Wednesday, February 03, 2010

A Better NHL Standings Representation

The way that we currently represent team won-loss records in the NHL bothers me. Every newspaper and website in the world shows five columns in the standings: GP-W-L-OT-PTS. When mentioning a specific team's record, sources will usually either report merely W-L-OT, or perhaps W-L-OT (Pts).

This is a problem, for four reasons:

  1. The ultimate determinant of a team's success is points, yet it's listed fifth. If the standings also list expanded statistics, like goals against and goals for, then points is buried in the middle of the standings despite being the most important.
  2. Fans are conditioned via other sports to think of a ".500" record as depending only on the W and L columns. Unfortunately, this is incredibly misleading; as I write this, only five teams in the entire league are "below .500," which is of course impossible.
  3. Reporting individual team records as (for example) 27-25-4 is similarly misleading, and if points are included, i.e. 27-25-4 (58 pts.), then this includes redundant information.
  4. Records are not transitive. For example, the Wild is 1-0-1 against Detroit, but Detroit is 1-1 against the Wild. Teams often have a "winning" record against each other (Team A could be 1-0-2, while Team B is 2-1, against each other).
Clearly, the problem here is the representation of the overtime/shootout loss; it is worth one point, a win two, and so we're conditioned to think of an overtime loss as "half a win." This isn't correct; it's a loss with a "loser point" tacked on top.

We need to stop thinking of the overtime/shootout loss this way, and the fastest way to do so is to change how it's represented in the standings. Here's how I'd like to do this.

First, the PTS column in the standings would be given primary importance, and be moved up to second in the order. The L column would include all losses, not just regulation losses, and the OTL column would be eliminated completely (or at least moved away from the main standings.)

Here you can see what I mean, with Tuesday's Northwest standings:


With the same number of columns, we represent more information (how many overtime games has a team played) without losing any information. We correct the problem of inflated records. And we move points, the most important factor, up to the beginning. (NOTE: Some representations shade or otherwise highlight points in their current representation; do that, and I can see them staying to the right of losses, though I feel strongly that they'd have to be to the left of the overtime record.)

With regard to specific teams, I'd like to see W-L (Pts) become the standard. For example, Minnesota would be referred to as 27-28 (58 pts). This, as you can see, does not lose any information - it's easy math to find the number of overtime/shootout losses - but weights wins and losses correctly. Records are shorter, and just as easy to understand.

Heck, if you want, you could invent your own punctuation system. A few possibilities:
58 (27-28)
27-28 -58-

And so on, and so forth.

Hockey is already seen by many Americans as the most confusing of the four major sports. Simplifying and improving the standings is a small step towards ending that confusion.


Old Man said...

All I want to know usually is how many points they're (WILD) out of a playoff spot. Your system makes it much easier to determine. Of course, hockey won't do it because they view themselves as knowledgeable insiders who don't need to make it easier for those 'average' fans to follow their sport.

USAFA Bulldog said...

Brilliant, fax this out to all media outlets. This is the way it should be done.

Now as an aside, I will say that the points (and "division titles") in all the North American major sports leagues is stupid, since all have unbalanced schedules.

Unless all teams play all teams the same number of times, saying one team is better (more points) is stupid, since they didn't play the same teams the same number of times.