With free agency slowing down (some might question if it ever picked up), attention has turned towards the National Hockey League’s next regular season. We know few details at this time – the raging coronavirus pandemic has predictably created deep uncertainty for the planners within the league.

That said, we can draw reasonable inference as to what the NHL will plan for if we understand the issues at play. The most obvious constraint concerns the international border between Canada and the United States. Should it remain closed, an all-Canadian division would be an option.

Creating that division is easy. The hard part is figuring out how to redraw the remaining divisions, especially if the appetite to enter virus-proof bubbles for a regular season isn’t there.

There are a few ways the league can draw these divisions; much of it will depend on whether or not the league considers regional bubbles (or similar controls of that nature) in order to minimize the risk of virus transmission. If we believe that geographic proximity will be something that dictates the redraw of divisions, we can figure out how to segment the remaining 24 teams.

To create this sort of regionality, I grabbed every airport’s latitude and longitude and created nautical mile distances between every possible American city pair. The below table shows the estimated time between pairs:

Embedded Image

There is a lot of data in there, but this is the launch point for what will end up being our solve – an attempt to create regionality in the most optimal format knowing that seven Canadian teams, spread from coast to coast, will be unable to participate in those divisions.

That is an important restrictor for teams like Buffalo, who in every scenario would be logically paired with their geographic rival in Toronto – a current divisional foe found through a short drive (or a ridiculously short flight) up the QEW.

What we need to do now is strike any possible Canadian team and create only American team versus American opponent scenarios. The table below shows every team’s seven closest opponents – their optimal region, so to speak:

Embedded Image

For most teams, you can create regional divisions that make all of the sense in the world. In our Buffalo example earlier, losing Toronto as a division rival doesn’t mean all that much – there are more than seven possible opponents in and around that region, and a team like Columbus (which currently occupies a slot in the Metropolitan Division) is barely a 40-minute flight away.

This example, of course, hits on another issue: the NHL has a swath of teams in close proximity to one another in the northeastern United States, whereas other teams – think the Minnesota Wild and Dallas Stars – really are standalone franchises. This naturally creates inefficiency in creating the regions. Time to get creative.

If we break away from the current format of 8-8-8-7 (number of teams respective to each division), we can come up with five regions carrying an average travel time of an hour or less:

Embedded Image

Do I anticipate a fully functioning regular season, flush with teams travelling week after week to NHL destination cities? No. But I do think this is the NHL’s best opportunity at redrawing divisions that can be managed on the fly.

If the league is reticent to employ a tightly controlled bubble format (in which case you can draw the divisions up however you see fit), something like the above five-division format with an intra-division schedule is probably the most sensible way to bracket the teams.

It allows the league to consider things like “playing weeks” and “rest weeks” as well – a way to not only control the testing and isolation that’s going to be required for the regular season (which will be of high importance to the league), while creating real travel and logistical flexibility for the players and support staff.

This is one hypothetical solution of many, and the league is surely exploring all avenues right now. But if I know one thing, it’s that the usual boundaries the league has held onto dearly have dissipated into thin air.

Anything goes in 2020. And probably in 2021, too.