/cdn.vox-cdn.com/uploads/chorus_image/image/35611246/20140201_lbm_sx9_250.0.jpg)
The question was raised this Spring as to whether better defensive teams have a negative effect on save percentage. The notion was that better teams block more low percentage shots. The shots that make it on goal are therefor more dangerous. I doubt that this is the case, but let's dig into it.
One approach would be to look at where shots get blocked. Unfortunately, the data for this is mostly non-existent. The little data that exists is poor quality and not likely to yield a valid answer. I took the approach of looking at how many shots get blocked and how that relates to various other measures of team defensive performance.
Some Global Numbers
Total Corsi Events: 273180
Total Blocked Shots: 70455
Team with most Corsi Against: Toronto 10627
Team with least Corsi Against: New Jersey 7189
Blues are 4th at 8819
Team with most CA-Blocked Islanders 2724
Team with least CA-Blocked New Jersey 1708
Blues are 21st with 2223
On average, the probability of blocking a Corsi Against is 0.2579068746
Montreal has the highest probability at 0.2910816908
Ottawa has the lowest probability at 0.2278819821
The Blues are 5th at 0.2760751681
Do teams differ in the percentage of shots they block?
I ran a logistic regression to see. Teams do differ significantly.
> LinearModel.1 = lm(bllogit ~ Team, weights=CA, data=Team3yrs)
> anova(LinearModel.1)
Analysis of Variance Table
Response: bllogit
Df Sum Sq Mean Sq F value Pr(>F)
Team 29 2008.21 69.249 4.3414 7.832e-07 ***
Residuals 60 957.05 15.951
Looking at the coefficients, we see that Ottawa, Los Angeles, Detroit, Phoenix, and New Jersey are worse than average, and the Rangers, the Islanders, San Jose, and Montreal are better than average.
Coefficients: |
Estimate |
StdError |
tvalue |
Pr(>|t|) |
(Intercept) |
-1.056876 |
0.007641 |
-138.311 |
2e-16 |
OttawaSenators |
-0.182065 |
0.042561 |
-4.278 |
6.89e-05 |
LosAngelesKings |
-0.148983 |
0.045969 |
-3.241 |
0.001946 |
DetroitRedWings |
-0.122067 |
0.045479 |
-2.684 |
0.009388 |
FloridaPanthers |
-0.111726 |
0.042757 |
-2.613 |
0.011325 |
PhoenixCoyotes |
-0.102334 |
0.041939 |
-2.44 |
0.017656 |
NewJerseyDevils |
-0.109428 |
0.04772 |
-2.293 |
0.025359 |
BuffaloSabres |
-0.077568 |
0.041098 |
-1.887 |
0.063946 |
CarolinaHurricanes |
-0.071811 |
0.040934 |
-1.754 |
0.084488 |
ColumbusBlueJackets |
-0.058729 |
0.043181 |
-1.36 |
0.178901 |
BostonBruins |
-0.046062 |
0.04343 |
-1.061 |
0.293122 |
ColoradoAvalanche |
-0.041534 |
0.041449 |
-1.002 |
0.320349 |
VancouverCanucks |
-0.033905 |
0.04359 |
-0.778 |
0.439725 |
TorontoMapleLeafs |
-0.026719 |
0.039489 |
-0.677 |
0.501252 |
TampaBayLightning |
-0.001522 |
0.042267 |
-0.036 |
0.971403 |
ChicagoBlackhawks |
0.00367 |
0.044682 |
0.082 |
0.93482 |
EdmontonOilers |
0.009968 |
0.041059 |
0.243 |
0.809019 |
DallasStars |
0.018265 |
0.042059 |
0.434 |
0.665652 |
AnaheimDucks |
0.020553 |
0.042208 |
0.487 |
0.628061 |
NashvillePredators |
0.030684 |
0.04181 |
0.734 |
0.465873 |
PhiladelphiaFlyers |
0.039444 |
0.042661 |
0.925 |
0.358877 |
WinnipegJets |
0.040547 |
0.041712 |
0.972 |
0.334918 |
MinnesotaWild |
0.042426 |
0.042374 |
1.001 |
0.320741 |
WashingtonCapitals |
0.041859 |
0.041655 |
1.005 |
0.318983 |
PittsburghPenguins |
0.0441 |
0.043044 |
1.025 |
0.309702 |
CalgaryFlames |
0.072423 |
0.041333 |
1.752 |
0.084852 |
St.LouisBlues |
0.081774 |
0.044989 |
1.818 |
0.074109 |
NewYorkRangers |
0.110108 |
0.042527 |
2.589 |
0.012056 |
NewYorkIslanders |
0.114277 |
0.041246 |
2.771 |
0.007438 |
SanJoseSharks |
0.158662 |
0.041983 |
3.779 |
0.000364 |
MontrealCanadiens |
0.164789 |
0.042059 |
3.918 |
0.000231 |
Blocks and Team Defense
There is a strong relationship between the total number of shots blocked and the total number of Corsi events against. (CABperG is Corsi Against-Blocked, i.e., shots blocked, per game, and CAperG is Corsi Against per game.)
There is a small positive relationship between the percentage of shots blocked and the number of Corsi events against per game. Teams that give up a lot of Corsi events tend to block a slightly higher percentage of those events than average teams. (CAperG is Corsi Against per Game and BLPCT is Blocked Shots/Corsi Against).
There is no relationship between the percentage of shots blocked and the number of shots against per game. (SAperG is Shots Against per Game and BLPCT is Blocked Shots/Corsi Against).
There is no relationship between the percentage of shots blocked and the number of goals against per game. (GAperG is Goals Against per Game and BLPCT is Blocked Shots/Corsi Against).
There is no relationship between the percentage of shots blocked and team Even Strength Save Percentage.