When I started this message, I felt that the current PP% system was
best.  I felt that PP% based on time played was not fair: When you
score 1:00 into the PP, it would come out as 200%. (!)  On a night that
you score one PP goal in three opportunities (and that goal is at 0:30
of the PP), your PP% for the night is .444!

For curiosity, I tried to apply the above method to see how much the
numbers would inflate.

First, consider how far into a PP is the average goal is scored.
Because seemingly few are scored in the first 30 seconds, I'll take a
stab at the average being 1:15 (or 1.25 minutes).  For simplicity, I
need to toss out majors, presuming they won't affect this much.

Say a team is 25 for 100 in today's system.  That is 200 PP minutes
minus the amount of time they were cut short due to scoring those 25
goals at 1:15 of the PP.  Remove .75 min * 25 times = 18.7 minutes.
So, they have 25 goals in 181.25 minutes.  Similar to GAA (see also
baseball ERA), divide 181.25 by 2 because there are 2 minutes per PP
and viola we have 25/(181.25/2) = .276 is the PP% based on time played.

The difference is fairly small, and gets smaller the worse a team is.

Now, my turning point was when I realized that the current system
penalizes a team for scoring early on a PP.  The team's PP% is charged
with the entire 2:00, even though they only played on PP for 1:15, or
0:30, or 0:06.  (Except that it is the opposite because it _helps_ his
GAA) it would be like giving Sioux goalie Kollar credit for allowing
only two goals in a game even though he was pulled at 6:15 of the first
on Dec. 1.  His true GAA for that game is calculated to be 52.64 while
using the PP% method, he would only get 2.00.

Back to my example:
On a night that you score one PP goal in three opportunities (and that
goal is at 0:30 of the PP), your PP% for the night is .444!  Yes, that
is because .444 goals were scored for every 2:00 of PP time played.

Tradition and recordkeeping in the same format for historical
comparison are completely other discussions which would make me
reluctantly decide, in the end, to keep the current scoring rules--
these statistical exercises become moot.  However, I have convinced
myself that scoring PP% by minutes rather than opportunities would more
accurately reflect how often a team scores when on advantage.

Fortunately, unless a team had a habit of scoring the vast majority of
PP goals early or late in their 2:00 advantages, rankings of teams will
not change at all and this discussion becomes even more... goal-less?

Kirk