Everyone knows that I am huge fan of statistics, especially when it comes to predicting the games. But that doesn’t mean that I can’t criticize the numbers from time to time. And the problem that I have is all of the best rankings have the same glitch – I have a suspicion is because an over-weighting of strength of schedule but I can’t exactly put my finger on it.
The team that was creating the outlier for me was West Virginia. But the last few games they have played have been strong. Maybe that will come for this team as well. But it is messing up my bracketology analysis.
That team, of course, is Arkansas. The NET has 19-12 Arkansas as the 18th ranked team in the country. Historically, this means that Arkansas is a lock to be in the tournament. I think that they probably should be considered a bubble team. The problem I have with Arkansas (and to the same extent, West Virginia) is that they are 8-10 in the SEC (for a tie for 9th in the standings).
Lets look closer at the SEC NET rankings of the teams ahead or tied with Arkansas in the standings.
2) Alabama (26-5, 16-2 – 1st) – 4th in AP
3) Tennessee (22-9, 11-7 – tied 4th) – 17th in AP
18) Arkansas (19-12, 8-10 – tied 9th)
20) Kentucky (21-10, 12-6 – 3rd) – 23rd in AP
23) Texas A&M (23-8, 15-3 – 2nd) – 18th in AP
32) Auburn (20-11, 10-8 – 7th)
46) Mississippi State (20-11, 8-10 – tied 9th)
48) Missouri (23-8, 11-7 – tied 4th) – 25th in AP
60) Florida (16-15, 9-9 – 8th)
82) Vanderbilt (18-13, 11-7 – tied 4th)
I understand that the SEC has an unbalanced schedule. Arkansas even was probably on the wrong end of that as they had to play Alabama, Texas A&M and Kentucky twice – which is a tough schedule. But still, Texas A&M finished 7 games ahead in the standings and is being ranked 5 positions worse than the Razorbacks. Missouri is 3 games ahead in the standings, ranked in the 25th in the AP poll, and yet is 30 spots behind in the rankings the committee uses to rank games.
This difference is so much that when Tennessee hosted the 9th place team in the SEC, it was a Quad 1 win. When Tennessee hosted the 4th place team in the SEC that is also ranked 25th in the polls, it was a Quad 2 loss. Let that sink in for a second. The big reason why Kansas is likely going to be a 1 seed is because of their 15 Quad 1 wins (more than anyone else in the country). And yet, when Tennessee hosted a team ranked in the top 25, not only is that considered a loss, it is actually considered a bad loss since it isn’t to another Quad 1 team.
Now, Tennessee doesn’t need to worry about this being a bad loss – they are ranked in the polls and somehow 3rd in the NET. But a team closer to the bubble like Auburn doesn’t get the same level of credit for beating a ranked team.
It might work out when you fully look at Arkansas’ schedule, that it makes sense. When you have to play on the road at Alabama, Baylor, Texas A&M, Kentucky, Tennessee, and Missouri, you are going to pile up some losses. But from first look, something feels off. And then if I can’t trust what games are Quad 1 (am I really supposed to believe that playing any road game in the Big 12 and playing at home against anyone in the Big 12 with a 9-9 conference record is considered a Quad 1 game, but playing Missouri at home is not).
I have to keep reminding myself that the committee doesn’t care about my opinion – the rankings are what they use. And they have their own personal eye tests that I have no clue about (otherwise, I could predict them better). But it doesn’t make me feel good.