Clifton D. Sutton
I grew up in Roanoke, Virginia, and except for four years in California while
in graduate school, I have lived in various places in Virginia my entire life.
I earned a B.S. degree in Applied Mathematics from the University of Virginia,
and then went to Stanford University, where I earned M.S. and Ph.D. degrees
in statistics. My dissertation pertained to random packing, an area of
geometric probability, and specifically one of the topics I investigated was
random packing in high dimensional spaces.
Since being at GMU (I've been here for about 21 years now),
my chief interest has become computer-intensive,
robust, and nonparametric methods for applied statistics, although I still love probability,
theoretical statistics, and advanced mathematics. Currently I'm being supported to work with a
research team looking into using statistical methods for writer verification
and writer identification. (We use measurements made on handwritten
characters to try to determine if two handwritten documents were written by the same person, or to try to match the
unknown writer of a handwritten document to someone in our database.)
In my spare time I like to play no limit Texas Hold'em in the Capital Poker League,
and on Fridays I like to go to happy hour and shoot pool. (In the poker league
one earns points (and sometimes money) by finishing well in tournaments --- the
better your finish the more points you get, and the points are also proportional to
the number of players entered. For the four scoring periods that the league had from
May 2006 through November 2007, I finished with the most points each time, and currently I'm in first place again,
even though lately I've been playing only about one night a week on average.)
In January 2007, a reporter for The Mason Gazette wrote an
article about my poker playing.
I like to watch sports on TV, and I like to investigate statistical issues
involving sports. I'll give four examples below.
- When baseball announced its
All-Century team in 1999, I thought that at
least one or two players from the old
Negro Leagues should have been included.
Possibly, in addition to lack of knowledge about these players, some
people didn't
vote for them because statistical comparisons are difficult. I plan to use
statistical methods to relate Negro League statistics to those from the National
and American Leagues.
- Jim Larranaga, the head coach of
George Mason University's men's basketball team asked for some
assistance with data analysis so that he can learn more about his team,
and in August 2002 I did a fairly thorough analysis of the team's
2001-2002 season. Basically, I looked for relationships amongst
variables such as field goal percentage, percentage of field goals for
which a player was credited with an assist, percentage of offensive
rebounds obtained from all rebounds at each end of the court, by GMU and
their opponent, numbers of turnovers, numbers of fouls, etc., and tried
to determine how each of these variables influenced whether GMU won or
lost. I presented some of my work and conclusions at a
statistics seminar in November, 2002.
(A reporter for a GMU publication saw the seminar announcement and wrote
a
story about my interactions with Coach Larranaga. Also, my dean
interviewed me, along with Coach L, on an episode of his Tech Horizons TV show.
- I wonder if the systems used in fantasy leagues to rate player production
are as good as they could be. For example, in basketball, should a rebound
count more or less than an assist, and how should these relate to both made
and missed field goals?
I've developed a
system for measuring player
performance, and have created a
a web page giving various measure of
player performance for the GMU men's basketball team and a
a web page giving various measure of
player performance for the GMU women's basketball team.
(Unfortunately, I just haven't had time to keep these pages about the GMU teams up to date.)
- GMU's Director of Athletics, Tom O'Connor, is on what is known as the selection committee for
the NCAA men's basketball tournament. Before he left to go to the meetings to select the teams
for the 2006 tournament he requested that I supply him with my input on the RPI. (He explained
that each year the committee likes to discuss the methods they use to evaluate and compare the
teams from around the country.) I didn't have a lot of time to do a thorough analysis, but I was
able to look into some aspects of the RPI, and I created a
a web page pertaining to the RPI. I guess Tom forwarded the
information to someone in the NCAA, because within 12 hours of when I finished my web page I got
feedback from someone who inidcated that I made some valid points pertaining to the RPI (but at
the same time seemed a bit defensive about the current use of the RPI). I plan to do much more
with an analysis of the RPI when I have a chance.
- I've developed a system, based on logistic regression, to rank college football and college basketball teams.