Generally, Canadian teams are compared within the same league that they play. On occasion they get compared against other select leagues. Last year Brotherly Game, a sub-site of sbnation.com focused primarily on the Philadelphia Union, continued their ongoing ELO ranking of top American and Canadian teams. Unfortunately, the site completely ignores many Canadian 3rd & 4th division teams by excluding Premiere Ligue de Soccer du Quebec (PLSQ), League 1 Ontario (L1O), and the Pacific Coast Soccer League (PCSL). Additionally, many elements, including the Amway Canadian Championships and some regular season games were missed in calculating the rankings. As a result, I started from scratch with results from the beginning of the 2014 season, and calculated the ELO rankings for Canadian Teams.
I’ll get into the math of ELO ratings and the assumptions I made further into the article, but for people just looking
for the rankings: the MLS teams are tops, followed by FC Edmonton of the NASL, and then some surprises happen. The NASL’s Ottawa Fury performed poorly throughout most of the Fall Season, and as a result good performing teams in L1O & PLSQ managed to leapfrog the higher division team.
The Ottawa Fury sit at a distant 12th place, showing just how well the Canadian 3rd division teams rack up against the newest Canadian NASL franchise. They’ll need to do quite well this season in order to pull themselves up. Last year’s champions for L1O & PLSQ sit right below FC Edmonton, which is fitting given their impressive records during the season.
Most of the PDL teams listed no longer exist in 2015. But a whole bunch of new teams will be added including 3 USL, 3 L1O, 2 PLSQ, and 2 USL PDL teams. The PCSL has yet to officially announce their teams for this year, but it looks like there will be 1 or 2 teams added to replace the loss of the teams from Vancouver Island. This should make the ELO ratings jump around a fair bit over the course of the season, and may dramatically change the results for 3rd & 4th division teams. Keep reading Northern Starting XI for monthly updates on team ELO ratings.
If seeing the ratings is good enough for you, I’d recommend you stop here. The remainder of this article talks about how we got these numbers. It’s all math and statistics stuff, so many of you will not be interested in the background.
What is ELO & how is the math done?
In short, ELO is a way to statistically say which team is historically the best. To put it another way, ELO rankings are a way to calculate the relative skill of two teams in a competitive sport. It’s adapted from a method meant to compare chess players, and is used by FIFA for ranking Women’s International Soccer. The difference in ratings helps to determine who would likely win in a match, and in our case a home-field advantage is taken into account. The resulting outcome of the match alters the ELO ratings of the two teams involved, which will change the predictions for the next match of each team.
ELO ratings are considered strong because they use the team’s historical results, the results of each match, goal difference, and the weight of the outcome based on importance and expectations. However, they suffer from a few shortcomings. The importance of the match is determined by the rating designer, as is the starting point for each team. The strength or weakness of a team’s schedule will also make a big difference in the outcome.
The ELO ratings and calculations have been tweaked to try to match it to the intended positions of the leagues. I chose to start my ranking with the 2014 season, because it was L1O’s first season of operation, and PLSQ’s scores are a little spotty prior to 2014. I’ve also taken the assumption that all 3rd Divisions are equal (which they’re not), because the Canadian Soccer Association wants them to be on the same level. Furthermore, “friendlies” are not counted toward ELO ratings since both teams use these matches to test out players they wouldn’t normally use. The starting assumptions are:
To calculate a new ELO, we use the following formula:
Rn = Ro + K * G * (W – (1/10(-dr/400)+1))
Where:
Rn = The new team rating
Ro = The old team rating
K = Weight index of the match
G = Index of goal differences (see chart above)
W = Match result (1 = win, 0.5 = draw, 0 = loss)
dr = Difference rating, +100 for the home team
If you take the time to punch through the numbers, you should end up with results similar to mine. For the purposes of tracking numbers through 2014, I included the US Open Cup, but many of the teams had “untracked” ratings. There are so many American teams in the lower levels of play that never see a match against a Canadian team that tracking them would have been unwieldy and produced a pretty insignificant difference. If you’re looking for perfect numbers, feel free to contact me and I’ll send my spreadsheet to shortcut some of the calculating time and score-hunting. In the meantime, this should suffice for a good & solid accounting of Canadian team ELO ratings as of the end of 2014.