A few days ago I looked at how some of the top offensive stars in the Southeast Division had scored inside their division versus outside their division. What I found was that, to a man, they were more effective scorers inside their division than they were against other teams. This wasn’t an especially surprising discovery; there has been significant anecdotal evidence to suggest that players from the Southeast have it easier than players elsewhere in the league, at least when it comes to putting up offensive numbers.

However, looking at five top scorers certainly wasn’t enough to prove anything. I’ve seen it argued that divisional games are “clutch” games and that the stars naturally perform better in them. A more rational argument would be that a sample of five players isn’t big enough to prove anything.

With that in mind, I decided to expand my sample to include every player who played more than 10 games inside the Southeast division last season; a total of 95 players. Then I took the totals and divided them down to an 82-game season for the sake of simplicity. Here’s what I found:

Position Div. G Div A Div PTS ODiv G ODiv A ODiv PTS % Chg.
Forward 20 29 49 17 24 41 -16.3
Defence 8 22 30 7 20 27 -10.0
All 16 27 43 14 23 37 -14.0

That’s pretty definitive; I looked at just under 100 players of all different types and found that on average their performance dropped by 14.0% when they played outside the division. There were some exceptions last season, notably Ilya Kovalchuk, Vincent Lecavalier, and Mark Recchi, but they were the outliers and in the case of both Kovalchuk and Lecavalier they’ve had their numbers improve inside the division for most of their career.

What does it all mean? For general managers, I think the implication is obvious: don’t acquire a player from the Southeast and expect him to match the offence he posted in that division. The reverse holds true for Southeast G.M.’s; players acquired from outside the division can be (as a rule) expected to exceed their results from outside the division.

Comments (9)

  1. For completeness I recommend doing this comparisson for every division. Although likely that the SE division is weaker perhaps being more familiar with a team makes it easier to score against them?

  2. Trenton: That’s why it’s Part I. I’m already confident off the answer, but other divisional numbers will be coming over this next week.

  3. With the reduction in divisional games from 32 to 24, and the fact that teams play 20 games against the other divisions in their conference, combined with the evidence from your post on the best division in the league, is it not more of an issue of conference rather than division?

  4. Coaltrain: Yes and no. The Conference imbalance is an important one, but because the league doesn’t seperate the results in and out of conference, it’s a hard one to get a read on. Meanwhile, divisional imbalance still matters; 24 games represents nearly one-third of the schedule.

    But it’s a good point you’re making, I just don’t know how to measure it.

  5. Right, but a SE team would play 24 games against other SE teams, while a NE or Atlantic team would play 20 games against SE teams. That’s only four extra games that a SE division player would play against SE teams. How much is four more games going to affect a player’s stats over the course of an 82-game season?

    Your mention that general managers shouldn’t acquire a player from the SE an expect him to match the numbers he posted “in that division.” Does the player being “in that division” really matter that much when the significance of being there only means four more games against SE teams?

    Since he’s the guy at the top of your post, let’s take Eric Staal’s 75 points last year. Staal posted 27 points in 24 games (1.13) against SE teams and 48 points in 58 games against non-SE teams (.82PPG).

    Let’s apply Staal’s divisional and non-divisional point-per-game rates to what would be his schedule playing for an Atlantic or Northeast team–20 SE games and 62 non-SE games. You wind up with 73 points. A decrease of two points.

    So while Staal did see a huge increase in points-per-game while playing SE teams, it really didn’t matter much in comparison to players not playing in the SE as he only got an extra four games against SE teams.

    That’s what makes me scratch my head when I see all these posts about how much SE players benefit from playing “in the SE”. It’s four extra games!

  6. Kahz: I suppose really, what we’re measuring isn’t the effect of the division, but the difference between a only-against-dvision schedule and a 40/18 conference split schedule with the rest of the league.

    Thanks for your comment – it’s made me think of this in a whole new light.

  7. I agree with some of these other comments. It’s not about the division, that’s a small part of a mostly balanced conference schedule. I would look at that angle, although each conference will wax and wane. A good GM knows a good player. I remember back in the 80′s PIttsburgh would play these stiffs with Lemieux and then parlay them into really good players in trade. That’s what it’s about.

  8. The more I think about this, the more I feel that I’m approaching it from the wrong angle. There’s definitely a division imbalance (SE < NE < ATL = WC) but the real imbalance is West vs. East and that’s what I need to try and find.

    Thanks for the feedback everyone – I will do more on this, although it’s going to be time intensive so that post could very well be delayed for some time.

  9. I know I’m a bit late on this one but I was wondering how you came to your stats. Did you average the contribution of each player (i.e. there were 95 players so you pro-rated each guys’ “in-divison” and “out-divison” stats to 82 games and then took the average) or did you just add up the totals (some guys played 20 games, some 40, some 60 and the more games you play the more your contribution is valued in the totals) and then pro-rate those to an 82-game schedule.

    Also, and this sounds a bit silly, but couldn’t you just add up the total goals-against for each Conference/Divison to come up with an estimate. So the divisions stack up like this for last season:

    Central – 1156
    Northwest – 1173
    Atlantic – 1183
    Pacific – 1185
    Northeast – 1207
    Southeast – 1261

    So if a player played all of his games evenly distributed against teams in the Southeast division you’d expect him to have 5.5% more offence than if he played the games, evenly distributed, against teams in the Atlantic and Northeast. Similarly, a player who plays his games, evenly distributed, against only teams in the Eastern Conference should have 3.9% more offence than a player who did the same against the Western Conference.

Leave a Reply

Your email address will not be published. Required fields are marked *