Thursday, September 15, 2005


Have they?

If you don't get the bi-weekly Western Standard magazine, one of the handful of things you're missing is the sports column by Edmonton's own Colby Cosh. Low on outrage, high on original thought - basically the polar opposite of most sports opinion. (And when you do get outrage, it's unadulterated).

I bring this up because the concluding paragraph of this issue's column, on the prospects of the three western Canadian NHL teams this season, got me thinking a bit:
Surveying the whole picture, I make the order of finish for this season to be Calgary, Vancouver, Edmonton [sounds about right to me - ed.]. The really good news is, they might just end up the top three teams overall in the conference. Stranger things have happened.

My instinctual reaction to this is "not bloody likely", but have stranger things happened? I think the answer is probably Yes - you make the call.

The new NHL schedule is fairly unbalanced: each team plays the other 4 teams in their division eight times (32); the other 10 teams in the conference four times (40); and 10 teams in the opposite conference once each (10), for a total of 82 games. (This year, Northwest Division teams play a home game against each team in the Northeast, and visit each team in the Atlantic; the only way the Flames and Lightning face each other this season is in a Cup rematch).

The important number here is the 32 games in-division, which have a zero-sum outcome**. 32/82 = 39% of games are against division rivals. Is it possible for three teams in one division to lead the conference under these conditions? Let's try some dreaded research.

In 1997, four teams from the NFC Central made the NFL playoffs - there was the 3 NFC division winners, and the other three NFC playoff teams were all from the Central division. Back in the 30-team, 6-division NFL, teams played 8/16 games (50%) against their own division.

In this year's National League East, all five teams have been above (or at least around) .500 for the bulk of the season, despite the fact that 76/162 games (47%) are played against each other.

It's hard to find relevant precedent in the NHL. For example, in 97/98, the top three teams in the West were all from the Central Division; however, there were only two divisions then, not three, and the schedule was not as unbalanced then as it will be this season.

The sense I get browsing around (and from memory) is this: good teams will be good, regardless of schedule balance. If there are several good teams in a division, they will split up most of the intra-division wins, and make a lot of hay with the non-division portion of their schedules.

But to end up with the result Cosh proposes, that's just not enough. A good team in another, weaker division will dominate its intra-division schedule, and pile up too many points for the 3rd team in a stronger division to match.

That said, this is not an easily searchable query. If anyone is aware of any historical precedent that would suggest that the Flames, Oil, and Canucks could have the top 3 point totals in the West this season, I'd be damn interested to learn about it.

And there is one final point to be made here: for this to have any chance of happening, Colorado and Minnesota will have to be bad, bad teams, at least to the extent that they are dominated by the Canadian teams (much like the Bears went 4-12 in '97, 1-7 against their four division rivals who made the playoffs). I'm not at all prepared to write off the team that's won 8 division titles in the past 9 years, let alone Jacque Lemaire's boys. In fact, it feels very wrong, right down to my cockles, to stand behind any prediction that places the Avalanche 4th in their division.

**Footnote: the point for the OT/SO loss (i.e. the "3-point game") makes mincemeat out of the numeracy underlying the zero-sum concept. However, since it doesn't do it in any foreseeable way (we can't, nor should we, predict that games the Flames, Oilers, & Canucks play with each other will be tied after regulation), it's still roughly useful as a concept.

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?