When working on an update to my previous article about the demographics of the current D-I basketball players - to look at which schools had the most in-state players on the roster and which schools had the least in-state players - I discovered an error. An error can be a little error or a big error, and I had no idea which one it was.
Oh boy, I thought. This is going to blow up big and make a bigger fool out of me than I already am. What had happened was that during the data preparation the players from Santa Clara in California were attributed to South Carolina - the Gamecocks all of a sudden had their roster size doubled with a lot of non-South Carolina players added to their team. I corrected that error, but braced myself when I recreated the regression. What would the new results be?
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
Total Projected Players = 3.10*Number of D-I High Major Players + 4.09*Number of D-I Schools in State + (1/1,219,967)*(2012 Population of State) - 4.25
Commenter bukuma in my previous post commented that:
"Also, why would state population not matter? That is counter to logic, more players, more competition to shine those diamonds."
He is referring to the low state coefficient in the old formula, where you need a state population of 23,676,596 to add one more projected D-I player. The coefficient in the newer formula implies you need a population of 1,219,967 to add one more projected D-I player. The corrected regression makes state population much, much more significant. It's not as significant as the coefficient that Drew Cannon had in the original article, but it means that the corrected formula shows a real contribution now between a state's population and the number of projected D-I players.
Did the new formula change the states in the Top 5 Over-recruited and Top 5 Under-recruited states? Except for one state, no. All the new regression did was change their order:
Top 5 Under-recruited
1. New York
5. South Carolina
The former list was Alabama, then New York, Iowa, South Carolina and Texas.
Top 5 Over-recruited
Virginia falls to #6 with the new formula - over-recruiting is still a problem there according to the model. California switches from #6 to #3 in our corrected list.
California's appearance wouldn't surprise Drew Cannon, the man who came up with the idea for this regression. According to Cannon, when recruiters think of the West Coast, they usually just think of California and forget about Washington and Oregon.
There was some interest expressed in seeing the entire list of schools, their total players, high major players, projected D-I players, etc. I don't know if our template at Swish Appeal will hold it all, but here it is.
|State||StateName||Total Players||High-Major Players||D-I schools||State Population||Predicted_Updt||Difference||% in-state|
|DC||District of Columbia||13||3||4||632,323||26.2||-13.2||0.0%|
Percentage of state players on local rosters
Looking at all this demographic data, we can now look at individual schools and see how much in-state recruiting they do. There are 345 D-I schools in the data provided by wbbstate.com.
There are 22 Division I schools that did not have a single in-state player on their roster according to wbbstate.com. (We count the District of Columbia as a state.)
Florida Gulf Coast
Note that all four DC schools - American, George Washington, Georgetown and Howard - are on the list. This could be an anomaly - DC has about 43,000 public school students of all ages. It could be that there are students who go to a DC high school but are technically counted as being from Maryland or Virginia.
Brown, Columbia, Dartmouth, Harvard, and Yale are all Ivy League schools with their own recruiting challenges. Likewise, recruiting for Navy means that a player has to do more than just dribble a ball to play for the Naval Academy.
There are three schools that have all of their students as in-state students:
Cal State Fullerton
These three schools are in California and Texas, which might as well count as countries. Even though about 50 percent of California (or Texas) D-I players end up in-state, a Top 10 list of the most in-state centric schools would have six California schools:
North Carolina Central
San Diego State
Cal State Fullerton
Seeing the Golden Bears on the list leads me to think that you could very easily get a national championship contender just by assembling an All-Star team of D-I Californians.
Outside the US
If you're looking for a foreign pipeline, the best place to look would be at America's neighbor to the North, Canada. There are 44 D-I players from Ontario alone, a greater number of D-I players than those provided by 22 individual American states! Combine the D-I players from Vermont, Alaska, Rhode Island, Wyoming, North Dakota, Maine, and Hawaii and you still wouldn't equal the combination made from Canada's most populous province. (A trip to Toronto might be worth taking to see some basketball.)
Is Ontario over-recruited or under-recruited? I really can't say. The regression isn't stretched to include foreign countries. That would require an expert on Canadian girl's basketball, which is a niche market just waiting to be filled.
Quebec adds 10 players, British Columbia adds 8 Alberta adds 5, and Saskatchewan adds two. Manitoba, New Brunswick and Nova Scotia add one each, bringing the total number of Canadian players to 72.
After Ontario, the next biggest sources of foreign players are Australia (37) and the United Kingdom (16). This makes sense. One would expect the biggest sources of foreign-born D-I players to be from English-speaking countries.
Non-English speaking countries provide players as well. Europe has a club system which can send kids from secondary school straight to the pros or to national teams, so there's not as much need for a European kid to go to an American college. The two non-English speaking countries that provide the most D-I players are Latvia (12) and Sweden (12). I would probably have to ask the writers at lovewomensbasketball.com why these are such popular pipelines.
The complete list of countries and the number of D-I players on rosters as of the 2012-13 season follows:
|British Virgin Islands||1|