NHL Goaltender Strength of Schedule (2010-11 through 2011-12)

Doctor No · Jan 17, 2013

This is the first step of something that's going to end up on my goaltender page (since I could do what I describe below as far back as I have game logs, so to the mid-1980s):

http://hockeygoalies.org

(Go to biographies, and then each goaltender)

This is a first attempt to quantify the differing strengths of schedule that goaltenders face in the NHL (for the 2011-12 season here).

First, I developed an estimate of each team's strength - using their entire (regular season plus postseason data), I started with each team's goal differential (GF minus GA), and then adjusted for schedule (each team's average opponent's goal differential). This is an iterative process, but does converge to a metric that estimates how many goals better (or worse) a team is compared to average during the 2011-12 season. To give you an idea of the endpoints, Boston was at the top with a rating of 0.655, and Columbus was at the bottom with a rating of -0.666. I also calculate home-ice advantage as worth 0.286 goals in the 2011-12 season.

Next, I take each goaltender's opponents, weighted by minutes against each, and find the average strength of opponent. If they played on the road, I add in half of the home ice advantage (and if they played at home, I subtract).

Doctor No · Jan 17, 2013

Here are the results (showing only goaltenders who played at least 10 NHL games in 2011-12). The value reflects the average strength of opponent, in terms of how many goals they would be favored to beat an average team (so, if Goaltender X had a strength of schedule of 0.2, then their average opponent would be 0.2 goals better than average):

Doctor No · Jan 17, 2013

And for the 2012 playoffs (showing only goaltenders who played at least 3 NHL games). Note that because all of a goaltender's opponents are playoff teams, the average strength of opponent will usually be higher:

Doctor No · Jan 17, 2013

Some observations:

Surprised that Tim Thomas was the only goaltender to have an average below-average postseason opponent (if barely) - as it turns out, Washington comes out as -0.214 on the road, and 0.071 at home, and that was Thomas' only 2012 Cup opponent.

Playing Boston and the Rangers in the playoffs gives Braden Holtby the toughest playoff schedule award.

Interesting regular-season juxtaposition between Harding and Backstrom, Thomas and Rask, and Clemmensen and Theodore.

Alex Auld has the toughest 2011-12 regular season schedule by a country mile - he played Philly, Pittsburgh, Montreal, Calgary, St. Louis, Nashville and Boston at home, and Toronto, Washington, Buffalo, Vancouver, Pittsburgh, New Jersey, and Los Angeles on the road.

Doctor No · Jan 17, 2013

One thing I'd like to do next (aside from filling out the remainder of seasons that I'm able to do) is to see what the correlation is between strength of opponent and save percentage.

That may allow for some future "adjusted save percentage" based on the quality of opposition.

Kane One · Jan 17, 2013

Just to help me understand this, the lower the number, the tougher the schedule?

Doctor No · Jan 17, 2013

kaneone said:
Just to help me understand this, the lower the number, the tougher the schedule?

The higher the number, the tougher the schedule (the number represents how many goals better (or worse) per game the average opponent is).

Kane One · Jan 17, 2013

Taco MacArthur said:
The higher the number, the tougher the schedule (the number represents how many goals better (or worse) per game the average opponent is).

If a goalie wins like 3-0, wouldn't this then make their strength of schedule easier than it was before the game?

And also, doesn't this just compare the goalie to their opponent's goalie instead of comparing them to the forwards?

For example, the goalies in the Atlantic Division last year were Lundqvist, Brodeur, Bryzgalov, Fleury, and Montoya. The Pacific Division had Quick, Hiller, Niemi, Smith, and Lehtonen. Using this formula, Lundqvist had an easier schedule than Quick, even though the top scorer in the Pacific was Thornton with 77, while Malkin lead the Atlantic with 109. Six out of the top 10 point scorers were in the Atlantic while 0 were in the Pacific. I don't get how Quick's schedule could have possibly been harder than Lundqvist's.

You're a goalie, right? If you had to choose, which division would you rather play in?

Doctor No · Jan 17, 2013

You seem to be looking at this as a Rangers fan. I have no horse in any Lundqvist/Quick debates (although since it seems to matter, I prefer Lundqvist).

Having said that, I'll point out that (as I mentioned above) the team strength considers both goals scored and goals allowed of opposition (and not just goal scoring). If you view this strictly as a measure of offense, you'll be half disappointed.

To your other point, if you shut out a team by 3-0 margin, it will affect their rating yes. However, it's only 1 of 82 games (more for playoff teams), so it won't affect it very much.

I would also suggest that the differences in Quick's strength of schedule and Lundqvist's strength of schedule are essentially meaningless (about one goal every 40 games).

Doctor No · Jan 17, 2013

For tracking purposes, here are Lundqvist's 2011-12 opponents (and the strength I ascribe to each).

Date | Opponent | Location | Opp Strength
20111007|Los Angeles|A|0.607
20111008|Anaheim|A|-0.155
20111015|NY Islanders|A|-0.469
20111018|Vancouver|A|0.651
20111020|Calgary|A|-0.154
20111022|Edmonton|A|-0.188
20111027|Toronto|H|-0.556
20111029|Ottawa|H|-0.066
20111103|Anaheim|H|-0.441
20111105|Montreal|H|-0.340
20111109|Ottawa|A|0.220
20111111|Carolina|H|-0.552
20111115|NY Islanders|A|-0.469
20111123|Florida|A|-0.174
20111125|Washington|A|0.071
20111126|Philadelphia|H|0.196
20111129|Pittsburgh|H|0.474
20111203|Tampa Bay|A|-0.454
20111205|Toronto|H|-0.556
20111208|Tampa Bay|H|-0.740
20111211|Florida|H|-0.460
20111213|Dallas|H|-0.254
20111215|St. Louis|A|0.615
20111220|New Jersey|A|0.353
20111223|Philadelphia|H|0.196
20111226|NY Islanders|H|-0.755
20111230|Florida|A|-0.174
20120102|Philadelphia|A|0.482
20120106|Pittsburgh|A|0.760
20120110|Phoenix|H|0.048
20120112|Ottawa|H|-0.066
20120115|Montreal|A|-0.054
20120117|Nashville|H|0.191
20120119|Pittsburgh|H|0.474
20120121|Boston|A|0.798
20120124|Winnipeg|H|-0.462
20120201|Buffalo|A|-0.035
20120205|Philadelphia|H|0.196
20120207|New Jersey|H|0.067
20120211|Philadelphia|A|0.482
20120212|Washington|H|-0.214
20120214|Boston|A|0.798
20120219|Columbus|H|-0.809
20120221|Pittsburgh|A|0.760
20120225|Buffalo|H|-0.321
20120227|New Jersey|H|0.067
20120302|Tampa Bay|A|-0.454
20120304|Boston|H|0.513
20120306|New Jersey|A|0.353
20120309|Chicago|A|0.219
20120311|NY Islanders|H|-0.755
20120317|Colorado|H|-0.291
20120319|New Jersey|H|0.067
20120321|Detroit|H|0.334
20120323|Buffalo|H|-0.321
20120324|Toronto|A|-0.270
20120327|Minnesota|A|-0.446
20120328|Winnipeg|A|-0.176
20120330|Montreal|H|-0.340
20120401|Boston|H|0.513
20120403|Philadelphia|A|0.482
20120407|Washington|H|-0.214

(Note that you have to weight these by minutes played in each to get the final number above, but this is essentially his schedule)

Doctor No · Jan 17, 2013

And for Jonathan Quick:

Date | Opponent | Location | Opp Strength
20111007|NY Rangers|H|0.237
20111013|New Jersey|A|0.353
20111015|Philadelphia|A|0.482
20111018|St. Louis|H|0.330
20111020|Phoenix|A|0.333
20111022|Dallas|H|-0.254
20111027|Dallas|A|0.031
20111029|Phoenix|A|0.333
20111030|Colorado|A|-0.005
20111103|Edmonton|H|-0.473
20111105|Pittsburgh|H|0.474
20111107|San Jose|A|0.308
20111110|Vancouver|H|0.365
20111112|Minnesota|H|-0.732
20111116|Anaheim|H|-0.441
20111117|Anaheim|A|-0.155
20111119|Detroit|H|0.334
20111123|Dallas|A|0.031
20111126|Chicago|H|-0.067
20111128|San Jose|H|0.022
20111201|Florida|H|-0.460
20111206|Anaheim|A|-0.155
20111208|Minnesota|H|-0.732
20111213|Boston|A|0.798
20111215|Columbus|A|-0.523
20111217|Detroit|A|0.619
20111219|Toronto|A|-0.270
20111222|Anaheim|H|-0.441
20111223|San Jose|A|0.308
20111226|Phoenix|H|0.048
20111228|Chicago|A|0.219
20111231|Vancouver|H|0.365
20120102|Colorado|H|-0.291
20120105|Phoenix|H|0.048
20120107|Columbus|H|-0.809
20120112|Dallas|H|-0.254
20120114|Calgary|A|-0.154
20120115|Edmonton|A|-0.188
20120117|Vancouver|A|0.651
20120119|Calgary|H|-0.439
20120121|Colorado|H|-0.291
20120123|Ottawa|H|-0.066
20120201|Columbus|H|-0.809
20120203|St. Louis|A|0.615
20120207|Tampa Bay|A|-0.454
20120209|Florida|A|-0.174
20120211|NY Islanders|A|-0.469
20120212|Dallas|A|0.031
20120216|Phoenix|H|0.048
20120218|Calgary|H|-0.439
20120221|Phoenix|A|0.333
20120222|Colorado|A|-0.005
20120225|Chicago|H|-0.067
20120227|Nashville|A|0.477
20120303|Anaheim|H|-0.441
20120306|Nashville|A|0.477
20120309|Detroit|A|0.619
20120311|Chicago|A|0.219
20120313|Detroit|H|0.334
20120316|Anaheim|A|-0.155
20120320|San Jose|H|0.022
20120322|St. Louis|H|0.330
20120324|Boston|H|0.513
20120326|Vancouver|A|0.651
20120328|Calgary|A|-0.154
20120330|Edmonton|A|-0.188
20120402|Edmonton|H|-0.473
20120405|San Jose|H|0.022
20120407|San Jose|A|0.308

Doctor No · Jan 17, 2013

By the by, I love this table format - with the ability to sort columns. Thanks, pnep!

Doctor No · Jan 17, 2013

Here's Alex Auld - the regular season leader:

Date | Opponent | Location | Opp Strength
20111008|Toronto|A|-0.270
20111015|Washington|A|0.071
20111018|Philadelphia|H|0.196
20111111|Buffalo|A|-0.035
20111120|Vancouver|A|0.651
20111125|Pittsburgh|A|0.760
20111208|New Jersey|A|0.353
20111216|Pittsburgh|H|0.474
20111227|Montreal|H|-0.340
20111230|Calgary|H|-0.439
20120123|Los Angeles|A|0.607
20120207|St. Louis|H|0.330
20120209|Nashville|H|0.191
20120225|Boston|H|0.513

Doctor No · Jan 17, 2013

And Tuukka Rask - the regular-season trailer.

Date | Opponent | Location | Opp Strength
20111010|Colorado|H|-0.291
20111018|Carolina|H|-0.552
20111029|Montreal|A|-0.054
20111107|NY Islanders|H|-0.755
20111110|Edmonton|H|-0.473
20111117|Columbus|H|-0.809
20111125|Detroit|H|0.334
20111203|Toronto|H|-0.556
20111206|Winnipeg|A|-0.176
20111210|Columbus|A|-0.523
20111213|Los Angeles|H|0.322
20111223|Florida|H|-0.460
20111228|Phoenix|A|0.333
20120110|Winnipeg|H|-0.462
20120116|Florida|A|-0.174
20120121|NY Rangers|H|0.237
20120124|Washington|A|0.071
20120202|Carolina|H|-0.552
20120208|Buffalo|A|-0.035
20120217|Winnipeg|A|-0.176
20120224|Buffalo|A|-0.035
20120303|NY Islanders|H|-0.755

Doctor No · Jan 17, 2013

I posted Auld's and Rask's schedules so that you can see what a meaningful difference might look like. I think we can agree that Auld's schedule is significantly tougher (even if you may disagree with how I rank the teams).

Doctor No · Jan 17, 2013

Here are the numbers that enter into my 2011-12 calculations above (both regular season and postseason).

Different people will come up with different numbers. As I described above, these numbers are the average margin of victory for a team in the year, adjusted by their average schedule (which is the average average margin of opponents' victory - that's a mouthful). It's a standard "power rating" design - which could be improved upon, but it's simple to explain and non-proprietary.

Last but not least, home-ice advantage is calculated by this method to be worth 0.286 goals (the way I've implemented this above is to increase the home team's rating by half of this value, and to decrease the road team's rating by half of this value).

Doctor No · Jan 17, 2013

Here are the 2010-11 team rankings - there's a bit more spread here than in 2011-12 (and I'm hoping that when I run the SoS for 2010-11 goaltenders, it's a bit more interesting).

(2010-11 home ice advantage: 0.195 goals)

Doctor No · Jan 17, 2013

Hmmm...seems like I can crank these out decently fast.

Here is the 2010-11 regular season:

And the 2011 Stanley Cup playoffs:

Doctor No · Jan 18, 2013

Just compiling observations here as I note them - I was trying to figure out why, in the playoffs particularly, backup goaltenders had a tougher strength of schedule than starters.

Now it's obvious (one of those things that's obvious once you figure it out :laugh:

).

In the playoffs, backup goaltenders are likely to only play in relief, in a losing effort. This is more likely to happen against better teams.

(This also happens in the regular season, but because the schedule varies more in the regular season, coaches can also spot start the backup against lesser opponents if he chooses to do so. That can't happen in the playoffs.)

Doctor No · Jan 18, 2013

Once I have a few more years in the hopper, I'm looking forward to looking at this by head coach - do certain coaches start weaker goaltenders against lesser opponents more often? Younger goaltenders? Start them at home, and the veteran on the road? (Anecdotally, we know that this is "true" - but it should be seen here if so. And do the perceived evidence line up with the data?).

I think that there's a lot of mining potential here.

Czech Your Math · Jan 18, 2013

Interesting to see Rask and Schneider having among the easiest schedules in each of '11 & '12, since many are expecting quite a lot from them in the near future.

It probably doesn't make much difference, but are Shootout goals factored out of team goal differential? As someone already mentioned, team GF might be another relevant metric for difficulty of schedule for goaltenders. GF/GA ratio would be a third metric (although it wouldn't vary that much from differential). I guess arguments could be made for each of the three metrics.

Doctor No · Jan 19, 2013

I like the goals scored by opponents angle; I could estimate how many more goals a team scored than would be expected by their schedule.

I was getting caught up in trying to figure out what needed to iterate/converge, but I realize now that convergence isn't necessary.

I'll try and work something up tomorrow - thanks! :handclap:

seventieslord · Jan 20, 2013

If I understand this correctly, you're using a team's goal differential to calculate their strength?

If the endgame is to use this result to help understand the results a goalie posts against a certain competition level, wouldn't it be better to only look at the offensive strength of the teams they are facing, and not the overall strength (half of which is based on their defensive strength, which does not affect the goalie they're going up against)?

Ohashi_Jouzu* · Jan 20, 2013

seventieslord said:
If I understand this correctly, you're using a team's goal differential to calculate their strength?

If the endgame is to use this result to help understand the results a goalie posts against a certain competition level, wouldn't it be better to only look at the offensive strength of the teams they are facing, and not the overall strength (half of which is based on their defensive strength, which does not affect the goalie they're going up against)?

I would tend to agree that offense of the opponent is the greatest obstacle faced by a goaltender. Getting wins against top teams is important, but stopping shots against top offenses is also important. Almost need to "hybridize" those two elements to get an idea of how tough the schedule of a goalie (specifically) was, imo.

Doctor No · Jan 20, 2013

There are two different things that I think would be interesting to know:

Do some goaltenders play a disproportionate share of their games against better opponents (through coaching or random fluctuation)?

Do some goaltenders face stronger shooters than other goaltenders?

The first would require an overall opponent strength (as I've calculated), and the other would require an offensive opponent strength (such as normalized goals scored, or normalized shooting percentage).

I'm working on a technique for the second, although I spent most of my page time yesterday trying to catch up on this year's AHL game logs. Fun Saturday.

NHL Goaltender Strength of Schedule (2010-11 through 2011-12)

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

Registered User

I am lizard king

Registered User

Student Of The Game

Registered User

Registered User

Ad

Ad

Ad