FanPost

The Caps' Slow Start, PDO and Identifying Lost Opportunities

This is the first in what I hope will be a periodic analysis throughout the season of the league's standings relative to each team's All-Strength PDO. Much of my work on this subject is simply an application of the analysis done by Patrick D. (SnarkSD) over on nhlnumbers.com using this season's statistics. I have collected the required data from the official NHL website to compute the data presented below.

A quick refresher on what PDO represents:

PDO is computed as the sum of save percentage and shot percentage. Since the league average of this sum has to be one, it provides an easy reference point to compare a particular team's result and how they match up to an imaginary NHL average team. Much work has been done to calculate PDOs for individual players and teams both for an individual game and for a season. Many people refer to the statistic as a representation of "luck" in that there is plenty of variation in the number from game to game that appears to be related to just chance as opposed to skill. The work done by SnarkSD referenced above breaks PDO down into how much of the statistic is related to chance and how much is apparently related to skill. The net result of the research is that on small enough samples, chance can easily overwhelm the skill aspect. It is only over the long NHL season that the skill factor begins to catch up, in terms of influence, on a team's overall PDO.

It should be noted that many other articles and posts about PDO refer to Even Strength PDO (namely the shooting % and save % related only to even strength conditions on the ice). The work done by SnarkSD shows that the Even-Strength PDO actually is influenced more by randomness than the All-Strength PDO. Since this article is about team PDO (not player PDO) and is also interested in projections over a whole season, for the remainder of this article, when PDO is mentioned, it should be assumed to be All-Strength PDO unless specifically called out otherwise.

One of the most interesting finds in SnarkSD's analysis is the following plot:

(Click on the plot for a larger version from nhlnumbers.com)

What this chart is telling us is how likely a particular team's PDO is high/low just due to chance. As the number of games played increases, the likelihood that the team PDO is where it is due to chance and skill can be determined by examining the plot above. For example, after 30 games, a team that has a PDO higher than 1.04 (or less than .960) is most likely to have earned that PDO through skill (or lack thereof) and not to have obtained it through chance (but if they had a PDO of 1.04 after just 10 games, it is possible that it was that high just by chance).

Where do we stand at this point in the season?

It is VERY early in the season but I wanted to get things set up so I could track the PDOs over the course of the season. My results for the league (as of 10am on 10/12/2013) are as follows:

Team Name

All Strength
PDO

Z-Score

Current
Standings Pts

Adjusted
Standings Pts

Expected
Standings Pts

Actual minus
Expected

ST LOUIS

1.094

2.1497

6

6

5.8

0.2

COLORADO

1.078

2.2079

8

8

7.2

0.8

SAN JOSE

1.075

1.9970

8

8

7.1

0.9

NY ISLANDERS

1.051

1.3356

5

5

6.2

-1.2

DALLAS

1.046

1.0957

4

4

4.5

-0.5

TORONTO

1.043

1.3779

8

7.5

7.4

0.1

TAMPA BAY

1.034

0.8755

6

5

5.6

-0.6

PITTSBURGH

1.033

0.9074

6

6

5.6

0.4

BOSTON

1.024

0.5870

4

4

4.0

0.0

MONTREAL

1.019

0.5159

4

4

5.1

-1.1

COLUMBUS

1.015

0.3599

4

3.5

3.7

-0.2

CALGARY

1.010

0.3098

8

9

6.0

3.0

WINNIPEG

1.008

0.2290

4

4

5.9

-1.9

ANAHEIM

1.005

0.1335

6

5.5

4.6

0.9

OTTAWA

1.001

0.0353

4

5

3.4

1.6

VANCOUVER

0.996

0.1308

6

5

5.4

-0.4

PHOENIX

0.994

0.2030

6

6

5.3

0.7

CHICAGO

0.992

0.2145

5

5.5

4.2

1.3

CAROLINA

0.990

0.3247

6

7

5.1

1.9

LOS ANGELES

0.974

0.7677

6

4.5

4.4

0.1

DETROIT

0.971

0.8040

4

3.5

3.4

0.1

WASHINGTON

0.970

0.8488

2

1.5

3.4

-1.9

NEW JERSEY

0.969

0.8612

3

4.5

4.2

0.3

PHILADELPHIA

0.967

0.9987

2

2

4.1

-2.1

BUFFALO

0.963

1.1256

1

1.5

3.9

-2.4

FLORIDA

0.958

1.2747

4

4

3.7

0.3

MINNESOTA

0.958

1.0642

4

5

3.0

2.0

NASHVILLE

0.946

1.4706

2

2

2.6

-0.6

EDMONTON

0.942

1.5836

2

1.5

2.4

-0.9

NY RANGERS

0.919

2.2292

2

2

1.6

0.4

The first column is the PDO for the team in question. It is computed from the teams current shooting percentage and save percentage thus far in the season.

The second column is a Z-score. This is computed as a ratio of that teams deviation from the norm (1) to the known standard deviation. The higher the Z-score (namely over 2), the more likely their PDO is where it is due to factors other than chance. For example, St. Louis and Colorado have high PDOs that are sufficiently high to suggest that they are better than the average team and not just incredibly lucky. Similarly, the NY Rangers poor PDO is also likely due to poor play and not a lot of bad luck.

The third column is the team's current point standings.

The fourth column is adjust point standings as discussed by SnarkSD in his posting. Namely, anytime a team gets into overtime, they are awarded 1.5 standings points. This removes the question of lucky streaks in overtime and shootouts which can have an undue influence on their standings points.

The fifth column is what a team's standings points should be if they are following a nominal win/loss path for their current team PDO. This is calculated using the coefficients provided by SnarkSD.

Finally, the last column is the difference between where the team currently is in the standings compared to where they should be. My interpretation of this column indicates whether a team is outperforming/underperforming their expected results for their PDO.

What does this mean for the Caps?

Well, so far, the Caps have basically missed an opportunity for an additional 1.5 points. They are currently 2 points under where they should be based on their PDO (however, they won the shootout against Cagary which saved them a half-point). I think most of us agree that they were close to winning at least one more of the games they have played thus far. The bad news is that when I looked at underperforming teams in previous seasons, it is rare for the lost points (what I refer to as "missed opportunities") to be made up in significant numbers as the season goes on. The good news is that at the moment, we are talking about a single game which could be made up by getting fortunate bounces. But if the trend continues and the team continues to underperform their PDO, historically it appears that a significant number of standings points rarely gets made up.

Summary

This is a work in progress and the season is still very young. I hope to write up additional FanPosts throughout the season with updated PDOs to see how the Caps (and others) are performing in the standings relative to where they should be. I am more than open to comments and suggestions! I am not well versed in statistics so if anyone notices errors in my analysis, I am more than welcome to hear their inputs.

If this FanPost is written by someone other than one of the blog's editors, the opinions expressed in it do not necessarily reflect those of this blog or SB Nation.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Japers' Rink

You must be a member of Japers' Rink to participate.

We have our own Community Guidelines at Japers' Rink. You should read them.

Join Japers' Rink

You must be a member of Japers' Rink to participate.

We have our own Community Guidelines at Japers' Rink. You should read them.

Spinner

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9355_tracker