How to Improve on the NHL Standings

Andrew Perry — Tue, 19 Feb 2008 13:33:08 -0500

A comparison of three systems for Ranking NHL teams

Here is a table comparing the NHL standings they would be if a logical system were used. The first set are the standings after Feb 18, the second uses 3 points per game, and the third is a straight up won lost %. Note that the tie between Bos. , Buff. and Car. Is resolved when Car is moved to their Free Spot.

I think the most interesting thing here is just how little movement there is in the standings. In the East, while there is some churn the playoff teams remain the same and no team moves more than two places: in the West various combinations of PHX, CGY, VAN and COL battle for the last two positions.

The clearest analysis comes from the 3^rd with W/L% only this one eliminates the GP variable, with TO having played 57 and DET at 62 a variance of 5 GP.

conf Team GP Pts Team 3PTS Team w%act

1 OTT 59 73 OTT 128 OTT 0.576

2 NJ 60 73 NJ 120 NJ 0.567

3 MTL 60 73 Pitt 118 Pitt 0.559

4 Pitt 59 71 PHIL 116 MTL 0.533

5 NYR 61 67 MTL 114 PHIL 0.508

6 PHIL 59 65 CAR 112 BOS 0.500

7 BOS 58 64 NYR 106 NYR 0.492

8 CAR 62 64 BOS 106 CAR 0.484

9 BUFF 59 64 BUFF 102 ATL 0.475

10 ATL 61 62 NYI 100 BUFF 0.475

11 NYI 59 62 WAS 96 NYI 0.475

12 WAS 60 62 FLA 96 WAS 0.467

13 FLA 61 60 TB 94 FLA 0.443

14 TB 59 56 TO 90 TB 0.424

15 TO 57 54 ATL 88 TO 0.421

conf Team GP Pts Team 3PTS Team w%act

1 DET 62 89 DET 156 DET 0.677

2 DAL 63 79 DAL 132 DAL 0.587

3 ANA 63 73 MINN 122 MINN 0.576

4 MINN 59 72 ANA 118 SJS 0.525

5 SJS 59 70 SJS 116 ANA 0.524

6 NASH 61 69 NASH 114 NASH 0.508

7 PHX 64 66 PHX 110 CGY 0.492

8 CGY 59 66 COL 108 VAN 0.492

9 VAN 59 66 CGY 104 PHX 0.484

10 COL 64 65 STL 104 COL 0.469

11 CBJ 61 63 VAN 102 STL 0.466

12 STL 58 63 CBJ 102 CHI 0.466

13 CHI 58 60 CHI 96 EDM 0.450

14 LA 62 53 LA 88 CBJ 0.443

15 EDM 60 59 EDM 82 LA 0.403

The worst, for my money, aspect of the present Fractional Point System is that no two teams have seen the same point totals awarded in their games. The league average as of my writing is 2.195 Pts per game, However, Philly has seen 2.119 P/G while EDM has seen 2.30 P/G, a net variance of 8.26% hardly a negligible value. Teams playing PHI for instance will have below average pt totals, while in the NW the pt totals will be seriously wonky, It seem likely that the three NW teams that sit on the cusp are there because of this..

More pernicious is the effect these points have on Management and fans, especially of teams near the bottom, who may put off improvements in the delusion of making the playoffs, and at the top where a team may decide it is too good, and wind up underestimating a playoff opponent. I am sure we can all think of an example of each.

Amusingly as well, if we define .500 as 50% of the points awarded in your games, then that number at any point of the season will be different for each team.

And it still makes more sense than this drawn from my spreadsheet.:

Team

Pts

OTW

OTL

PAWPG

NHL%

Opp%

50%

PHX	64	31	29	66	7	4	2.172	0.516	0.570	69.5
Team	GP	W	L	Pts	OTW	OTL	PAWPG	NHL%	Opp%	50%

Note that we have the absurdity of both PHX and their opponents being above .500 by NHL reckoning. The last column represents the Point total PHX needs to have 50% of the points awarded in their games.

I hope I haven’t bored anyone to death, this is a subject that has bothered me for a while, because I believe that a better , more easily understood standings and stats system will make the game more accessible to new fans, help old fans to a better understanding, and prevent the league from perpetrating any more frauds on a rightly suspicious fandom. Failing all else, the existing system seems to have been built to confuse, to generate false hope in some places, it combines I believe the worst aspect of every other system.

Bleacher Report - Articles by Andrew Perry

How to Improve on the NHL Standings