Calculations find Bruins consistently inconsistent

  Gilbert Quinonez The stat geek dreams of
having his own computer rankings used by the BCS and spends his
Friday and Saturday nights analyzing the world of sports stats.
E-mail him at [email protected].

Matt Barnes is overrated. Dan Gadzuric is overrated
too.

No, I’m not a drunken maniac from USC. I’m the
statistical geek, armed with his first column, the Superman-like
power of being able to stay awake in math class and God’s
gift to mankind: a calculator.

UCLA fans have been outraged at the men’s basketball
team’s recent losing skid, going 3-4 in the last 7 games. The
Bruins lost to the 12th best team in the Big East (Villanova), lost
by 29 to a team that didn’t even make the NCAA Tournament
last year (Oregon), beat the former No. 1 team in the country
(Kansas) and beat the current No. 7 team in the country (Alabama).
What’s wrong with the Bruins? The stat geek knows.

Barnes does not need the ball. The team only seems to do worse
when he takes control of the game. Barnes does score more when the
team loses than when it wins (13.3 in wins, 14.9 in losses), but
his overall play gets worse. He takes more shots when the Bruins
lose, and his field goal percentage drops significantly (51.1
percent to 47.3 percent). In Barnes’ top two scoring
games in his career (34 against USC, 32 against Stanford last year)
the Bruins lost both games. When Barnes scores at least 20 points,
UCLA has gone 1-3 this season. His non-scoring stats take a hit
too. Barnes averages more turnovers (2.5 to 2.9), fewer assists
(3.6 to 2.9), fewer rebounds (6.1 to 5.0), and more fouls (1.9 to
3.0) when the Bruins lose.

A bad shooting performance shouldn’t affect how you help
your team. Jason Kapono shoots for a lower percentage when the
team loses (49 percent win, 44.3 percent loss), scores less (17.9
to 17.3) in games the Bruins lose, as would be
expected. However, the rest of his statistics don’t
suffer from the poor shooting. Kapono commits fewer turnovers
when the team loses (2.1 to 1.3), fewer fouls (2.1 to 2.0) and more
assists (2.4 to 2.6). Kapono’s rebounds drop (5.8 to
5.3), but not to the extent Barnes’ totals do.

Just like the basketball team’s performance lately, other
stats can be very confusing, even for the stat geek. Many fans
have thought Dan Gadzuric’s lack of playing time due to his
constantly being in foul trouble greatly hurts UCLA. However, the
stats indicate otherwise. Gadzuric plays less minutes when the
Bruins win (23.1 to 23.4) than in a loss. It’s all about
quality, not quantity. Gadzuric just plays terrible when the Bruins
lose. Gadzuric’s scoring goes down (11.3 to 8.1), turnovers
go up (1.2 to 2.1) and his free-throw shooting becomes Shaq-like
(51.6 percent to 33.3 percent). Although it would help, the Bruins
don’t need Gadzuric on the floor more. They just need him to
play consistently well.

The Bruins’ inconsistent play has also led to some very
weird stats. They actually average more turnovers when they
win (15.4 to 15.0) and less blocked shots (2.2 to 2.3). Remember,
these stats aren’t summed up in one bad game. UCLA has
lost seven games.

Just how inconsistent is UCLA? One way to calculate this is
using something I learned in statistics class, the correlation
coefficient. A correlation coefficient of 1.0 would indicate
perfect consistency, and a 0.0 would indicate no consistency at
all. This stat is calculated by comparing the margin of victory
with the opponent’s winning percentage in each game, along
with more complicated standard deviations, sigma notation, and
fractions. After extensive use of Microsoft Excel, UCLA’s
correlation coefficient is 0.36, meaning UCLA is about as
inconsistent as international figure skating judges.

With this great consistency, UCLA will probably beat Duke in
March, only to lose to UC Irvine in the next round. I’d be
calculating correlation coefficients for weeks. Of course, I do
stuff like that anyway.