# Which goalies face the easiest and hardest shooters?

Discussion in 'By The Numbers' started by Hockey Outsider, Aug 25, 2005.

Not open for further replies.
View Users: View Users
1. ### Hockey OutsiderRegistered User

Joined:
Jan 16, 2005
Messages:
4,655
639
Trophy Points:
139
Not every goalie faces shots of the same difficulty. Shots from certain players are more difficult to stop than average, while others are easier. I am attempting to adjust for this.

For example, if Brendan Shanahan shoots, he has an 8.9% chance of scoring (25 goals on 280 shots). Thus, an average goalie will allow 0.089 goals per shot, and make 0.911 saves. If a goalie faced nothing but Shanahan shots, he'd be expected to save 91.1% of them. More complex example: Brendan Shanahan, Kris Draper and Nicklas Lidstrom have shooting percentages of 8.9%, 15.5%, and 5.2%, respectively. If a goalie faced five shots from Shanahan, two from Draper and three from Lidstrom, he would be expected to allow 0.911 goals and make 9.089 saves. Since the goalie faced ten shots, this corresponds to an expected save percentage of 90.89%. A high expected save percentage means the goalie faces easy shots, and a low expected save percentage means the goalie faces hard shots.

I compiled this data for every goalie for 2004. Here are the results:

Expected Save Percentage
(Ranked from highest/easiest shots to lowest/most difficult shots)

1 TOSKALA_SJS 91.5%
2 HEDBERG_VAN 91.3%
3 KIPRUSOFF_CGY 91.3%
4 DIPIETRO_NYI 91.3%
5 TUREK_CGY 91.3%
6 SAUVE_COL 91.2%
7 THIBAULT_CHI 91.2%
8 LEGACE_DET 91.1%
9 BRATHWAITE_CBJ 91.1%
10 NORONEN_BUF 91.1%
11 GERBER_ANA 91.1%
12 MARKKANEN_NYR 91.1%
13 VOKOUN_NAS 91.1%
14 NABOKOV_SJS 91.1%
15 MCLENNAN_CGY 91.1%
16 GRAHAME_TBL 91.1%
17 LEIGHTON_CHI 91.1%
18 BOUCHER_PHO 91.0%
19 FERNANDEZ_MIN 91.0%
20 JOSEPH_DET 91.0%
21 ESCHE_PHI 91.0%
22 ROLOSON_MIN 91.0%
23 LALIME_OTT 91.0%
24 BURKE_PHO 91.0%
25 DENIS_CBJ 91.0%
26 KOLZIG_WAS 90.9%
27 CONKLIN_EDM 90.9%
28 KHABIBULIN_TBL 90.9%
29 AEBISCHER_COL 90.9%
30 CECHMANEK_LAK 90.9%
31 DUNHAM_NYR 90.9%
32 TURCO_DAL 90.9%
33 CARON_PIT 90.8%
34 GARON_MTL 90.8%
35 FLEURY_PIT 90.8%
36 LUONGO_FLA 90.8%
37 HUET_LAK 90.8%
38 NURMINEN_ATL 90.8%
39 BRODEUR_NJD 90.8%
40 AUBIN_PIT 90.8%
41 ANDERSON_CHI 90.8%
42 GIGUERE_ANA 90.8%
43 SALO_EDM 90.8%
44 CLOUTIER_VAN 90.8%
45 BELFOUR_TOR 90.7%
46 THEODORE_MTL 90.7%
47 WEEKES_CAR 90.7%
48 DAFOE_ATL 90.7%
49 BIRON_BUF 90.6%
50 HACKETT_PHI 90.6%
51 OSGOOD_STL 90.6%
52 RAYCROFT_BOS 90.5%
53 SNOW_NYI 90.5%
54 POTVIN_BOS 90.4%
55 PRUSEK_OTT 90.2%

Quick observations:
- At the top of the list we see a lot of goalies from San Jose and Calgary. Clearly, these teams do an excellent job of protecting their goalies.
- It was a bad year to be a goalie in Boston. Raycroft and Potvin are both in the bottom four in terms of shot difficulty.
- Theodore was left for dead by his defense. How things have changed in Montrealâ€¦ I wonder what Ken Dryden would say.
- Itâ€™s interesting to note that DiPietro faced among the easiest shots in the league, while Snow faced some of the hardest. They played behind the same defense, so why is Snow facing much tougher shots? Did he consistently play against better teams? Did the Islanders open up when he was in net?
- This data doesn't support the "Brodeur faces really easy shots" theory. In fact, it shows us that Brodeur faces tougher shots than most goalies.

2. ### PecaFanRegistered User

Joined:
Nov 16, 2002
Messages:
8,942
14
Trophy Points:
141
Location:
Ottawa (Go 'Nucks)
I generally like your stuff, but I'm not so sure on this one, Outsider.

First of all, with only one percent separating over fifty goalies, it looks to me like this is just simple random noise or variation at work here. Two goalies on the same team, both playing behind the exact same defense, one at the top of the list (Hedberg), and one at the bottom (Cloutier) supports this as well.

I'm not sure I buy the basic theory as well. Kris Draper has a higher shooting percentage than Shanahan. So a goalie who faced 10 shots from Draper would have a toughter time than a goalie who faced 10 from Shanahan? I don't think so.

3. ### Snap WilsonRegistered User

Joined:
Sep 14, 2003
Messages:
5,838
0
Trophy Points:
0
Not necessarily. This might be due to the quality of competition faced by Cloutier and Hedberg. Hedberg could be used against some of the lesser teams in the league for instance.

There will be a certain amount of noise over any given season. I'd like to see how it plays out over two or three seasons.

HO, could you combine the stats by team instead of individual goalie? I think that would be more informative.

4. ### Hockey OutsiderRegistered User

Joined:
Jan 16, 2005
Messages:
4,655
639
Trophy Points:
139
Seeing the data so closely grouped together also made me wonder how valid this theory was. I was definitely expecting a larger variation.

It's possible that for Vancouver, the Canucks had more confidence and opened up when Cloutier was in net, whereas they played a more defensive game when Hedberg was in goal. Or, Hedberg (I think) played against easier opponents. Of course, this could all be random variation, too.

The Shanahan/Draper probably isn't a good one (it's my fault for picking those players), because Shanahan had an unusally bad year offensively while Draper had by far his best season. But in 2004, as a goalie, I'd rather face an average Shanahan shot than an average Draper shot.

Same. Unfortunately, I'm pretty sure this data (ie matching each shooter to each goalie) exists only for 2004. It's important to see if this stay relatively consistent from year to year.

Here are all teams sorted by Expected Goals Allowed (shots faced * shot difficulty). This takes into account both shot difficulty, and number of shots allowed.

Team Shot Exp Goal Exp Sv%
DAL 1894 174.0 90.8%
CGY 2082 182.8 91.2%
NJD 1997 186.4 90.7%
TBL 2078 188.1 90.9%
OTT 2034 189.5 90.7%
DET 2143 192.3 91.0%
PHI 2090 195.0 90.7%
EDM 2172 197.2 90.9%
STL 2098 197.4 90.6%
VAN 2181 198.1 90.9%
TOR 2164 200.8 90.7%
LAK 2203 201.7 90.8%
CAR 2142 203.6 90.5%
NYI 2243 203.8 90.9%
NAS 2323 208.1 91.0%
SJS 2370 208.5 91.2%
COL 2318 208.9 91.0%
MIN 2379 214.3 91.0%
BOS 2276 216.3 90.5%
MTL 2333 216.7 90.7%
BUF 2336 216.8 90.7%
NYR 2395 218.1 90.9%
ATL 2397 220.0 90.8%
ANA 2436 221.9 90.9%
PHO 2493 225.4 91.0%
CHI 2510 226.1 91.0%
CBJ 2613 235.5 91.0%
WAS 2577 237.8 90.8%
PIT 2707 248.1 90.8%
FLA 2822 260.5 90.8%

The teams at the top of the list make sense: New Jersey, Tampa Bay, Dallas, Calgary, Ottawa, etc., all have reputations for being great defensively. I was a bit surprised to see St. Louis in the top ten, however.

Florida, Pittsburgh and Washington each allowed a ton of shots, and they were more difficult than average. Rangers, Buffalo and Atlanta are near the bottom too. Seeing Boston near the bottom makes Raycroft's performance even more impressive.

These numbers make Luongo look really good. He faced 2,475 shots with an expected save percentage of 90.8%. He was expected to allow 228 goals, but in reality allowed 172. So by this measure, Luongo saved his team 56 goals (compared to an average goalie)!

Shot difficulty on its own isn't overly useful. I think combining it with the number of shots allowed gives us interesting, reasonably accurate results.

Last edited: Aug 25, 2005
5. ### JonathanKMcOptimistic

Joined:
Aug 1, 2005
Messages:
3,073
1
Trophy Points:
96
Location:
Edmonton, AB
IMO (no offence) I think that somebody has wayyyyyyyyyyyyyyyyy too much spare time.

6. ### me2Calling out the crap

Joined:
Jun 28, 2002
Messages:
33,585
1,511
Trophy Points:
214
Location:
Blasting the bull***
I've decided I dislike sv%. It distorts the truth. You might be food that "90% fat free" because that sounds good. But write "10% fat" on the label and that sounds bad. ITS SAME THING.

[assuming sv% is meaningful without looking too deep at the shooting percetages faced]

For example say two goalies face the same 2000 shots over the season. One goalie finishes with a sv% of 90.0 the other 91.0%. So is goalie #2 1% better goalie or 10% goalie. Personally I'd say he's 10% better since he allowed 10% less goals (200 goals vs 180).

Goalies should be looked at based on fail percentage. Goalie #1 has a 10% fail%, goalie #2 has 9% fail%. Luongo has 6.7% fail% (93.3 sv%).

Compare Loungo to goalie #1. 6.7 vs 10. It now becomes obvious goalie #1 allowed 50% more goals than Luongo. 50% more!!!! Vote for fail%, you know it makes sense.

7. ### me2Calling out the crap

Joined:
Jun 28, 2002
Messages:
33,585
1,511
Trophy Points:
214
Location:
Blasting the bull***
I calculated shooting percetages using a slightly different method based on shot distance.

8.96 HEDBERG_VAN
9.25 AULD_VAN
9.76 CLOUTIER_VAN

I found quite a few of the backup had a much easier time than the starters.

8.82 TOSKALA_SJS (I think Toskala played 30 games)
9.56 NABOKOV_SJS

Easier opposition combined with the teams playing more controlled protective hockey with their backups would be my thinking.

Joined:
Aug 8, 2005
Messages:
2,024