Back when we thought the new divisional alignment was going to come into play, no team was happier about the proposed change than the Dallas Stars. How do I know that? Because they said so: “no team can be happier about the change than the Dallas Stars.”
They also said:
“By far, the biggest issue the Stars face by playing in the Pacific Division is time zones. Dallas is the only team in the league that plays over 80% of its divisional road games two times zones behind. That translates into too many western trips that create difficult travel (to and from) during a long 82-game schedule. No other team in the league faces the travel challenges that are inflicted on Dallas from a mileage and time zone perspective.”
This is probably true. The Stars go on to show that they are closer to six other teams than they are to any divisional rival, and that they often have to stay an extra day when travelling home. That last part is true for anyone coming back from the west coast, but the Stars deal with that more often than anyone else (or so they claim, but I don’t doubt they do).
The Canucks have also brought in “a sleep consulting firm” to assist in minimizing the rigours of travel, so clearly this is an ongoing issue for teams who travel frequently, like Dallas or Vancouver (or others).
But what effect does this travel actually have on a jet-lagged team’s results?
Checking into Jet Lag
As far as I know nobody has studied this over a large number of NHL games. Some have studied it in baseball (see here and here), but hockey only has to fit 82 games in a season, not 162, so there is generally more rest time. It’s been a topic of study in football, too, and others have looked at some NHL results (here, which found almost no effect, and here). But we’re going to take a complete view here, and look at all NHL games played, including playoffs, since the two California teams joined in 1967, through the end of the 2011 playoffs. (Big thanks to Hockey-Reference.com for their list of games going back that far.)
The first step is to assign every game to a time zone, and to make this significantly easier, we’ll ignore daylight saving time. Most of the time DST won’t matter in terms of relative time zones travelled, but I should point out that we may misidentify teams travelling across time zones during the clock switch (in March/April or October/November), as well as those travelling to or from Arizona during DST.
Once that’s done, for each team’s game, we’ll figure out how many time zones they crossed, if any at all, since their last game, as well as how many days they have had off between games.
We’ll then look at all games involving a jet-lagged visiting team and a rested home team. “Jet-lagged teams” are those who played their preceding game at least two time zones away, and have not had the generally accepted “one day per time zone” rest period to adjust. So Thursday in New York and Saturday in Los Angeles counts; Wednesday and Saturday does not. “Rested teams” are those who played in the same time zone, at least two days ago.
For example, on February 28, 2011, Detroit played in Los Angeles, and two days earlier they played in Buffalo, three time zones away. The Kings, meanwhile, last played at home on the 26th. So this game counts: the Wings are considered jet-lagged, and the Kings are considered rested. These are not very common games: there have been just 416 of them since the 1967-68 season and before the end of the 2011 playoffs, or just under 10 per year.
If the (tired) road team generally does worse in these games, we can say there might be an observable jet-lag effect.
Of course, home teams generally tend to perform better anyway, so to control for home-ice advantage, we’ll also look at a control group of games involving two rested teams — that is, two teams who both last played in the same time zone, at least two days before.
The second group, all games featuring two rested teams, is much larger as you’d expect: 13,933, or 324 per year.
Here is a summary of results from those games (all from the perspective of the visting team):
Road Home N Win% GF GA GoalDiff Jet-lagged Rested 416 .427 2.91 3.36 -0.45 Rested Rested 13,933 .437 2.93 3.33 -0.40
(Winning percentage counts ties and shootout results as half-wins and half-losses, and all other overtime results as full wins or losses.)
Not much of an effect, is there? The additional 10 points of winning percentage that the jet-lagged teams give up (.437 minus .427) is about 0.06 goals in the current goal-scoring environment, which fits with what we see in the goal differential numbers (0.05).
One goal for every 20 games is not that big of a deal, especially when nobody plays more than a few of these games per year. The most any team has had in any one season is four, which last happened to Anaheim in 2008-09. So according to these results, that scheduling cost them … well, one-fifth of a goal.
Actually, that effect is even smaller than it looks. We’re assuming the visiting team and home team are generally of equal quality in each group but that’s not necessarily true. And in fact it isn’t. We can use a simple team ranking — SRS, or literally “simple ranking system” — to control for the quality of the two teams in each game, to isolate further on the jet-lag effect.
Going back to our February 2011 example, the Red Wings’ SRS was +0.34 that year (the link says +0.27, but I’ve included playoff games too), and the Kings were +0.17. That gives Detroit a 0.17-goal advantage right off the top, so we expect the goal differential to be +0.17 in their favour, ignoring home-ice advantage. They actually won by three, 7-4, so that game goes down as +2.83 goals for Detroit.
Doing that for all the games and averaging out the results gives us this:
Road Home N GoalDiff Expected Difference Jet-lagged Rested 416 -0.45 -0.05 -0.40 Rested Rested 13,933 -0.40 -0.01 -0.39
This means in games where the home team plays a jet-lagged visiting team, the visitors are generally a little bit worse anyway (about 4 goals over the course of a season), regardless of their recent travel patterns. And the jet-lag effect now looks like almost nothing: .39 minus .40 is, truly, too small to care about.
We should point out that, when teams travel four time zones or more (e.g., Europe back to North America) and play a team who did not, they have been outscored 77 to 53 in 25 games (ignoring shootouts), or by a goal per game. We can’t draw conclusions from just 25 games, but that result certainly won’t have teams falling over themselves to sign up for the next Europe trip.
– Back to Dallas for a second. You may have noticed that the above definition of jet-lagged teams ignores those travelling home, so we’re not considering any games where the Stars have stumbled, tired and exhausted, back into Texas, and have barely been able to hold practices. According to them, anyway.
So let’s see how much that hurt them. Dallas joined the Pacific division in 1998-99. Since then they’ve played 33 home games following a Pacific Time road game, with no more than one day off between games – exactly the situation they describe. In those 33 games, they won 23. They also tied or went to a shootout in three, leaving seven losses. They outscored their opponents by 1.4 goals on average.
In other words, these games hurt them so much, they won or tied 79% of them.
– It is well-established that some teams face higher travel burdens than others. And there are cost concerns to the extra travel, both in time and money, as the Stars’ PR department pointed out in great detail. And playing divisional games at late hours is bad for TV ratings. And so on.
We can’t see an effect here, though, in terms of the on-ice results for that team’s first game after crossing multiple time zones. Granted, we’re looking at only 400 or so games and not taking into account anything more advanced than the final score, and we’re also measuring travel distance in terms of time zones crossed, not miles flown. Still, overall, if there’s such a competitive disadvantage to playing these games, we would see some effect at the macro level.
But we don’t. We see no effect. (We see something happen when teams cross an ocean, but nothing in terms of the usual travel.) Perhaps, after those regular ET-to-PT trips, teams think they’re more tired than they are, but when they get on the ice, muscle memory, adrenaline, and their combined decades of aerobic training and hockey experience take over.
Whatever the reason, it doesn’t appear to be the case that regular coast-to-coast travel significantly reduces a team’s chances of winning its first game on that road trip.