|
|
| A STATISTICAL APPROACH TO DANCING
|
|
March 12, 2007
|
|
Jonathan, SCS.com Editor
|
Contact
Jonathan
|
There are many methods you can employ when filling out your NCAA bracket. Many of us who keep up with all 300-plus teams throughout the season sort of have a "feel" for each one by this point in the year and could already tell you who we liked as sleeper picks before the bracket was even released.
There are others who don't pay attention all season long but pick just based on the reputation of certain teams, and that's often not a bad way to go about it.
Then there's the lady at the office who hasn't watched a game since last March (if she even watched then) but picks a stunningly-successful bracket just because she liked so-and-so's uniforms or thought so-and-so had a cuter mascot. Fascinating!
In any case, no matter how ridiculous, everyone has their own method. Among those techniques used can be a more statistical approach, looking at things from a historical perspective. Back about five years ago when this site first started, I did a statistical analysis on data from 1990 up until 2001, analyzing which upsets occur most frequently, how many upsets occur each year on average, which seeds are most likely to make a Final Four appearance, and which seed most often wins the whole darn thing.
By popular demand, those stats have been updated through last March and will be distributed below. Keep in mind that these stats go back just 17 years, to 1990, but in my opinion, that's a pretty good timeframe to look at. It goes back far enough to observe trends but not so far that the game itself has changed dramatically since then.
In an effort to keep those sitting on the edge of their seat from falling on the floor, here it is:
FIRST ROUND UPSETS
| | | |
| 1st Round Upsets per Year |
| 1990 |
7 |
1998 |
9 |
| 1991 |
9 |
1999 |
12 |
| 1992 |
8 |
2000 |
3 |
| 1993 |
6 |
2001 |
13 |
| 1994 |
9 |
2002 |
7 |
| 1995 |
8 |
2003 |
8 |
| 1996 |
9 |
2004 |
4 |
| 1997 |
7 |
2005 |
8 |
| |
|
2006 |
9 |
| Average 1st Round Upsets per Year: 8.0 |
|
This is where the excitment is, where March Madness makes its money. Everyone loves the first round upsets, so (almost) everyone naturally picks at least a few bracket-busters when making their annual selections.
But how many should you pick? Glad you asked. As the chart to the right shows, there have been an average of exactly 8.0 upsets per season in the first round since 1990. Keep in mind that there are 32 first round games each season, so in other words, exactly 25% of the first round battles each March have ended in an upset over the course of the past 17 years. That number has not gone higher than 13 nor lower than 3 in a single year since 1990.
 If you're more of a picture person, perhaps the chart to the left will help you spot a trend. As that shows, the three-year period from 1999 to 2001 was odd in terms of the number of upsets, but the pace seems to have leveled off back to a more normal number over the past four seasons.
BUT WHICH FIRST ROUND UPSETS?
| | | |
| Average # of 1st Round Upsets per Year |
| 16 over 1 |
0.00 |
12 over 5 |
1.47 |
| 15 over 2 |
0.24 |
11 over 6 |
1.06 |
| 14 over 3 |
0.59 |
10 over 7 |
1.76 |
| 13 over 4 |
0.76 |
9 over 8 |
2.06 |
| Average Total 1st Round Upsets per Year: 8.0 |
|
Again, glad you asked. That's where the chart to the right will be very helpful. First of all, I don't recommend picking a 16 over a 1. Why? Because it's never happened before. Not since 1990, and not before that either.
And did you know that since 1990, the 9 seeds have actually beaten the 8 seeds more often than not? True. A 9 seed has won 35 of the 68 games between the two over the past 17 years.
The next most common upset is naturally the 10 over the 7, which happens almost half the time (30 of 68 times since 1990). But then here's another interesting tidbit: A 12 actually defeats a 5 far more often than an 11 knocks off a 6 (25 times compared to just 18 since 1990).
So the data shows you should pick a total of about eight first round upsets, and according to the numbers above, you should pick them in this manner:
2 instances of a 9 over 8
2 instances of a 10 over 7
1 or 2 instances of a 12 over 5
1 instance of a 6 over 11
0 or 1 instances of a 13 over 4
0 or 1 instances of a combo of 14 over 13 or 15 over 2
0 instances of 16 over 1
WHO'S GOING TO ATLANTA?
| | | |
| Average # of Each Seed in the Final Four per Year |
| 1-Seed |
1.60 |
4-Seed |
0.53 |
| 2-Seed |
0.81 |
5-Seed |
0.24 |
| 3-Seed |
0.62 |
6 or lower... |
0.25 |
|
Since 1990, at least one 1-seed has made the Final Four every single season except for last spring. In eight of those 17 seasons, two or more number one seeds have made the Final Four. So what does that tell us? Be sure to have at least one, and possibly two or three, top seeds headed to Atlanta, but don't ever put all four number ones in the Final Four. All four number one seeds have never gone to the Final Four in the same season, and in fact, three number ones have gone to the Final Four just three times over the past 17 years.
The rest of the Final Four is usually made up of 2-, 3-, and 4-seeds, while a five or lower occasionally makes a trip as well.
AND WHO WILL WIN IT WHEN THEY GET THERE?
| | | |
| # of National Titles per Seed Since 1990 |
| 1-Seed |
11 |
4-Seed |
1 |
| 2-Seed |
3 |
5 or lower... |
0 |
| 3-Seed |
2 |
|
|
|
A one-seed has made the national championship game 12 of the past 17 seasons. Rarely, though, do two 1-seeds play each other in the final. In 9 of the 12 years a 1-seed has appeared, that team has played a non-one seed for the title.
It's worth noting, however, that a one-seed has won the Big Dance in 11 of the past 17 years. Who won it the five other years? Three 2-seeds, two 3-seeds, and a 4-seed. No team seeded five or worse has won the national title over the past seventeen years.
USE IT WISELY
Maybe all of that info will help your bracket some. Then again, maybe it won't. In either case, it will be interesting to see which of these trends continue and which are strayed from once the dancing starts on Thursday.
And oh by the way, if you're the guy who works in the office with that lady who likes the team with the cute uniforms, we feel for you. She'll probably find a way to beat you again. Even if you use these stats.
|
|
|
