Tuesday, August 19, 2008

Recruiting vs. NCAA Tournament Success - The McDonald's All Americans

Time to answer the question: what do Ivory Latta, Kasha Terry, Alison Bales, Camille Little and Ann Strother all have in common?

Answer: all of them were McDonald's All-Americans.

Ann Strother: 2002
Ivory Latta: 2003
Camille Little: 2003
Alison Bales: 2003
Kasha Terry: 2003

Some background about the McDonald's All-American team. Since 1977, McDonald's - the guys who make the Big Mac - have been naming high school All-Star teams. Being like everyone else in organized basketball, they didn't even notice the girls until 2002, when the first girl's McDonald's All-American team was named.

Sports America makes a list of 1200 girls, which is pared down to 150 players. The 150 players are then reviewed by a selection committee that shrinks the roster down to its final size. An "East" and "West" team are chosen, but more for balance than for geographical correctness.

These McDonald's All-Americans are considered to be the cream of the crop that the country has to offer, and as far as I know, all are seniors. These kids are heavily recruited by college programs, and most of the kids already know where they're going by the time the next college season rolls around.

If you look at some of these lists, they look like a WNBA roster. The 2004 McD All-Americans were pretty impressive. This list had Candace Parker, Crystal Langhorne, Nikki Anosike, Sylvia Fowles, Essence Carson, Alexis Hornbuckle, Tasha Humphrey, Candace Wiggins, and a couple of others that are on rosters today.

Of course, this doesn't mean that every McDonald's All-American will be a WNBA All-Star. Look at the Atlanta Dream, which has five McDonald's All-Americans on the roster, but only one of WNBA All-Star quality.

Once I learned about the McDonald's All-Americans, I asked myself, "how much does having a McDonald's All-American on one's college roster have to do with tournament success?" If I were, say, Coach Joe Blow from Wassamatta U., and I could recruit two McDonald's All-Americans , what would be my chances of success in the NCAA Tournament if I could get that far?

I first drew up lists of all of the All-Americans from 2002 to the present 2008 class. I looked up their college records. Some McD All-Americans, like Lindsay Richards (2002), got two years before injuries ended their career (Richards played at Iowa in the 2002-03 and 2003-04 seasons.) Others, like Brooke Smith (2002) would transfer after one year. Richards played for Duke from 2002-03, then transferred to Stanford. NCAA rules require that students transferring sit out one year. (Smith concluded her career at Stanford in 2007.)

Other students were medical redshirts, due to blown out knees and the like. Others couldn't cut the mustard academically (Erica Brown in 2005) or questioned their commitment to basketball (Elena Delle Donne in 2008).

The first All-Americans that made it through four years were the 2002 class, and it wasn't until 2006 that a college team could have a McDonald's All-American on their roster for four years. This gave me three NCAA tournaments to review - the 2006, 2007 and 2008 tournaments.

I accounted for three cases:

a) teams that went to the tournament and had an All-American on the roster,
b) teams that went to the tournament and did not have an All-American on the roster - these teams were the majority of teams, from the mid-majors and automatic byes, and
c) teams that had an All-American on the roster, but did not go to the tournament. Texas Tech had Erin Grant (2002), Brooke Baughman (2003) and Darrice Griffin (2004) but still couldn't make the tournament in 2006.

I wanted to look at games played in the NCAA tournament as opposed to games won. Looking at games played made more sense to me, as games played seemed to apply to all Division I teams, whereas games won restricted the group (in my mind) only to teams that made the tournament. A team that lost in the first round has one game played, a team that didn't make the tournament has zero games played.

We then restrict to teams that have McDonald's All-Americans.

The next step is to perform what is called a linear regression. Linear regression answers the question: "if the relationship between All-Americans and games played is linear - if having x All-Americans means you get to play y games in the NCAA Tournament - then what is the relationship?"

In real life, the relationship is not strictly linear, but we can statistically draw a "best fitting line" that best approximates the data.

Here are the "best fitting lines" for the 2006, 2007 and 2008 tournaments.

Let "x" = # of McDonald's All-Americans on your team.
Let "y" = # of expected games played in the NCAA Tournament that year by your team.

2006: y = 0.630x - 0.237
2007: y = 0.705x - 0.004
2008: y = 1.116x - 0.910

Note that for each of the years, the slope of the line - the number next to x - increases year per year. For each successive year in the analysis, having a McDonald's All-American has more impact.

Why is this? My hypothesis is that the selection committee has gotten better over time. Remember that the 2006 tournament represents the very earliest McDonald's All-Americans whereas the 2008 tournament reflects a McDonald's All-American committee with more experience in picking players. 2008 represented the strongest correlation between roster and tournament success.

Here is the best fitting line representing three years of data:

all years: y = 0.798x - 0.337

If your team has zero All-Americans on it, you will play an "estimated" -0.337 games in the tournament. If you have two All-Americans, you will play an estimated 0.798 * 2 - 0.337 = 1.259 games. Two All-Americans on the roster might get you past the first round.

If this relationship is correct, or even close to correct, it explains why these girls are so highly recruited.
Note that

in 2006, you needed three McDonald's All-Americans to make the NCAA finals (Maryland)
in 2007, you needed five McDonald's All-Americans (Tennessee, Rutgers)
in 2008, you needed six McDonald's All-Americans (Tennessee, Stanford)

In all three years of the NCAA Tournament for which McDonald's All-Americans play a fact, only one team in each of those three years that made it to the Elite Eight had no All-Americans on it.

2006: Utah
2007: Mississippi
2008: Texas A&M

Neither Utah, Mississippi, or Texas A&M was able to make it to the final four. Three years isn't much of a sample, but recruiting success appears to be highly correlated with tournament success.

So which teams have the most McDonald's All-Americans on them now? Which teams are the ones to look out for in 2009?

As you can see, Tennessee and Pat Summitt have an amazing nine McD All-Americans on the roster:

Fuller, Cane, Baugh, Bjorklund, Gray, Johnson, Manning, Stricklen, Brewer

Up next is Rutgers with eight All-Americans:

Vaughn, Prince, Rushdan, Lee, Sykes, Dixon, Pope, Speed.

As a matter of fact, one could probably pass off the list above as a pre-season Top 25 list and the casual observer wouldn't give it a second glance. Except, of course, for the observers who live in Connecticut.

POSTSCRIPT: The third column of the graphic should indicate Tournament Games, not wins. Since I'm at work, I'll fix it when I get home.

1 comment:

Anonymous said...

I would be very interested to see the correlation between WNBA longevity and success and prior McD All American status...because my hunch would be that the correlation is low for the first few years but now going up as the quality of the HS game is improving (so that someone cannot just be a superstar among weak play anymore). Fascinating to think that Ann Strother was HS Player of the Year when she came out....and she may still develop into a solid WNBA star (she is getting a ton better this year in Atlanta)....but she is definitely not at the same level as more recent McD stars