Dec 072005

David Singer over at The Ice Block has posted some interesting numbers about how the goals per game in the NHL has dropped over the first couple months. As you may know, I have been keeping track of this stuff as well and have a few interesting tidbits to add. While I don’t dispute his numbers I don’t necessarily think that it means production in the NHL is on a downward spiral.

The chart below shows a comparison of how the goals per game production has changed over the first 408 games of the NHL season and how it compares with the 2003-04 season (the two reddish lines, 2003-04 being the lower one). The green and blue lines represent 20 game averages (i.e. the average goals per game of games 1 thru 20). (Note: shutout goals not included)

As you can see, there has been substantial fluctuation in the 20 game averages with this season fluctuating a bit more than 2003-04. Around December 1st we saw the 20 game average hit a season low of around 4.5 goals per game. This is actually quite low, even for 2003-04 standards. The goal drought that caused this ran from Nov. 27 to Dec 1st and then quickly popped back to more normal levels that past several days. This short-term drop off would have a reasonably significant impact on David Singer’s lower goals per game averages for November and December.

The reality is that that the goals per game average for the season is slightly higher now than it was from October 16-19 (bottomed out at ~6.05). There does seem to be a very slow downward trend from a peak of around 6.37 goals per game around November 3rd to 6.14 where it is today. Where things settle out is yet to be determined but I think claiming a downward trend, especially a significant one, might be a bit premature.

Nov 142005

A little over a week ago I posted a ranking of the most entertaing games and teams in the NHL. That post generated a fair amount of interest so I decided to create a new and improved ranking system – Entertainment Index Version 2.0.

I believe that fans want to see 3 things in a hockey game:

1. Lots of offense and scoring
2. Physical play and intensity
3. A close game

With this in mind I created the following point scoring system to evaluate every game played in the NHL so far through Sunday, November 13th. Here is the scoring system.

2 points for each goal scored
1 point for every shot over 50 shots (2 team total)

Physical Play
1 point for every hit above 30 (2 team total)
2 points for every fight

Close Game
35 points for a game ending in OT or a shootout
30 points for a 1 goal game
25 points for a 2 goal game
20 points for a 3 goal game
15 points for a 4 goal game
10 points for a 5 goal game
5 points for a 6 goal game

Since watching to see if a goalie is going to get a shutout, I have also given a bonus 2 points for a shutout.

Using that forumla the most exciting games so far are:

1. Boston 4 at Toronto 5 – Oct 24 (shootout) (Gamescore = 103)
2. Toronto 5 at Ottawa 6 – Oct 10 (shootout) (Gamescore = 97)
3. Los Angeles 4 at Detroit 5 -Nov 9 (overtime) (Gamescore = 86)
4. Florida 4 at Montreal 5 – Nov 1 (overtime) (Gamescore = 85)
5. Boston 3 at Montreal 4 – Oct 18 (Gamescore = 85)
6. NY Islanders 3 at Montreal 4 – Oct 22 (Gamescore = 83)
7. San Jose 7 at St. Louis 6 – Oct 8 (Gamescore = 82)
8. Philadelphia 5 at Ottawa 3 – Oct 30 (Gamescore = 82)
9. Montreal 3 at Ottawa 5 – Oct 27 (Gamescore = 82)
10. Boston 7 at Pittsburgh 6 – Oct 8 (overtime) (Gamescore = 81)
11. Chicago 3 at San Jose 4 – Oct 15 (Gamescore = 81)

And the least entertining games:
1. Tampa Bay 6 at Atlanta 0 – Oct 20 (Gamescore = 22)
2. Detroit 6 at Columbus 0 – Oct 22 (Gamescore = 22)
3. Los Angeles 0 at Phoenix 4 – Nov 3 (Gamescore = 25)
4. Carolina 6 at New Jersey 1 – Oct 15 (Gamescore = 27)
5. Atlanta 9 at Carolina 0 – Nov 12 (Gamescore = 27)

Next I calculated team scores by averaging the game score for all the average game score for each game the team played in. The following table shows the results.

Rank Team Entertainment Index
1 Toronto 64.5
2 Ottawa 62.3
3 Montreal 61.8
4 Boston 61.6
5 Philadelphia 61.2
6 Florida 57.9
7 NY Islanders 57.5
8 Dallas 56.6
9 Pittsburgh 55.9
10 Phoenix 55.5
11 Buffalo 55.2
12 San Jose 55.2
13 Calgary 54.2
14 Washington 54.0
15 Los Angeles 53.9
16 Anaheim 53.4
17 Carolina 52.9
18 Chicago 52.8
19 Vancouver 52.6
20 St. Louis 52.1
21 Colorado 52.8
22 Nashville 51.9
23 New Jersey 49.9
24 NY Rangers 49.6
25 Detroit 47.9
26 Edmonton 46.8
27 Atlanta 46.6
28 Minnesota 46.6
29 Comumbus 45.1
30 Tampa Bay 44.7

The northeast division is clearly the most entertaining division with the top 4 teams being from the northeast. Make sense because games between Toronto, Ottawa, Montreal and Boston have all been fairly high scoring and generally close. Plus these teams are among the most physical teams in the NHL.

(Statistics used were gathered from browsing box scores at and as well as fight data at

Nov 042005

A few days ago I was watching the Tampa-Atlanta game and noticed how little hitting there was in that game. That got me to thinking if this was normal for Tampa (and it was). Well over the past few days I went through every box score on and extracted the hits each team had in every game and I came up with the following results.

-Detroit had the fewest hits in a single game with just 2 hits in a game against Columbus. Pittsburgh and Atlanta were next with just 3 hits in a game.

-The fewest hits in a game by both teams is 10 which happened 3 times. Pittsburgh(3) vs New Jersey(7), Phoenix(4) vs Anaheim(6) and Columbus(4) vs Edmonton(10)

-The most hits in a game by a single team is 37 done by the Islanders (vs Montreal) and Toronto (vs Philadelphia)

-The most hits in a game both teams combined is 65 done in 2 games. Florida(34) vs Toronto(31) and NY Islanders(37) vs Montreal(28).

-Toronto is the most physical team averaging 24.4 hits per game and surprisingly Montreal is the next most physical team with 21.5 hits per game. Toronto is a big team so that makes sense, but Montreal is a smaller team so it is surprising to see them average so many hits per game.

-The least physical teams are the Tampa Bay Lightning with just 8.4 hits per game and the New York Rangers with just 10.7 hits per game.

-The team least likely to be hit is the Pittsburgh Penguins. Opponents only hit a Penguin an average of 9.4 times per game. I guess no one wants to hit Lemieux or be the guy to give Crosby an injury. Or maybe teams just felt sorry for them with the bad start they had and didn’t want to kick them while they were down.

-The team most likely to get hit was the Montreal Canadiens. Opponents hit a Montreal player on average 21.2 times per game. Next was the Ottawa Senators who were hit 20.7 times per game. It seems having speed doesn’t mean you can avoid hits.

-There doesn’t seem to be any correlation between playing physical and scoring goals or winning. Some good teams throw a lot of hits (i.e. Ottawa – 20.6 hits/game) and some good teams don’t throw a lot of hits (i.e. Carolina – 11.5 hits/game) and some good teams are right in the middle (i.e. Detroit – 15.1 hits/game).

The table below shows each teams goals for, goals against, total goals, hits for, hits against and total hits on a per game basis. The data is sorted by hits per game.

Comparison of goals and hits per game by NHL teams.

Home/Road game strategies

 David Johnson, NHL, Statistical Analysis  Comments Off on Home/Road game strategies
Oct 072005

This is my first, of hopefully many, article involving more in depth statistical analysis. I’d like to hear everyone’s thoughts on it and what sort of things they would find interesting to look at. I am working on devising a statistical method to accurately predict which team will win an upcoming game as well as a Power Rankings formula, both of which I hope to unveil in a few weeks time. As part of those projects I have tried to identify what teams win games. Do defensive minded teams win, or do offensive minded teams win. To do this I analyzed all the game results of the 2003-04 season and came up with the following results.

Home Team Record 589-470-171 .548 win%
Team with better offence 634-414-166 .591
Team with better defense 637-414-170 .591
Team with better offence
and defense
444-224-101 .643
Home team has better offence 351-180- 76 .641
Home team has better defense 349-178- 84 .640
Road team has better offence 283-234- 90 .540
Road team has better defense 288-236- 86 .543

A team was said to have better offence if they, over the course of the full season, had more goals for than their opponent. A team was said to have better defense if they had fewer goals against than their opponent.

What surprised me is how little it matters whether you have the better offence or better defense. So long as you are better than your opponent at one of them, you have a good chance at winning. Of course, being better both in scoring goals and stopping them is the best. Also, playing at home is a significant advantage too.

General hockey wisdom is that teams frequently play differently at home than on the road. At home they prefer to play an offense oriented game to put on a show for their fans while on the road they play a more disciplined defensive style game. The fact that home teams score about 10% more goals than road teams would seem to back this up, but it could also be due to the extra fatigue/stress caused by traveling and living out of a suitcase. Whatever the reason, home teams score more goals.

In the above table we used season long (82 game) goals for and goals against data, but what if we use home and road (41 game) goals for and goals against data. i.e. when determining which team has the best offence we look at the home teams goals for in games played at home and the road teams goals for in games played on the road. When we do this the results look quite different.

Team with better offence 656-381-166 .614
Team with better defense 539-514-170 .510
Team with better offence
and defense
284-146- 61 .641
Home team has better offence 446-253-110 .619
Home team has better defense 91- 20- 7 .801
Road team has better offence 210-128- 56 .604
Road team has better defense 448-494-163 .479

Now those are some dramatically different results. The shocking thing is, if you can play good defence at home or good offence on the road, you improve your chances of winning dramatically. Also, it appears to be a big mistake to play defensive on the road.

The five teams with the fewest goals against at home Dallas, San Jose, Detroit, Colorado, and Ottawa. The five teams with the most goals on the road are Colorado, Ottawa, Toronto, Tampa Bay, and Vancouver, all very good teams with very good records. All of those teams finished with 100+ points except Dallas which ended up with 97. Also, Detroit and Dallas had the two best home record and San Jose finished fourth and Ottawa fifth.

Conversely, the five teams with the fewest goals against on the road are New Jersey, Boston, Calgary, Vancouver and Minnesota. The five teams with the most goals for at home are Detroit, Ottawa, NY Islanders, Tampa Bay and Buffalo. Two of those teams didn’t make the playoffs and two others had fewer than 95 points.

Moral of the story: Playing offensive at home and defensive on the road isn’t necessarily the best strategy.

 Posted by at 5:23 pm