In Part 1 and Part 2 of our look at the NHL realignment work that we did in this paper, we explained how to estimate team and league travel without knowing the schedule ahead of time, and we presented some of our best six-division realignment solutions. In this article, we will
Discuss the proposed four-conference realignment with the more balanced schedule (We estimate that this proposal would require the league to travel about 100,000 more miles than our best six-division solution);
Give our top four-conference solutions;
Give solutions for some hypothetical franchise moves (like PHO to QUE) and expansion to 32 teams.
Best four-conference solutions
We now focus on four-conference solutions. The NHL's proposal, made last season, looked like this:
We are able to estimate the travel distances that would be required by this four-conference structure and the more balanced schedule that the NHL proposed to go along with it. This proposal would have required the league to travel 100,000 more miles than with our best six-division solution. Here we compare travel for each team under our best six-division solution and the NHL's four-conference proposal:
Black dots are the estimated miles for the NHL's proposal. The green arrows pointing down indicate teams with better travel (fewer miles) under our best six-division solution. Red arrows indicate teams with worse travel (more miles) under our solution.
Some teams in the center of the US would have worse travel, but more teams would have better travel. In particular, the teams that have the most difficult travel like the West Coast teams, and the Florida teamswould be much better off with our best six-division solution.
Also, the gap between the teams with the best travel and worse travel is smaller with our best six-division solution. If we think competitive balance is important, and that travel affects on-ice performance, then we would want the difference between the best travel and worst travel to be as small as possible. With the NHL's proposal, the difference would be about 30,000 miles, and with our six-division solution, the difference would be under 20,000 miles.
Of course, distance isn't the only consideration, and since the NHL may be set on a four-conference structure, we give our best four-conference solution:
I really like this solution, at least as far as four-conference solutions go, partly because I don't think it is something I would have come up with if I were doing this by hand and without the help of a computer algorithm. One downside for fans of the PHI/PIT rivalry is that those teams are not together in this solution. We note that even our best four-conference solutions require 85,000 more miles of travel than our best six-division solution.
If we include the same constraints as we did before (like forcing PHI and PIT together), we get this solution:
This one looks closer to the NHL's proposal. The main difference is that TB and FLA aren't with the teams from the current Northeast Division. One downside might be that the current Atlantic Division has been split up: PHI and PIT are in the purple conference, and NYR, NYI, and NJ are in the red conference.
Since our methods can be easily adapted to other situations, we thought we would have some fun with some hypothetical situations. Phoenix (PHO) has been in the news seemingly for years because their ability to stay in PHO has been questioned. So we decided to ask our Futurama-style what-if machine a few questions, like "what if PHO moves to Quebec?" The best six-division solution is below.
We used the same constraints as before, except we allowed the orange division in the northeast to have four Canadian teams. We also used "PHI must be with PIT," and that is why PHI isn't with the New York teams and CAR is. It looks kind of silly, so it might be better to remove that constraint.
In our paper, we also give solutions for these hypothetical moves: PHO to Houston (HOU), PHO to Kansas City (KC), and PHO to Seattle (SEA). We also tried PHO to Southern Ontario, which led to the same solution as above, except with Southern Ontario replacing Quebec.
It wouldn't be surprising if the NHL expanded to 32 teams within the next few years, so we thought we would look at a couple of hypothetic 32-team configurations as well. Suppose PHO moves to Las Vegas (LV), and new teams are added in Southern Ontario and Quebec. Suppose the NHL decides to go with eight four-team divisions, like the NFL has now. Our best solution is given below.
Four-team divisions make sense for the NFL because of the way their schedule is structured. They play two games against every divisional opponent, and that would be harder to do with eight-team divisions. But if and when the other leagues expand to 32 teams, eight-team divisions might be preferred. Here is the best solution for the NHL in that case:
This configuration would be pretty bad for TB and FLA. It might be best to force them to be in the east, with DET and CBJ likely paying the price. However, if the priority is to minimize travel for the teams that have it the worst, priority would have to be given to TB and FLA. In the NHL's current alignment, DET and CBJ still have better travel than TB and FLA, despite being in the west.
If you are interested in realignment in other sports, feel free to check out our paper. Our best MLB one is pretty interesting: in one of the cases we considered, we allowed teams to switch from the AL to the NL, and vice versa. Some people would be highly opposed to a radical realignment like that, but it's just a hypothetical situation, so hopefully it can be more fun than controversial.
As for the NHL, we will likely make some updates as the situation in the NHL changes. Hopefully, we will hear about that soon.
Brian Macdonald is an author of Hockey Prospectus.
You can contact Brian by clicking here or click here to see Brian's other articles.