We like to think that basketball coaches are masters of the recruiting game, and that there is no inefficiency in recruiting - there is no Moneyball-like solution to the challenges that recruiting presents. In June of 2011, Drew Cannon of Basketball Prospectus wrote an article about recruiting and geography. The article was so good that I wanted to apply its conclusions to women's college basketball. Are there areas of the United States that are significantly under-recruited or over-recruited?
Assumptions of the model
In order to answer this question, Cannon had to make some basic assumptions.
1. Talent can differ by state - some states might not produce a lot of Division I players because the state does not have a good basketball culture, or because some other sport is more popular, or any number of other reasons. Cannon assumes that the number of "high major" players a state produces is a good estimate of the state's talent level. By "high major" we mean the traditional power conferences - the ACC, the Big 10, Big 12, Pac-12, Big East and SEC.
2. We have to deal with the issue of walk-ons. This can be handled by looking at the number of Division I schools in the state, each of which might have some local walk-ons on the roster.
3. Talent might be dependent on the state's population. This might clash with assumption #1 above, that each state has its own level of talent and can produces as many (or as few) D-I players as necessary. We will test this assumption in our regression.
Cannon then ran a multiple regression on these data points - basically, the idea is that if you know the number of D-I high major players from a state, the number of D-I schools from a state, and the state's population, then those three numbers should mathematically determine the fourth number which is the total number of D-I players that a state should produce - if the regression states that there is some relationship. We compare that number with the actual number of players produced by the state, and that number will tell us if a state is over-recruited or under-recruited.
Getting the data
All of the data was obtained from wbbstate.com. (If you haven't paid for a site membership, I highly recommend it.) The data, unfortunately, isn't in one place and this required 345 individual cut-and-pastes. Each team's roster lists the player's name and what is listed as her hometown.
The data, however, was not perfect. Some players were listed twice under slightly different names. Other players had to be looked up individually due to data formatting issues. Sometimes when you visited a college's website to verify data, you found players on the roster that were not on wbbstate's roster, or vice-versa.
As we say in the risk profession, I do not believe that any data errors are material errors and should not affect the results too much. There was a wealth of data - 4,387 players in total - which might be the source of another article or two.
Results of the regression
Once you run the regression you get a formula that looks like this:
Total Projected Players = 3.20*(Number of D-I High Major Players) + 4.43*(Number of D-I Schools in State) + (1/23,676,596)*(2012 Population of State) - 4.25
Note that fraction, which is so close to zero as makes no difference - the effect of state population is negligible.
We then compare total projected players to actual players to come up with our conclusions. So which states are the most under-recruited?
1. Alabama (67 total D-I players, 110.6 projected)
2. New York (152 total, 194.4 projected)
3. Iowa (46 total, 78.7 projected)
4. South Carolina (50 total, 81.0 projected)
5. Texas (370 total, 398.5 projected)
In the initial article, Cannon had some theories for why certain states were under-recruited. His regression was performed on men's college basketball, but the same conclusions likely apply to the women's game.
* New York: Cannon's theory was that when a college basketball coach comes to New York, "New York" means New York City - it's an area of high population density with a lot of basketball talent. Why go to New York and not go to New York City? The problem with this line of thinking is that you miss the rest of the state, with cities like Buffalo, Albany and Rochester. The players in upstate New York get overlooked.
* Alabama and South Carolina: Cannon thought that the exclusion of southern US states might be attributed to bad recruiting habits. For most basketball recruiters "the South" means states like Georgia, Florida, Virginia, North Carolina and Tennessee. The states that are less populous or which don't have huge population centers are overlooked.
You might argue that the number of D-I players in Alabama is somehow overstated - perhaps lazy recruiters from Alabama schools pick up a bunch of kids off their doorstep that don't really belong in D-I. There are nine D-I schools in Alabama and it is true that Alabama is in the top 10 of states where in-state D-I prospects end up in in-state D-I schools, at 43 percent. But it's just barely there, at #10. D-I players from North Carolina end up at in-state D-I schools at a rate of 57.6 percent, and D-I players from Louisiana end up in the Bayou State at a rate of 71.3 percent! (*)
The two hardest states to explain are Iowa and Texas. Cannon said about Texas, "Other states with lots of events (e.g., Texas, California) are so big they can't possibly draw everyone from the whole state." But Texas and California are on different ends of the scale in women's basketball, with Texas an under-recruited state and California an over-recruited one. I would have to know more about the kind of AAU/evaluation events that are held in California and Texas to draw a conclusion.
So what about Iowa? The lack of a major metropolis on the scale of Atlanta might cause recruiters to give it a miss. It makes no sense to me, since Iowa has a long history of women's basketball success at just about every level.
Now that we know the under-recruited states, what are the states that are the most over-recruited?
1. Maryland (152 total, 110.6 projected)
2. Ohio (213 total, 173.4 projected)
3. Tennessee (140 total, 120.7 projected)
4. Indiana (131 total, 111.9 projected)
5. Virginia (156 total, 141.3 projected)
Virginia and Tennessee might be explained by the fact that they are traditional recruiting stops. According to Cannon, Indiana-Illinois-Ohio is considered to be a "recruiting swing" so those states are very closely looked at by recruiters.
Maryland is very hard to explain. According to the model, it is the most over-recruited state in the country. I have my own theories, but nothing solid to base any of them on.
Arguments against the model
Of course, the weaknesses of the model have to be considered. If you think an under-recruited state is not under-recruited, you are arguing that either:
1. That as Cannon said, "the state is top-heavy". Its distribution of talent within the state either
a) drops drastically after the best players are considered, or
b) the top players in the state end up at D-I major schools but aren't D-I major players.
2. The schools don't have a lot of in-state walk-ons.
3. The state is producing more high-level talent than it usually does for some reason.
There are also arguments to be made about the number of quality mid-major schools a state might have. Some D-I major level players might not go to a D-I major if there's a good D-I mid-major around. Some D-I prospects might go to a D-II school if the state has very good D-II programs, which could skew the numbers. There's a lot to consider, and I'm sure a lot of arguments could be made for claiming some schools might not belong on either list due to unique circumstances depending on the nature of basketball in-state.
So if you're a D-I recruiter who works at the mid-major or low-major level, and you want to find gems that are overlooked - you might want to give Alabama or upstate New York a shot this summer.
Recently, former Texas Tech head coach Kristy Curry was hired at Alabama. She stated that "We want to secure the borders of this state and not let anybody out. We want to make sure that we take care of home first." But if it's true that Alabama and Texas are both under-recruited, if she couldn't find the missing talent in Texas, then it doesn't look promising for her to find the missing talent in Alabama. At least within Alabama, you have less far to travel.
(*) Cannon thought that there was something unique to Louisiana that made players want to stay in-state. Compare that number to Minnesota, which has 79 Division I players but only 8.9 percent of those remaining in-state. (Minnesota only has one D-I school, where Louisiana has 12.)