TSL Roundtable: The BCS

TSL Staff, TechSideline.com, on December 19, 2013
Share on your favorite social network:
Share on FacebookTweet about this on TwitterEmail to someoneGoogle+share on TumblrShare on Reddit


Now that we are in the final year of the BCS, what are your favorite and least favorite things about the BCS?

Tafkam Hokie: Ah, the much-maligned BCS.  At times it feels like I am in a significant minority (if not outright lonely) in the opinion that I liked the BCS.

Part of it may be that I remember the system before the BCS.  That was when it was downright difficult for #1 to play #2 in a bowl and you typically ended up with a very scattered postseason due to conference tie-ins.  Take 1990 as an example.  Go back and watch the halftime show from your 1990 VT-UVA game tape.  That was an era of bowl bids being handed out with three or four weeks left in the regular season.  In the end, #1 played #5 in a bowl, #2 played #19, and #3 played #4.  How do you figure a champion with that mess?  After that jumblemuck, it is hard to fathom we have anything to complain about now.

The BCS was intended to cure that by actually allowing #1 to play #2 and decide the championship on the field.

And in my opinion, it worked.  I don’t just mean in 1999 and 2005 when two (and only two) major-conference undefeated teams were left standing at the end of the season.  I even mean it worked in years like 2004 where there were more than two seemingly qualified teams to play for the title.  The BCS filtered through the available information, added a purely quantitative element to bring a sanity-check to the process, and gave us the two teams that had the overall best body of work.  Sure, there were years when #3 complained, but that is what #3 is going to do whenever only two teams get into the “playoff.”

Honestly, I can only think of two gripes I’ve had with the BCS system over the years.  One is how much they kept tweaking the formula early on every time they got a result someone wasn’t happy with.  Personally, I liked it better when there were separate components for strength of schedule and quality wins, and the human polls were only about half of the formula (instead of 2/3 like it is now).  One could actually look at the formula and discern what your team needed to do, and who you needed to root for to achieve your desired outcome.  Today, we are essentially left with “impress the voters” as the criteria for ending the season #1 or #2.

My other gripe is the system with only two playoff participants didn’t give any specific preference to conference champions.  I will certainly admit it is possible for the two best teams in the country to reside in the same division of the same conference, or for one of the two best teams to have tied for their conference championship yet lose out on a tiebreaker.  But when your playoff spots are THAT exclusive, I really think extreme preference needs to be given to those who win a conference championship.

I am actually kinda sad to see the BCS die and be replaced with the four team playoff.  Maybe as a Virginia Tech fan, it is difficult for me to put any faith in a selection committee after how we have typically been treated in March.  The BCS gave us a quantitative element combined with human polls where the final votes were public record.  In the impossible task of comparing 120+ football programs who only got to play 12 games, the system at least had some transparency and accountability.  But a closed-door selection committee is almost guaranteed to give us just as many, if not more, head-scratching results than the BCS ever did.  And now we will get to listen to #5, #6, and #7 complain instead of just #3.

So long BCS.  You were better than you will ever get credit for.

Upwind of uva: The BCS did a pretty good job at promoting a national title game above the back-door shenanigans where bowl selection was dictated by tie-ins and economics. It introduced some much-needed (if imperfect) objectivity to the process. It also paved the way for the playoff, which may be good or bad depending on your perspective. By “paved the way” I mean that I cannot imagine the powers that be leaping from the free-for-all bowl selection system that existed prior to the BCS, directly to a playoff. The first, necessary step was to change the status quo to establish something that valued legitimizing the national championship over existing bowl traditions (*cough cough* Rose Bowl *cough*). At first I disliked the idea of the playoff, but I’m more on board today, so that step was a good one in my world.

I have a couple of significant criticisms of the BCS. The constant tinkering created the perception that the outcomes were more gerrymandered than a Boss Tweed election. How many of the computer rankings had to be “adjusted” to reflect not the mathematical theory they were based on, but rather the criteria that the BCS wanted? The power conferences were still making back-room deals that ensured they got what they wanted. The evolution of the selection formula also led to over-emphasizing human polls and basically set up a built-in bias when one conference is perceived as dominant.

Take, for example the 2011 title and Alabama. I can accept that one of the best two teams in the country may not win their conference championship, but at the end of the day the results on the field should matter more than the “eye test”. Alabama playing for the national title after they couldn’t win their own division is a disgrace. They, by not winning their own freakin’ division, sat out the most difficult “round” of pre-title competition (conference championships), avoided a game in which the odds of losing are much higher, and thereby enabled the SEC to secure both spots in the title game. I believe you should win your conference, not to mention your division, if you want to compete for a national championship in football. Emphasizing conference championship as a criteria maintains the relevance of the regular season while maximizing the probability that the best teams make the playoff. So at least under the playoff scenario the hand-picked teams will be forced to win two games rather than just one, and any selection bias at least gets tested on the field. Where it matters.

Baltimore Hokie: What I liked about the BCS:

First of all let’s recall that the BCS did not pave the way for No. 1 vs. No. 2 matchups.  The main problem early on was the bowl tie-ins that prevented a true championship game. That was mitigated in large part by the Bowl Coalition and Bowl Alliance in prior years, using the regular opinion polls.  The BCS added the algorithm that selected the top two teams – which convinced the Pac 10 to join in.

In the beginning I liked the fact they had an algorithm which took into account a lot more than the opinion polls.  Computer polls are inherently ‘smarter’ than opinion polls as they take into account more factors and eliminate regional/professional bias.  I felt this would guarantee objectivity in team selection – you would get more ‘deserving’ teams.  The original algorithm rewarded winning and winning big, and I thought this was a good thing.  Back then losing was the worst thing you could do, even worse than playing a weak schedule (as our selection in ’99 proved out).  The original algorithm never put a major one-loss team ahead of a major undefeated team.

VT benefited HUGELY by the ’99 algorithm, as our margin of victory bonuses far outweighed our low strength of schedule.  Even then, we barely beat out a 1-loss Nebraska for the NC game.

What I disliked about the BCS:

In hindsight I’m not sure that something like the Bowl Alliance wasn’t all that was really needed.  The complexity of the BCS made team selection more systematic, but also more controversial.

Ultimately, I strongly disliked that the purpose seemed to evolve from ‘picking the two most deserving teams’ to ‘validating the opinion polls.’  Every time the BCS produced an outcome that didn’t align with the opinion polls, the powers-that-be decided the BCS needed an overhaul.  That in of itself ultimately killed it.

I had lost track of all the changes so I had to look them up.  As it turns out, the algorithm was significantly changed after the ’98, ’00, ’01, ’03, and ’04 seasons, each change (except ’98) triggered by a controversy over team selection (there was controversy in the ’06 BCS, but no changes were made).  During this period the BCS encountered several unique situations – such has having to choose between three 1-loss teams in 2003.  The BCS did what it was designed to do, but since it didn’t jibe with the opinion polls it needed ‘fixing.’  What these situations REALLY did was highlight the flaws in the opinion polls – like punishing teams for losing late in the season.  But the BCS was evolved to place MORE emphasis on opinion polls – and opinion polls can be influenced by campaigns, media bias, etc.

Perhaps the ‘refined’ algorithms were never tested properly with prior years’ data.  It was widely reported that the changes made after ’00 would have put Nebraska in the ’99 NC game over VT had they been in place at the time.  That alone should have raised a red flag.

I think the computer polls were the most contentious thing about the BCS because no one understood them.   It seems odd to me that the BCS had to ‘force’ computer poll owners to adjust their algorithms.  This more than anything proved to me that the BCS folks had no idea what criteria were important – they were simply ‘outsourcing’ that part of the process and hoping for the best.  If they knew what they really wanted they’d have no need of outsourcing – simply hire someone to create a custom BCS Computer Poll and be done with it.

Nova Hokie 95: I think the BCS served its purpose — it provided a much better way to provide for a national championship at the end of the season than was previously in place, and it provided a season’s worth of talk for college football. And really, that’s what it’s all about — increasing the Q rating. The more people talk about you, the more people talk about you. And the more people talk about you, the more money comes pouring in. BCS=publicity.

But at its core, I think the BCS was a positive step and quite honestly would have been fine if it’d continued in some form or another. As others have said, I disliked the constant tinkering of the process that made it seem like an attempt to curry favor or provide favoritism to one team/conference or another. I thought the earlier editions were just fine with the computer rankings and strength of schedule and margin of victory. And I was definitely not a fan of the over-reliance on the human polls — especially given the way they were run.

One change I would have made — and would have likely helped my final objection — was to make it mandatory for the human polls involved in the BCS to refrain from doing the polls until the first BCS ranking. I liked that the BCS didn’t provide a ranking until mid October; I would have greatly preferred to have the human polls involved do the same. I think that would have avoided some of the objections about “started the season too low” or “is only the highest one-loss team because of the initial rankings.” Put more meaning back into the human polls — give them some actual this-year knowledge prior to polling — and you have a stronger poll and thus a stronger BCS.

I’m not sure yet how this playoff is going to work out. I’m pretty sure I don’t want to see it expand past four, but I’m also pretty sure I don’t trust any committee to act without preconceived notions. And we’ve already seen threats of lawsuits against the BCS; can you imagine if Boise State or Utah is #4 in the polls at the end of the year and left out of the playoffs because of some committee decision? We’re back to the original issue that started pushing the playoffs. Fairness.

No process is going to be fair, quite bluntly, when there are so many variables in the sport — what conference, what OOC teams, conference championship games, preseason polls, etc. It’s impossible to compare a wide variety of one-loss teams. I’m not convinced that four people (or however many it is) will be able to do it better than a group of computers combined with human polls, but I guess we’re going to find out.

Will Stewart: What I liked about the BCS:

I like that it put an end to the discussions about who was “really” number one by setting up a game between the top two teams — or what it had decided were the top two teams. That eliminated a lot of pointless, stupid postseason arguments.

And I like that it served as a necessary step in the push towards a playoff. I think I’m on record in the past as saying that I preferred bowls to a playoff, but I’d like to give it a shot, because in the long run, it’s good for the sport of college football.  It increases the money flow into the sport, and the attention paid to the sport.

The cute little four-team “playoff” that they have set up starting next year is the next step towards what should be the ultimate goal: an eight-team playoff. And playoffs have their own downsides. Whatever the downsides of the NCAA basketball tournament are, the football playoff will have the same problems: committee bias, coaches losing their jobs because they don’t make the playoffs, fans getting ticked because they don’t make the playoffs, etc.

But on balance, a playoff is probably good for the sport, and the BCS was the first step in that direction.

One more thing: I really, really, REALLY liked that the BCS didn’t involve RPI calculations like the men’s basketball tournament. Because the RPI algorithm is a terrible, terrible algorithm.

What I didn’t like about the BCS:

A lot of little things.

I didn’t like that so much of the college football world, once the initial BCS standings were released in October, focused on who was going to finish #1 and #2. (I’m looking at you, ESPN.) You’ve heard me say it before: college football is a rich tapestry of teams, people, and narratives. To focus on the top five or so teams, to the detriment of every other team on the planet, was tiresome. I remember when VT cracked the rankings back in 1993, what a huge deal that was. Now, no one pays much attention to #20-25.  Sure, as I’ve said before, you’re included in the highlight packages, and you scroll across the bottom line, and those things are good, but that’s it; then the talking heads go right back to debating which of the top five teams will make it through unscathed and play for the championship.

I also really hated that they removed margin of victory from the computer calculations. That’s just stupid. Beating FAMU 7-6 is not the same as beating FAMU 70-0. The first team sucks, because they barely beat FAMU. The second team is pretty good, because that’s what good teams should do to FAMU.

“Oh, but including margin of victory encourages good teams to run up the score on bad teams!” Waaaaaah.  It’s football, not knitting. If you don’t like it, then include MOV, but cap it at 20 points, so that guys like Bobby Bowden realize that there’s no difference between beating FAMU 49-0 and 70-0, so they won’t keep throwing the bomb with two minutes left on the clock. Feel better?

But guess what? Margin of victory matters to the Harris and USA Today poll voters. So removing it from the computer calculations was just stupid and pointless, and in my eyes, emasculated the computer calculations.

And I spent WAY too much time ranting about that. Sorry, it’s a pet peeve, and Jeff Sagarin and Kennethy Massey agree. So there.

As others pointed out, the constant jiggering of the algorithm to get it to match the human opinion polls was also very annoying, and along with removing MOV from the computer algorithms, ultimately weakened what I considered to be an encouraging goal of the BCS: removing human bias from the calculation. The more they watered things down, the more pointless it seemed to have computers and calculations at all.

Follow Tech Sideline on Facebook and Twitter.

Share on your favorite social network:
Share on FacebookTweet about this on TwitterEmail to someoneGoogle+share on TumblrShare on Reddit