Tuesday, January 12, 2016

Best Title Game Ever?

Note: Below is the original write-up from 2016.  Updates to the results are through the end of the four-team era.

Perhaps you caught last night's title game between Alabama and Clemson.  I sure did!  And because I am me, I immediately spent time after the game trying to determine if it was the best title game ever.  I did come up with a purely subjective list, but I figured that wouldn't be interesting enough to share with the world.  What would be interesting enough?  Why, something that's all math-y and such.

Before I get to my methodology, let's talk about what makes an ideal title game.  The most obvious trait of a good game is some degree of closeness.  It can be fun to watch a great team enforce its will, but only if your team isn't the one getting destroyed (I have some experience on both sides of that, so I understand arguments in both directions).  That said, closeness isn't everything.  If the teams involved in a close game aren't very good, things like this can happen.  Of course, title games generally don't involve bad teams, but there is still clearly a difference between 2001 Miami and some of the lesser title game participants.  For my simple model, I will weight these two factors equally.  Here is what went into those two factors:

Margin

The easiest way to measure the closeness of a game is obviously the final score.  Still, we can do better.  What I did was track down every title game box score* and determine the average scoring margin over the whole game.  This should give us a better measure of actual evenness by weeding out games where the losing team puts up points in garbage time.  To get my final number for margin, I took this number and divided it by the total points score in the game (as it's easier for a 10-7 game to be close throughout, so we need to adjust for that), and then found the z-score for that result so we can compare it to metrics on other scales.  Finally, I weighed this metric against the adjusted z-score of the actual final score by a ratio of 3:1.**  Final margin still has usefulness as it can indicate a thrilling finish, so I wanted to give that at least some consideration.

*Finding box scores from 1998-2002 was shockingly hard.  I mean, the internet existed then, right?  How did people track sports scores in the early days of the internet?  Newspapers?  Gross.

** Every weighting here is inherently subjective.  If you don't like it, well, I can send you my spreadsheet and you can change things around.

Team Strength

I measured team strength in two ways.  The first and more important way, raw strength, accounts for 2/3 of this aspect.  To do this, I simply added the SRS ratings of the two participants together and took the z-score (SRS is a basic rating system that accounts for score and opponents that can be used to compare teams across time...I would have used something more complex, but most of those systems don't go back to 1998).  If a particular set of teams in a given year is stronger than your average top-ranked teams, that will be accounted for here. 

The second aspect I measured was comparing the title game participants just to the teams within their season.  For this, I took the ordinal rank of each team in SRS for that year, and added those figures together.  I took another z-score, and this figure then accounted for the other 1/3 of team strength.  Even if a season didn't have any historically great teams, a title game that at least matched up the two actual best teams is a good thing that should be rewarded in these rankings.

The Results:

#26 - 2022: Georgia 65, TCU 7 (-2.44 rating, 25th in Margin, 22nd in Strength)

The bitter irony of the final two years of the four-team era is that we finally got amazing semi-final games, but at the expense of dreadfully boring title games.  Georgia breaking the bowl record for largest margin of victory made for bad TV, but hey, we'll always have the multi-year-spanning Peach Bowl that proceeded it.

#25 - 2007: LSU 38, Ohio State 24 (-2.44 rating, 15th in Margin, 26th in Strength)

Similar to the irony of the couple bad four-team era title games, the best season in recent memory produced the worst title game.  Surprisingly, the main reason for this was the relative weakness of the teams, rather than the margin (I felt that this was way more of a blowout than it actually was).  This Ohio State squad was actually the worst title game participant, per SRS, which isn't shocking considering that they lost to Illinois at home.  Ron Zook is a wizard.  A poor one, but a wizard nonetheless.

#24 - 2012: Alabama 42, Notre Dame 14 (-2.18 rating, 23rd in Margin, 21st in Strength)

Yeah, let's not talk about this one...

#23 - 2001: Miami 37, Nebraska 14 (-2.05 rating, 24th in Margin, 18th in Strength)

....or this one, either.

#22 - 2006: Florida 41, Ohio State 14 (-1.97 rating, 20th in Margin, 25th in Strength)

Fun fact: In spite of all the hype about Troy Smith and Ohio State, and Florida starting the SEC "dynasty," neither of these teams were in the top two of the SRS.  Rather the top two were USC and....Louisville.  So yes, this game is ranked this low in part because of Bobby Petrino's coaching acumen.

#21 - 2000: Oklahoma 13, Florida State 2 (-0.96 rating, 22nd in Margin, 11th in Strength)

An ugly game gets boosted up by a pretty strong Seminole team.  The adjustment of the margin measure to the score of the game makes it hard for such a low-scoring game to rank any higher than this.

#20 - 2023: Michigan 34, Washington 13 (-0.92 rating, 18th in Margin, 23rd in Strength)

The other recent dud that followed a couple of classics.  As much as I loved this Husky team, and as much as they deserved to be there, this was a relatively weak team that drags down the strength component of this rating system.

#19 - 2011: Alabama 21, LSU 0 (-0.91 rating, 26th in Margin, 5th in Strength)

Yes, the game that was so bad that it created a playoff ranks above nearly one-third of the title games.  How?  Well, Alabama and LSU were really freaking good.  I mean, allowing only 2.7 yards per play on defense is pretty darn special.  Sure, defensive prowess doesn't always make for the most attractive games on the surface, but historically awesome defense is still something to appreciate.

#18 - 2003: LSU 21, Oklahoma 14 (-0.57 rating, 17th in Margin, 17th in Strength)

A perfectly fine game takes a bit of a hit because the actual best team of the year (USC) was relegated to the Rose Bowl.  Yep, I went there.

#17 - 2002: Ohio State 31, Miami 24 (-0.47 rating, 9th in Margin, 24th in Strength)

So this is a bit of a head-scratcher, but it makes sense...trust me.  First, the game falls just outside of the top five in margin because of the relatively low score (just 34 points in regulation).  There have been a lot of other close games with a little more back and forth than this game, and those get credit for that.  Second, Jim Tressel's squad is the second worst title-winning squad per SRS, falling just behind the 2007 LSU team with two losses.  Perhaps the exciting ending should be rewarded more than it is, but I think I'll leave this here for now and move on.

#16 - 2004: USC 55, Oklahoma 19 (-0.38 rating, 21st in Margin, 7th in Strength)

SRS agrees with popular opinion that 2004 USC is one of the best teams of the BCS-Playoff era, which propels this blowout a little higher than it probably should be.  If you want to switch these last two games in your mind, I won't complain.

#15 - 1998: Tennessee 23, Florida State 16 (-0.05 rating, 8th in Margin, 19th in Strength)

The first title game was a perfectly fine affair, but Tennessee's relative weakness prevents this one from reaching the top echelon.

#14 - 1999: Florida State 46, Virginia Tech 29 (0.21 rating, 12th in Margin, 15th in Strength)

Virginia Tech outscored the Noles 15-0 in the third quarter and took the lead into the fourth...that I did not remember.  This is a great example of why looking at the average margin throughout the game better reflects how close a game was instead of just the final score.  Florida State may have put this away late, but it was a great game most of the way.

#13 - 2013: Florida State 34, Auburn 31 (0.31 rating, 6th in Margin, 20th in Strength)

Here's another game that I had in my preliminary top five that didn't make it as high in my formal methodology.  Auburn caught fire down the stretch in 2013, but their overall resume that year was merely quite good, and that's the main difference between this game and the few above it.

#12 - 2017: Alabama 26, Georgia 23 (0.42 rating, 8th in Margin, 16th in Strength)

Very funny that neither overtime game makes the top ten.  Part of this is because there's been a lot of really good, close games that did not go to overtime.  But the bigger part is that the overtime games just happened to feature "lesser" teams than most other games.  Pretty easy to make the case this was Saban's worst title winner while Kirby's subsequent champions were better than this Bulldog team, too.

#11 - 2009: Alabama 37, Texas 21 (0.54 rating, 14th in Margin, 10th in Strength)

Despite Colt McCoy's injury on the opening drive, this game remained close until the final moments.  That, combined with a typically dominant Crimson Tide makes for a higher quality game than you probably would have guessed.

#10 - 2010: Auburn 22, Oregon 19 (0.60 rating, 7th in Margin, 14th in Strength)

The battle for all the Tostitos was a weirdly defensive game between teams that were better than I thought (3rd and 2nd in SRS).  I guess Oregon was 10th in the more familiar (to me, at least) F/+, hence the surprise on my side.  This is still the only title game decided by a kicker, although Miami did hit a field goal at the end of regulation in 2002.

#9 - 2014: Ohio State 42, Oregon 20 (0.60 rating, 13th in Margin, 12th in Strength)

It seems weird to put this Oregon game above the other one until you remember that these were fairly clearly the best teams at the end of the season, and that this game was pretty close until midway through the fourth quarter.  The rich man's Michigan-Washington, if you will.

#8 - 2019: LSU 42, Clemson 25 (1.00 rating, 10th in Margin, 8th in Strength)

There was a lot of "greatest of all time" hype following the Tigers' romp through the Playoff.  And to be clear, they were very good, very fun, and defeated a lot of the best teams on their way to 15-0.  But.  They finished #2 in SRS (and in other ratings) and never got the chance to play/beat #1.  This is my most Columbus-coded opinion.

#7 - 2021: Georgia 33, Alabama 18 (1.16 rating, 5th in Margin, 9th in Strength)

This tops the other Tide-Bulldog matchup because a) these teams were better, and b) this was a little more back-and-forth than the other game.  Yeah, Tua's walk-off bomb was a lot of fun, but this one was more compelling throughout.

#6 - 2018: Clemson 44, Alabama 16 (1.25 rating, 19th in Margin, 1st in Strength)

Could make an argument that this laugher deserved to be even higher if only for the schadenfreude, but ultimately this feels about right.  Both of these squads are in the running for 2nd best team of the century, which is enough to overcome a late blowout.  Also, this doesn't factor in my rating, but it doesn't hurt that this was actually a "close" game on the fundamentals.

#5 - 2015: Alabama 45, Clemson 40 (1.28 rating, 1st in Margin, 13th in Strength)

So here's where the bias of the design of my margin metric comes into play.  When I was figuring out the best way to measure this, I most certainly had this game in my mind, which caused me to focus both on average margin, and anchoring it to the total score (this was the highest scoring title game, and it involved Alabama!)  All that said, I think 5th feels perfectly fine for what was an amazingly entertaining game between two well-matched teams.  SRS doesn't like Clemson enough to put this game any higher (their 18.9 rating is most similar to 2013 Auburn, which is sort of creepy if you think about it), but that works fine for me, as I really love the games that rank above this one.

#4 - 2016: Clemson 35, Alabama 31 (1.59 rating, 4th in Margin, 6th in Strength)

The sequel to 2015 was a bit better game between slightly better teams.  Incredibly sensible ranking.

#3 - 2020: Alabama 52, Ohio State 24 (1.65 rating, 16th in Margin, 2nd in Strength)

The best team of the century played a decent opponent in a high-scoring game that was close until right before the half.  That's a perfect formula for a game not popularly considered as great to sneak up these rankings.  And while I will always cherish Justin Fields' destruction of Clemson in the semi-final, I really wish that would have happened a year earlier instead.

#2 - 2008: Florida 24, Oklahoma 14 (2.35 rating, 3rd in Margin, 3rd in Strength)

It should be enough that a pair of Heisman winners were involved in a tight battle that only finally resolved in the closing minutes*.  But what pushes this game over the edge is the insane quality of the teams.  SRS places these teams 8th and 12th best among the 52 title game participants in this post, and they're not alone in singing their praises.  I feel this game gets overlooked when looking at classic championship showdowns, but it most definitely should not.

*The game was tied for almost 40 minutes.  That's how you overcome a relatively low-scoring game in my Margin metric.

#1 - 2005: Texas 41, USC 38 (2.39 rating, 2nd in Margin, 4th in Strength)

Of course this game is #1.  I probably would have thrown out the rating system and started over if it hadn't been.  Other games had slightly more back and forth, and the 2008 and 2018 games had slightly better teams (maybe).  But no game combined excitement and historic levels of talent quite like this one.