January 14, 2008

South East Division

Below is a table that shows the odds that a given team will be in a certain position in the NHL. The chart is built using all the point rules (tie breakers) and also with the divisional champion rules (as you can see by the funny gap at the 3rd spot, due to the terrible South East). The table is coded to be darker for higher probabilities and thus it is extremely readable. It is sorted by expected points.


What is interesting, is that the entire South East division is predicted to have too few points to be in the top 8 spots, but due to the divisional champion have secured at least one playoff spot. There will likely only be 1 team from the South East and it is possible that the team who makes it doesn't have enough points to make the playoffs (they will still make the playoffs due to the divisional champion rules).

How bad is the South East?
Currently every team is -10 or worse in terms of overall goal differential and the entire division is a -85. Carolina, the current divisional leader, has struggled since mid November and likely wont be able to maintain the nice #1 spot. It looks as though Washington is the one team able to keep its head above water since their lack luster start and could in fact make the playoffs.

A lot will be determined in March (another divisional game month) when these team will play eachother a bunch of times. It's hard to say what will happen.

Either way 6th spot may just be the best place to be in the upcoming Eastern playoff race as a team who would normally qualify for 8th or 9th slides into the 3rd spot to play that lucky 6th place team. Of course this 3rd place South East divisional winner has had quite a bit of success in the past (Carolina and Tampa Bay cup winners). It just seems strange to give such terrible teams that advantage.


Anonymous said...

As a division they have a record of 54-70-14 against teams outside their division with Carolina having the best record at 13-13-3 versus non-divisional opponents. The best team in the southeast division is barely average. It means that the team finishing in 6th spot in the east is likely better off than the teams seeded 4th or 5th and that just isn't right either. Thanks Gary. You're NHL just makes so much sense.

(BTW, I think the NBA addressed this problem a year or two ago by dropping where the 3rd division leader down to 4th if they are below other teams. But that only partially fixes the problem because teams are still competing with other teams for playoff spots based on unbalanced schedules. If the teams in the southeast didn't play so many games against each other and played a tougher schedule like every other team in the NHL, would any of them deserve a playoff spot? Would any of them have above .500 records? Maybe not.)

Mark said...

Interesting stuff. What formula do you use to simulate the outcomes of future games?

JavaGeek said...

[not simple]

CHpct = AdjWin% (on my site) for home team
GFHpg = goals for home team per game
GFApg = goals for away team per game
GAHpg = goals against home team per game
GAApg = goals against away team per game

probability for a game = 1/(1+exp(A))

Using the above probability calculation for every game I simulate the season 1000 times. However, each season I adjust each team by a small amount of random variation [I do this adjustment to deal with lucky/unlucky teams in the first portion of the season]. This adjustment makes the team appear more average (and makes the results more believable).

Anonymous said...

If Tampa Bay, instead of Florida, had gotten Vokoun this offseason, they'd be one of the best teams in the NHL right now.

Earth to the TB Lightning franchise: Get yourself a goaltender for pete's sake!