Grabovski Hypothesis – Regression Refined (defense)

28 Aug

Regression Refined (Offense)

1st Regression and Post

How does Corsi and Fenwick help in goal prevention?  Here is a first run!

Again, I parceled out blocked shots for (UBS), missed shots for (UMS) and shots against.

More items were statistically significant this time.  Obviously, the less shots you have on goal against you the less you are scored against.  Also, in the obvious column is a keeper’s save percentage–the higher the better.

Yet, running the regression with blocked shots, missed shots, blocked shots for and missed shots for…none of the Corsi and Fenwick stats matter for preventing goals.  The “tilt of the ice”, or possession, doesn’t appear to affect being scored on or scoring against your opponent.  Well…

There is actually a positive coefficient on blocked shots for.  So, if you are being worked by the other team, even if you are blocking shots, you have a better chance of being scored on.  But, it is small.  For every 153 shots a team blocks (on average) they will have one goal scored against them.  This doesn’t suggest a team shouldn’t block shots, but that they shouldn’t be in a position where they have to block shots.  However, this doesn’t seem to mean simple possession is the answer to preventing goals.

What does matter is offensive zone faceoff percentage…the first time we saw this stat.  So, the more you are taking faceoffs in your end, the more you are preventing the opponent from scoring.  This regression says NO?!?  How can that be?

The regression shows that on average, a 16% increase in offensive zone faceoffs leads to one goal against.  I don’t pretend to understand this, but that is what the stats say.

However, this was one of the big assumptions made about Grabovski.  That in Toronto he was getting a lot less offensive zone faceoffs than what he will get in Washington.  The increase in offensive zone faceoffs actually leads to more goals being scored against you.  So, for some reason, offensive zone faceoff percentage is not good for defense and it has no affect on offensive scoring.


2007/08 through 2012/13

Estimate              Std. Error             t value                  Pr(>|t|)

(Intercept)          1.476e+03           2.571e+01           57.412                   <2e-16 ***

SF                           -4.354e-03           2.880e-03            -1.512                    0.1325

SA                           7.691e-02            2.981e-03            25.800                   <2e-16 ***

BS                           -1.455e-03           4.021e-03            -0.362                    0.7180

MS                         -2.449e-03           5.711e-03            -0.429                    0.6687

UBS                        6.544e-03            3.118e-03            2.099                     0.0374 *

UMS                      2.008e-04            5.317e-03            0.038                     0.9699

Sh.                          3.206e-01            2.897e-01            1.107                     0.2700

Sv.                          -1.617e+01          2.812e-01            -57.492                 <2e-16 ***

OZFO.                   6.196e-01            2.492e-01            2.486                     0.0139 *

DZFO.                    6.359e-02            2.003e-01            0.317                     0.7513

east                       -7.814e-01           1.268e+00           -0.616                   0.5387

west                      -8.145e-01           1.349e+00           -0.604                   0.5468

yr13                       -4.903e+00          2.608e+00           -1.880                    0.0619 .

yr12                       -4.675e-01           1.071e+00           -0.437                    0.6630

yr11                       -7.005e-01           1.068e+00           -0.656                    0.5130

yr10                       -2.205e-01           9.284e-01            -0.237                    0.8126

yr9                          -1.085e-01           8.114e-01            -0.134                    0.8938

yr8                          NA                          NA                          NA                          NA

Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: