NHL Power Rankings – Feb. 1, 2006

These are the first power rankings under my new algorithm which I think is a fair bit more accurate, and possibly more controversial, than the previous version. The improvements made include:

-Treating shootout wins/losses as ties. The shootout does not accurately reflect the ability of a team to win or lose games (when a skills compention is not required) and as we all know, there is no shootout in the playoffs.
-Overtime losses are treated as losses. No points for them.
-A significantly improved method of accounting for difficulty of schedule.

The result of these changes is that you will find teams in the stronger divisions being ranked higher than under the old system and teams in weaker divisions being ranked lower. The team with the most difficult schedule so far is Colorado while the team with the easiest schedule so far is the New York Rangers. As a result Colorado has moved up in the standings while the Rangers are sitting well below where many people would rank them in 22nd position. The reason is the Rangers have played teams ranked 1-15 just 16 times and in those games they are just 6-10 (and one of those wins was a shootout win). They will have to prove themselves down the stretch when they play Ottawa (3 times), Carolina (twice), Buffalo, Toronto (3 times) and Philadelphia 5 more times. For similar reasons Philadelphia and New Jersey are well below where many would place them.

Rank Last Week Team AdjWinP SchedStr Power Rank
1 1 Ottawa 0.704 0.520 0.755
2 4 Colorado 0.575 0.537 0.643
3 3 Carolina 0.683 0.479 0.637
4 5 Dallas 0.596 0.521 0.626
5 2 Calgary 0.596 0.519 0.618
6 6 Vancouver 0.575 0.523 0.605
7 7 Detroit 0.651 0.480 0.593
8 8 Buffalo 0.640 0.478 0.583
9 11 Edmonton 0.519 0.529 0.565
10 12 Nashville 0.615 0.476 0.552
11 9 Los Angeles 0.509 0.524 0.545
12 13 Toronto 0.500 0.521 0.541
13 10 San Jose 0.510 0.517 0.534
14 15 Philadelphia 0.615 0.460 0.531
15 16 Phoenix 0.472 0.532 0.517
16 14 Minnesota 0.472 0.526 0.513
17 17 Anaheim 0.500 0.515 0.511
18 18 Montreal 0.450 0.523 0.497
19 19 Tampa Bay 0.528 0.473 0.488
20 20 Boston 0.462 0.511 0.483
21 22 NY Islanders 0.441 0.497 0.455
22 21 NY Rangers 0.519 0.456 0.447
23 23 Atlanta 0.462 0.474 0.433
24 24 New Jersey 0.462 0.464 0.414
25 25 Florida 0.396 0.482 0.382
26 26 Columbus 0.365 0.498 0.365
27 28 Chicago 0.333 0.504 0.337
28 27 Washington 0.324 0.485 0.317
29 29 St. Louis 0.270 0.495 0.266
30 30 Pittsburgh 0.250 0.484 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 6 Comments

  1. I’m not sure if there is any obvious way. To really play spreads with the most benefit you really need to look a each teams offense and defense ability. I am planning on creating an offense and defense ranking system but I haven’t done it yet. Maybe in the next week or two I’ll get that done.

    Bowmans, which I am using to bet with, doesn’t have a straight moneyline for the Detroit/St. Louis game (only +/- 1.5) so I am not going to play that game at all. I think I am also not going to play the Rangers/Penguins game since the odds are so low (-450). I’ll probably omit playing games below -400 because the reward is so little and there is always a risk.

  2. Normally, you use power ratings to determine what the actual line for a game should be.

    For example…. look at tonight’s Ottawa vs. New Jersey game. Take Ottawa’s power number and divide it by the Devils power number and multply it by 100.

    So, .755 divided by .414 = 1.82 x 100 = 182 so, Theoretically, Ottawa should be favored by -182 in that game.

    Look at the Calgary game upcoming tonight….

    Calgary .618 divided by Colubmus’ .365 x 100 = 169 so the Flames should be a -169 favorite in that game.. Bowmans has them at -280 so strictly playing overlays and underlays…the Jackets are the play…. HOWEVER, where these ratings appear to be flawed is that there is no home-ice advantage factor weighted into the calculations so, -169 for Calgary is their power edge BEFORE factoring their home ice weighting.

  3. Interesting. I do take into account home/road advantage as well as some other factors (i.e. back to back games, etc.). It might be interesting to see if it betting would return better results if you bet according to your process. I could probably modify my predictions to output the power edge after these additional factors if you would be interested in tracking it.

  4. I’m down David. But also where does the ‘return from a long road trip’ theory (5 games for Calgary) factor into the calculations? I mean one could go on and on trying to calculate the actual line for a game using all the factors involved but it gets unrealistic IMO. The books set their lines for each game and how they arrived at it and what they took into consideration is beyond everyone else’s imagination.

  5. Ok, I’ll look into outputting the power edge rating with the gane predictions but it might not be until next week that I get around to it as I am going to be fairly busy between now and Monday.

    As for how the books set the lines, I assume it has less to do ‘reality’ than it does with what they think bettors will do.

Comments are closed.