computers are telling us they are. While it may all seem very egalitarian and objective, the computers are giving a skewed representation of the relative strengths of the conferences and their opponents. 

The reason that there's an SEC bias is because there's been an SEC dominance for the better half the last decade. Over that span, the elite conference is 8-2 against the best from around the country in BCS bowls. I don't recall a loss by a top-five SEC team in the regular season, but I do recall a few wins—one of which came this year. 

Bias isn't necessarily a bad thing. I have a pro-democracy bias. I have a pro-literacy bias. I have all kinds of biases that are perfectly reasonable. It's normal for informed people to have a bias. In fact, it can be argued that's the whole point of being informed, to form our biases. 

And, as an informed fan of college football, it's hard to not have an SEC bias. It has been the best conference for years. There's been nothing to change that perception this year, as the best teams in the SEC have beaten some of the best teams from the Big 12, the Big Ten, the Pac-10/12 and the Big East. 

The Big 12 has been a conference that historically runs up big scores against marginal teams and then gets beat when facing the elite teams from the other elite conferences. There's been nothing to change that perception this year either. Wins against marginal teams like BYU and Tulsa aren't convincing me that anything is different. 

Until there's more than questionable evidence from "unbiased" computers to say otherwise, I'll stick to my bias.