Even more stat geekery... Thanks to Kyle, I started thinking about a possible way to do a pythagorean based on goal differential. In baseball, the Pythagorean expected value for a team is calculated by
(Runs Scored)^2
-------------------------------------------
(Runs Scored)^2 + (Runs Allowed)^2
That works for baseball because the average team is at .500. However, as has been discussed, a winning percentage is impossible to calculate accurately, because teams are awarded for overtime losses, meaning not every game has an equal weight. So, I calculated the average number of points per game to be approximately 1.125, or 9/8. So, I decided to come up with a slight variation on pythagorean with the average PPG factored in (which will, obviously, fluctuate from week to week and even year to year). To calculate the pythagorean estimated points for a team, then, I went with the following formula:
(Goals Scored)^2
------------------------------------ * (Total Possible Points) * (Average PPG)
(Goals Scored)^2 + (Goals Allowed)^2 ...with the total possible points being twice the number of games played (i.e. a win every game). That produced what I'll call Pythagorean Points (PYPs). This was interesting, but still a little misleading, because there was a variation on the number of games each team played. So, for example, the Hawks PYP was 33.9, while Vancouver's was 36.2, but Vancouver had played 3 additional games, so on average the Hawks were still doing better. So, I took it a step further, and calculated the Pythagorean PPG (PY-PPG) in an effort to gauge the effective point rate of each team for comparison. Very similar to the prior formula:
(Goals Scored)^2
------------------------------------ * (Average PPG)
(Goals Scored)^2 + (Goals Allowed)^2 So, finally, I ran PYP and PY-PPG on every team, and calculated the difference between the expected and actual values (a negative value for Diff means they are under-performing their expectation, and a positive value means they are over-performing their expectation), and then sorted the teams in order of PY-PPG (the expected point rate based on pythagorean) to come up with the Adjusted NHL Standings:
Eastern Conference
Teams Division Games Points GF GA PPG PYP Diff PY-PPG PPG-Diff
Boston Northeast 27 42 94 59 1.556 43.6 -1.6 1.614 -0.059
Pittsburgh Atlantic 27 34 87 75 1.259 34.9 -0.9 1.291 -0.031
Montreal Northeast 26 35 79 69 1.346 33.2 1.8 1.276 0.070
New Jersey Atlantic 24 30 70 63 1.250 29.8 0.2 1.243 0.007
Philadelphia Atlantic 26 32 84 78 1.231 31.4 0.6 1.208 0.023
Washington Southeast 28 33 88 86 1.179 32.2 0.8 1.151 0.028
NY Rangers Atlantic 30 38 77 76 1.267 34.2 3.8 1.140 0.127
Ottawa Northeast 25 25 63 63 1.000 28.1 -3.1 1.125 -0.125
Buffalo Northeast 27 29 74 79 1.074 28.4 0.6 1.052 0.023
Florida Southeast 27 27 67 76 1.000 26.6 0.4 0.984 0.016
Toronto Northeast 28 26 84 98 0.929 26.7 -0.7 0.953 -0.024
Carolina Southeast 28 29 71 83 1.036 26.6 2.4 0.951 0.085
Atlanta Southeast 26 21 77 93 0.808 23.8 -2.8 0.915 -0.107
NY Islanders Atlantic 27 22 70 94 0.815 21.7 0.3 0.803 0.012
Tampa Bay Southeast 27 20 63 88 0.741 20.6 -0.6 0.762 -0.022
Western Conference
Teams Division Games Points GF GA PPG PYP Diff PY-PPG PPG-Diff
San Jose Pacific 27 46 102 64 1.704 43.6 2.4 1.614 0.089
Minnesota Northwest 26 31 72 56 1.192 36.5 -5.5 1.402 -0.210
Chicago Central 25 31 90 73 1.240 33.9 -2.9 1.357 -0.117
Detroit Central 26 40 95 80 1.538 34.2 5.8 1.316 0.222
Vancouver Northwest 28 33 86 74 1.179 36.2 -3.2 1.293 -0.114
Anaheim Pacific 28 33 82 79 1.179 32.7 0.3 1.167 0.012
Calgary Northwest 27 33 80 80 1.222 30.4 2.6 1.125 0.097
Los Angeles Pacific 26 26 69 72 1.000 28.0 -2.0 1.077 -0.077
Edmonton Northwest 26 28 74 78 1.077 27.7 0.3 1.066 0.011
St. Louis Central 26 27 76 81 1.038 27.4 -0.4 1.053 -0.015
Nashville Central 27 30 77 85 1.111 27.4 2.6 1.014 0.097
Colorado Northwest 27 27 72 80 1.000 27.2 -0.2 1.007 -0.007
Columbus Central 27 25 76 86 0.926 26.6 -1.6 0.987 -0.061
Phoenix Pacific 27 26 69 79 0.963 26.3 -0.3 0.974 -0.011
Dallas Pacific 26 24 72 93 0.923 21.9 2.1 0.843 0.080
Interesting data, but take it with a grain of salt, as I currently have no idea how much this fluctuates from day to day, let alone week to week. However, according to the data: - Minnesota is the team most under-performing their statistics, by a wide margin over the Blackhawks and Senators - Detroit is the team most over-performing their statistics, by a wide margin over the Rangers.