NHL Power Rankings – February 8, 2006

Last night Minnesota beat Los Angeles 5-1 and today these two teams are the big movers in our Power Rankings as they switched positions. Minnesota rose 4 places from 14th to 10th while Los Angeles fell 4 places from 10th to 14th. Others movers include Buffalo who rose from 8th to 5th and Carolina who fell from 3rd to 6th.

Rank Last Week Team AdjWinP SchedStr Power Rank
1 1 Ottawa 0.698 0.508 0.718
2 2 Colorado 0.562 0.542 0.640
3 4 Dallas 0.600 0.517 0.623
4 5 Calgary 0.591 0.522 0.618
5 8 Buffalo 0.651 0.484 0.611
6 3 Carolina 0.667 0.473 0.611
7 6 Vancouver 0.580 0.524 0.610
8 7 Detroit 0.664 0.481 0.603
9 9 Edmonton 0.518 0.532 0.566
10 14 Minnesota 0.509 0.525 0.548
11 12 Nashville 0.589 0.484 0.544
12 11 Toronto 0.509 0.512 0.537
13 13 San Jose 0.500 0.519 0.528
14 10 Los Angeles 0.500 0.520 0.527
15 15 Philadelphia 0.582 0.463 0.511
16 17 Anaheim 0.500 0.514 0.509
17 18 Montreal 0.454 0.522 0.505
18 16 Phoenix 0.456 0.531 0.499
19 20 Boston 0.464 0.510 0.487
20 19 Tampa Bay 0.536 0.467 0.485
21 22 NY Rangers 0.545 0.456 0.466
22 24 New Jersey 0.482 0.475 0.454
23 21 NY Islanders 0.426 0.494 0.435
24 23 Atlanta 0.455 0.470 0.420
25 25 Florida 0.411 0.476 0.387
26 26 Columbus 0.364 0.503 0.368
27 27 Chicago 0.343 0.500 0.340
28 28 Washington 0.324 0.485 0.318
29 29 St. Louis 0.283 0.497 0.281
30 30 Pittsburgh 0.241 0.495 0.248

AdjWinP is a teams winning percentage when shootouts are considered ties and there are no points awarded for overtime losses
SchedStr is an indication of a teams relative difficulty of schedule
Power Rank is the teams expected winning percentage if team played all .500 teams

This article has 20 Comments

  1. Wow, I’m shocked Carolina is in 6th place. Why the big drop?! I mean, they are the top ranked team in the East and the first team in the entire NHL to reach 80 points. I mean come on, they should be at least in the top 5…

  2. First, they are in a virtual tie with Buffalo (both have Power Rank of .611). Second, they have played an easier schedule than any of the teams above them and in fact the weakest of any of the top 14 teams. Third, in interconference play, the west showed that it is a better conference overall this the western conference teams get an edge. Finally, Carolina dropped from last week because they split 2 games against weaker opponents (loss to NJ and win over Boston).

  3. You give a higher weight to the stats that are more current. Basic sports mathematical models. Based on models by Massey, Sagarin, Dunkel ect…. Which BTW, you should read Masseys paper he wrote. Excellent information. A bit dated but the info hasn’t changed any.

  4. You said you would post more detailed information on your program. I for one would love to see it. How about blowouts? Do you use a Bayesian Correction method to take care of things like that? This has given me an itch to update my program. The last rewrite of my program was in the late 80’s when I implemented the Correction methods.

    Hell what launguage is your program written in. I started out Cobol, then basic for the home machine.

  5. I don’t give higher weight to more recent games. I only look at season totals. I’ve considered weighting more recent games more highly but generally I don’t think that gives you better results except when trades or significant injuries are involved (see Columbus re: getting back Nash and trading for Fedorov).

    I have browsed through some of Massey’s work and while his approach is much more statistically involved, I am not convinced that he gets better results. His top 10 teams are:


    The only team in his top 10 that isn’t in mine is Nashville. I have Nashville 11th and Minnesota 10th. He has Minnesota 14th.

    So, does Carolina deserve to be in the top spot ahead? I am not convinced that they should. I am not sure if you caught it but last week I posted an article about beating the best teams. At that time Carolina had 42 of their points against the bottom 10 teams in my power rankings while Ottawa had just 20, Colorado 11, Dallas 18, Calgary 14 and Buffalo 31. Carolina had just 11 points against teams in the top 10, while Ottawa had 16, Colorado 27, Dallas 26, Calgary 27. Buffalo had just 6.

    As for blowouts, I don’t even consider them. I don’t even look at goals scored or goals against as I don’t believe they are the best indicator of a teams ability to win (or not win). I only look at a teams won-loss record and difficulty of schedule.

    My program is written in C.

  6. One more thing. Compared to Massey’s system mine is super simple. If his is a 10 on the complexity scale, mine would be a 2 or 3. I could write an article describing my algorithm fairly quickly but I want to fully explain why I am doing things the way I am, why I am not looking at goals for/against, etc. That is what will take the most time as I would like to do some more concrete research in this area as well.

    I think what my algorithm does more than Massey’s or Sagarin’s is that I put more weight on difficulty of schedule. Dunkel seems to weight recent games much more significantly (hence Ottawa 8th, New Jersey 7th, Columbus 18th, St. Louis 22nd and Los Angeles 26th). Dunkel also makes game predictions and it would be interesting to see what his success rate is.

  7. Ok, you’ve pretty much told me everything what I need to know. BTW give C++ a shot. You will make things smoother for yourself.

  8. What does C and C++ mean? Is it something that Los Angeles citizens aren’t supposed to know about or something?

  9. C and C++ are programming languages. The program I use to calculate my predictions and power rankings is written in C.

  10. Colorado isn’t the 2nd best team, nor do they have the 2nd best goaltender. They were 2 years ago, but you could make a good team core out of kariya, selanne, forsberg, foote, etc.

  11. My Power Rankings:

    1. Ottawa
    2. Detroit
    3. Carolina
    4. Buffalo
    5. Nashville
    6. Dallas
    7. NY Rangers
    8. Vancouver
    9. Calgary
    10. Philadelphia
    11. Colorado
    12. New Jersey
    13. Anaheim
    14. San Jose
    15. Minnesota
    16. Tampa Bay
    17. Los Angeles
    18. Edmonton
    19. Boston
    20. Toronto
    21. Atlanta
    22. Florida
    23. Montreal
    24. NY Islanders
    25. Phoenix
    26. Columbus
    27. St. Louis
    28. Washington
    29. Chicago
    30. Pittsburgh

  12. Again, I refer you to the article I wrote last week on beating the best teams. At that time the Rangers had just 10 points against the top 10 teams in my rankings and 42 points against the bottom 10. Much of the Rangers success has come against the worst teams in the league. Beating Ottawa is a positive but Ottawa hasn’t beaten anyone not named Pittsburgh in their last 6 games and haven’t looked that great in most of those games (including portions of the Pittsburgh games). Should the Rangers be placed in the top 10 when they have only beaten a top 10 team a few times? I don’t think so. In my opinion, to get into the top 10 you have to beat top 10 teams.

    Atlanta is 1-0 against Ottawa, should they be a top 10 team? Of course not.

    The Rangers have played the easiest schedule in the NHL. The Avalanche have played the most difficult and thus their relative positions.

    New Jersey is playing better and have been and will continue to move up the list. The problem is that I only look at what a team has done the whole season thus a few teams get penalized because of injuries. New Jersey (Elias) and Columbus (Nash) are two such teams. But for the most part I think teams are fairly evaluated.

  13. Point taken. However, Western conference records could be a little high for some of the better teams in teh conference, as there isn’t as much of a logjam as there is in the Eastern conference. IMO, the West isn’t as deep as the East. So, the Rangers could be playing teams that might be better relative to the Avs, even though the records are lower from being Eastern conference teams.

Comments are closed.