I'm going to respond to this thread the same way I've responded to some
others: by saying that I'm much more knowledgeable about this kind of
question when it comes to baseball.  My instinct is that there is at least
some similar answers, though.  I have two reasons for thinking this: the
logic behind the analysis is the same, and the conventional wisdom is the
same in both sports (i.e. Defense Wins).  It would be interesting to dig up
the empirical evidence in hockey to find out if my suspicions are correct,
but I have neither the time or all of the proper training to do so.
 
The best work I've seen on this subject (as on many other subjects in
sabermetrics [the familiar term for this kind of quantatative baseball
research {should I nest my statements any farther?}]) was about ten years ago
by Bill James and the rest of the crew that worked on the Baseball Abstract.
 He tried to assess the accuracy ofthe old statement, "Pitching is 75% of
baseball".  He observed that baseball (much like hockey, or most team sports)
is a zero sum game; in the aggregate, for every game won, one is lost and for
every run scored, one is given up (by the other team).
 
His first problem was that, while everyone seems to use the cliched phrase,
no one ever seems to define it.  He then proceeded to run analyses based on a
large number of possible definitions.  Many of these related to the
performance of individual players and are not translatable into hockey terms
(though they're very interesting and cast doubt on another baseball cliche,
that "Good pitching beats good hitting".)  Some of the questions related to
team performances answered precisely the same questions that have been asked
here.  James used a number of different comparisons to measure the
correlation between offensive production and winning vs. the same correlation
with defense.
 
In sum, he found that defense was no better a predictor of team success
(measured either by winning percentage or championships) than was offense.
 
This doesn't mean that the same is necessarily true of hockey.  In addition
to the fact that the correlations might differ, teams could be assembled in
such a way that it is easier to build a team that wins with defense than
offense.  Also, the tactical differences might put a large element of doubt
in any such study; as Mike Machnik has pointed out, offense and defense
aren't as cleanly seperable in hockey as in baseball.
 
Nevertheless, I've been convinced of this issue in one sport.  Until it is
demonstrated otherwise, I'll believe that offense first and defense first are
equally valid approaches to building a successful hockey team.  Wayne Smith
was right, though.  Almost no one goes the distance by being one dimensional.
 
J. Michael Neal
 
HOCKEY-L is for discussion of college ice hockey;  send information to
[log in to unmask], The College Hockey Information List.