November 20, 2009
Numbers On Ice
Fatigue Worse Than Playing on the Road?

by Tom Awad

As a statistician, I always like discovering information that has some predictive value. The very first schedule trackers must have shouted “Eureka” a hundred years ago when they first discovered that, all other things being equal, the home team tended to win more games than the road team. Home advantage is a curious phenomenon, especially so since it exists across all sports: in baseball last season the winning percentage of home teams was 54.7%, in basketball 60.8%, in hockey 55.8%. It pays to bet on the home team.

This led me to wonder: are there any other “macro-events” that give a team an edge? While you can always analyze a roster in infinite detail to find small edges, are there any other factors that apply to all teams that could give them an advantage or a disadvantage? The only one that came to mind was fatigue. Hockey teams sometimes play on back-to-back nights, and since this is the exception rather than the rule, these teams should be at a disadvantage against their better-rested opposition. Time to put the theory to the test.

I have defined “Next Day Advantage” (NDA) in the same way as Home Ice Advantage (HIA): a team has NDA if they are not playing for the second day in a row and their opponents are. This means that, unlike HIA, there isn’t always a team that has NDA in a game.

I took every regular season game for 4 seasons, from 2005-06 to 2008-09, and simply counted wins versus games played. Slightly more than a quarter of the games (26.5%) featured a team with NDA. For simplicity, I have eliminated any notion of overtime or shootout losses: you either win or lose, so the average winning percentage should be 50%.

Here are the results:

Situation	          GP	Wins	Win %
HIA pos, NDA pos	1005	1176	58.5%
HIA pos, NDA neutral	3613	3984	55.1%
HIA pos, NDA neg	 302	 302	50.0%

What does this table tell us? More than three quarters of teams with NDA also played at home; put another way, most of the teams playing back-to-back games were doing so on the road. Exhausted and oppressed, they only managed to win 41.5% of their games, while home teams with tired opponents won 58.5% of their games. At the other end, home teams playing for the second night in a row won exactly 50% of their games; playing back-to-back games exactly cancelled out home ice advantage.

I decided to perform the same analysis with goals, to calculate the importance of each factor to the final result.

Factor	                Goal Differential Impact
Home Ice Advantage	0.2638
Next Day Advantage	0.2644

In goal differential, HIA and NDA are of the same size: this confirms what we saw above, that tired home teams lose their home ice advantage when playing back-to-back. In fact, negative NDA is marginally worse than being on the road. We often wonder what, exactly, comprises home ice advantage: adrenaline, last line changes, knowing the rink? Since home teams had an overall advantage of 0.301 goals against their opponents, I now have one portion of the answer: 12.5% of home ice advantage ((0.301 – 0.264) / 0.301) comes from having opponents who are more tired than you.

This is interesting, but does it matter? Surely in the grand scheme of things the number of back-to-back games that a team and its opponents play balance out over the course of a season?

Not quite. Here is a list of teams by total NDA the last 4 seasons:

Team	                 2006	 2007	 2008	 2009	Total
Nashville Predators	 3	 2	 15	 2      22
Colorado Avalanche	 2	 4	  5	 9	20
New York Islanders	 7	 5	  2     -3	11
Los Angeles Kings	 7      -1	  0	 3	 9
Vancouver Canucks	 4	 3	  0	 1	 8
Montreal Canadiens	 1      -3	  4	 5	 7
Pittsburgh Penguins	 2      -4	  3	 6	 7
Tampa Bay Lightning     -1	 2	  4	 1	 6
San Jose Sharks	         0	 5	 -5	 6	 6
Anaheim Ducks	        -1	 2	  6	-1	 6
Calgary Flames	         1	 2	  3	-3	 3
Edmonton Oilers	         2	 5	  0	-4	 3
New York Rangers	 3	 3	 -4	 0	 2
Florida Panthers        -4	 0	  4	 2	 2
Atlanta Thrashers	11	 0	 -8	-2	 1
Minnesota Wild	        -2	 1	 -3	 3	-1
Toronto Maple Leafs	-7	 4	  2	-1	-2
Phoenix Coyotes	         4	-4	  1	-4	-3
Washington Capitals	-2	-2	  3	-3	-4
Columbus Blue Jackets	-3	-2	  2	-2	-5
New Jersey Devils	 0	 0	  2	-7	-5
Boston Bruins	         2	-3	 -5	 0	-6
Philadelphia Flyers	-5	-2	  2	-1	-6
Carolina Hurricanes	-2	-3	 -3	 2	-6
Ottawa Senators	         0	 0	 -6	-2	-8
Chicago Blackhawks	-3	-4	 -3	 2	-8
St.Louis Blues	        -3	-1	 -4	-2     -10
Dallas Stars	        -4	-3	 -5	-1     -13
Buffalo Sabres	        -7	-2	 -6	-3     -18
Detroit Red Wings	-5	-4	 -6	-3     -18

At one end, the Nashville Predators faced tired opponents 22 times more often than they themselves were, more than one out of every 16 games they played. At the other extreme, the Sabres and the Wings had to play 18 more back-to-back games than their adversaries. This shows, among other things, that NDA is not just an artifact of the schedule, since the Wings and Sabres have not exactly been awful teams the last 4 years (for those whose memories don’t go that far back, the Sabres won the Presidents’ trophy less than 3 years ago). I was actually expecting this metric to work against West Coast teams, whose travel schedules are reputed to be horrible, but in face the Kings, Sharks, Canucks and Ducks all came out ahead, indicating that their isolated location could be a benefit as other teams play consecutive games on West Coast road trips.

Twenty-two games with NDA amounts to an extra goal differential of 7: not huge over 4 seasons. However, if I was an NHL GM, I’d try and lobby for my schedule to include games against as many next day opponents as possible. In a league where one point separates 8th place from 9th, that could be the extra little boost my team needs.

Tom Awad is an author of Hockey Prospectus. You can contact Tom by clicking here or click here to see Tom's other articles.