As The Computers See The SEC

VandyMania's Howell Peiser explains his computer ratings system and gives us predictions for Tuesday and Wednesday's SEC games.

As The Computers See The SEC
for
1/24 & 1/25

Tuesday

Wednesday

L S U 77 Georgia 63
Tennessee 77 Mississippi St. 70

How These Ratings Are Figured

A couple of basketball fans have E-mailed me asking me to explain how these ratings are derived. First, let me say that these are not the PiRates that I computed during football season. I am simply taking 14 of the most accurate different computer ratings that are available to me and applying a weighting with the most accurate ratings getting the most weight.

I continually grade how each of these 14 ratings pick the Southeastern Conference games and rate them in order of most accurate to least accurate. Then, for both the mid-week and weekend games, I select the 10 most accurate of these 14 as of the previous two weeks of games. I take these 10 ratings and apply a weight to each one. Let's look at how these ratings determined tonight's Kentucky at Auburn game.

Kentucky at Auburn: In order, from most accurate to 10th most accurate, this is what these ratings say will happen

#1 Kentucky by 3      20% weight      Kentucky .6
#2 Kentucky by 2      20% weight      Kentucky .4
#3 Auburn by 1         10% weight       Auburn .1
#4 Kentucky by 5      10% weight      Kentucky .5
#5 Auburn by 2         10% weight      Auburn .2
#6 Kentucky by 3      10% weight      Kentucky .3
#7 Kentucky by 1       5% weight       Kentucky .05
#8 Tossup                 5% weight        ---------------
#9 Kentucky by 12     5% weight       Kentucky .6
#10 Auburn by 5        5% weight        Auburn .25

If you add up the 10 ratings, Kentucky averages to a 1.8 (rounded to two)-point favorite.

Here is where I apply my own personal beliefs. Once I know the computer spread, I place a predicted score based on statistics and probably pace of the game. I think the pace will be above-average but the shooting percentages will be below average. Thus, my pick is Kentucky 72 Auburn 70.

Vanderbilt, According To The Computers

The computer ratings paint a gloomy picture for the Commodores. Taking all 14 ratings and giving them equal weight, the Commodores are picked to finish 6-10 in the SEC and lose in the opening round of the SEC tournament to Arkansas for a 15-13 regular season record.

Taking just the top 10 ratings without extra weight applied, the Commodores are picked to finish 6-10 in the SEC, win their opener in the SEC tournament against Ole Miss, then lose in round number two to Florida for a 16-13 regular season record.

However, when you take those top10 ratings and apply the same weight as the one above in the Kentucky and Auburn example, the Commodores are picked to win just the Georgia game at home. That's right, when the most accurate computer ratings are given more weight, Vandy is picked to finish 4-12 and lose in the opening round of the SEC tournament to Alabama for a 13-15 finish. I must point out that the of the nine losses called for, six of them are by three points or less. When you factor in the small sample size, these six games are considered tossups.

That's how I approach the remainder of the season. Vanderbilt has six tossup games and one game where they are expected to win with ease. These seven games will determine whether the Commodores will go to the Big Dance, the Big Apple, or the Big Goose Egg. The seven big games are:

Home--Kentucky, LSU, and Tennessee
Away--Georgia, Alabama, South Carolina, and Ole Miss

My personal belief is that Vanderbilt will win three or four of these games plus the Georgia game in Nashville for either a 7-9 or 8-8 SEC finish; if the Commodores could upset Florida in Nashville, then they could possibly go 9-7.

At 8-8, two SEC tournament wins would be needed to get into the NCAA Tournament. This would more than likely mean the Commodores would have to beat Auburn or Mississippi State in the opening game and then upset LSU in the second round.