Wednesday, November 9, 2011

NHL Alternative Power Rankings

So here I am going over ESPN's latest power rankings http://espn.go.com/nhl/powerrankings and I'm noticing how from week to week some of the teams wildly fluctuate.  Which makes me wonder what the hell good these weekly rankings are.  So over the course of a beer I get thinking "well what if we averaged the rankings, would that give us a better idea than week to week?"  And that led to a lot of paper scratching, some smart phone calculator usage, and the following four (yes four) power rankings.  Below their description:

1) Complete Average - I used all 6 weeks of ESPN's power rankings, including their pre-season rankings before any games were played.  This not only gives a sense of completeness, but also factors in the opinion of better sports minds than me on how teams were being evaluated at the outset based on coaching staffs, talent levels, and before injuries, suspensions, and schedule strength mucked with things.  I added up each team's rankings over the 6 weeks, divided by 6 to get a single score, and then ranked the teams from lowest score (highest rank) to highest score (lowest rank).  So if a team was ranked 1, 2, 1, 3, 1, and 5 over 6 weeks, they would've got a total of 13/6 = 2.17.  This would rank them higher than any team with a higher score.  Some teams divided out to identical scores, so I ranked them as tied in a single ranking spot.

2) Playtime Average - I did the same thing, but excluded the pre-season rankings and divided by 5.  This ranking only reflects weeks in which teams actually played.

3) Recent Average - Same thing, but this time I only used the last three weeks divided by 3.  This doesn't reflect early success, only how teams have been playing the past couple weeks.

4) Weighted Average - I took the 5 week total, then added the three week total in again, and divided by 8.  This still gives teams a benefit for early success, allowing for a team with a hot start to enter a slump they can dig out of, but does weight recent success as a better indicator of how a team will perform right now.

Full Rankings below the jump...



Complete Average Rankings
1 -  Capitals
2 -  Flyers
3 -  Penguins
4 -  Sharks
5 -  Red Wings
6 -  Blackhawks
      Sabres
7 -  Stars
8 -  Kings
9 -  Maple Leafs
10 - Canucks
11 - Avalanche
12 - Predators
13 - Bruins
14 - Ducks
15 - Wild
        Panthers
16 - Oilers
17 - Lightning
18 - Coyotes
19 - Blues
20 - Rangers
21 - Devils
22 - Canadiens
23 - Hurricanes
24 - Islanders
25 - Senators
26 - Flames
27 - Jets
28 - Blue Jackets

Playtime Average Rankings
1 - Capitals
2 - Flyers
3 - Stars
4 - Penguins
5 - Sabres
     Red Wings
6 - Sharks
7 - Blackhawks
8 - Maple Leafs
9 - Kings
10 - Avalanche
11 - Predators
12 - Canucks
13 - Oilers
14 - Wild
15 - Coyotes
        Panthers
16 - Ducks
17 - Lightning
       Devils
18 - Bruins
19 - Blues
20 - Canadiens
21 - Rangers
22 - Hurricanes
       Islanders
       Senators
23 - Flames
24 - Jets
25 - Blue Jackets

Recent Average
1 - Capitals
2 - Stars
3 - Sharks
4 - Blackhawks
5 - Penguins
6 - Flyers
7 - Maple Leafs
8 - Kings
9 - Oilers
10 - Sabres
11 - Avalanche
12 - Red Wings
13 - Coyotes
14 - Canucks
15 - Wild
16 - Predators
17 - Panthers
18 - Senators
19 - Lightning
20 - Rangers
       Devils
       Blues
       Canadiens
21 - Ducks
       Bruins
22 - Flames
23 - Hurricanes
24 - Islanders
25 - Jets
26 - Blue Jackets

Weighted Average Rankings
1 - Capitals
2 - Stars
3 - Penguins
     Flyers
4 - Sharks
5 - Blackhawks
6 - Maple Leafs
7 - Sabres
8 - Red Wings
9 - Kings
10 - Avalanche
11 - Oilers
12 - Predators
       Canucks
13 - Coyotes
14 - Wild
15 - Panthers
16 - Ducks
17 - Lightning
18 - Devils
19 - Bruins
20 - Blues
21 - Canadiens
22 - Rangers
23 - Senators
24 - Hurricanes
25 - Islanders
26 - Flames
27 - Jets
28 - Blue Jackets

This is what happens when you have ENTIRELY too much time on your hands.
10 -

No comments:

Post a Comment