Women's International Elo Ratings - FAQ

To view the latest rankings, You can view the ratings here.

What is Elo?

The Elo rating system is a rating system that awards points to teams based on competitive play. Originally designed for chess, Elo has since been adopted for other sports and has been viewed as an alternative ranking system to FIFA's rankings. The general idea is centered around stealing points from your opponents and adding them to your own score. If you beat a strong team, you take a large number of points. If you defeat an inferior team, you take only a few points from them. (Wikipedia)

FIFA has their own ratings for the women's game. Why create Elo ratings?

Most people are familiar with the Elo ratings we use on the men's side and FIFA's women's ratings are generally similar, they just use a different approach. FIFA actually uses a type of Elo ratings, as there isn't a "right" way to do compile them. However FIFA's approach to the women's ratings creates a noticeably different outcome than the method used for the men's Elo ratings. These rankings here are made to have a mirror to the men's rankings, without the headaches that FIFA institutes in theirs.

It should be noted that FIFA men's ratings are done different than their ratings on the women's side. It's not clear why they wouldn't use one system for both. Regardless, I'll refer to these ratings here as Elo because they're more in line with the commonly accepted version.

How are FIFA and Elo different?

In the top rankings, the order of teams are similar from one ranking system to another, although the gap between teams is different. As of August 2016, number ten rated Canada is 230 points behind the US in FIFA's rankings. In the Elo ratings, the gap is almost twice as much at 400, from USA and (#10) North Korea.

Beyond the top teams, there is a large discrepancy between rankings for the lower rated nations. Nations can regularly be rated a difference of 40 spots between systems and as high as 90. This is due to the starting rating, which will affect how many points teams gain or lose when they earn a result. Sometimes FIFA will give starting ratings as high as 2000, while other times they will start teams at 400, like they did with Mauritius. While starting ratings eventually lose their importance after hundreds of matches, most nations have not played a hundred games and the 1600 difference between teams can have major implications on future ratings. Elo starting ratings range from 600 to 1800 and are based on each nation's performance to avoid big swings in ratings.

The specific approach between the ratings are fairly different. For example, FIFA rankings can actually take away points from a winner if they do not win by enough, which Elo does not do. There also isn't a difference in winning 6-0 as there is 40-0 in FIFA's ratings.

Additionally there are odd restrictions to FIFA's rankings on the women's side. First, it only reaches back to 2003, despite having most of all international games on their site. Secondly, there is a stipulation that if a country does not play a game within eighteen months, they immediately drop to the bottom. As I write this, Argentina has recently dropped 99 spots to last place. (Technically they are not rated although they would be 133, after Botswana at 132.) Although apparently this is only temporarily, as Argentina are no strangers to the quick maneuver.

Argentina's FIFA ranking by year

Argentina's FIFA ranking by year

Overall, FIFA ratings favor European, Asian, and South American teams while North American and African teams are rated higher in Elo ratings. FIFA originally rated countries whose national team programs started earlier much higher than those that started late, while many of those older teams have low winning percentages (Estonia at 31%, but 87 spots higher in FIFA than Elo, as of October 2017) and have fallen down in the Elo ratings. Newer teams have not lost as many games and are still rated middle of the road, while FIFA has already downrated them before playing a dozen matches.

What are limitations of these Elo ratings?

Probably the biggest point for discussion is the base rating for a country. Every team has to start with a rating and this can range greatly. On the men's side, Elo ratings range from 1800 (great starting spot) to 600 (very bad). After a few hundred games, the starting point is relatively pointless and would only change a rating 200 games later by peanuts. However for teams that are low on games played, it does have a larger influence.

To give a clearer example, US women's team has played around 600 games, giving them a rating of 2176. Their base rating, before they played any games, was 1600. If I drop their base rating of 1600 to 600 their overall rating only drops to 2161. After 600 some odd games, their rating gets finalized.

An example of a country that hasn't played several hundred games is Equatorial Guinea. They have at Elo rating of 1231 over 44 games with a base rating of 900. Their rating of 1231 puts them 32nd in the world, but FIFA has them at 50st. FIFA has given them a lower base rating than a 900. If we move their base rating to 1200, they move to 1272, 29th overall. So to be honest I am not quite sure why Equatorial Guinea is so low in FIFA rankings. But all that to say, base ratings were tricky to finalize but ultimately not a big deal for countries who have played at least 50 games, which was most teams.

The only other hurdle I ran into was the K factor that Elo ratings use. The K factor weighs each game in importance. A friendly is basically worth one third of a World Cup game and everything else between that varies. (FIFA also weighs games less than Elo.) Again, there's reasoning behind doing so, it's just going to change the outcome if the math is different. (Read more about them Elo game weights here.)

How do the women's Elo ratings compare to the men's?

It is no secret that the men's game has more depth than the women's side and now we can see that more concretely. The top forty teams are separated by 400 points on the men's side, while the women are almost at 1000 points from USA to fortieth ranked Ivory Coast. Even inside the top ten we can see the disparity as USA is 400 points higher than North Korea.

However the women's Elo rankings are designed after the men's so they can be somewhat comparable. The lowest base Elo rating is 600, despite some 40 teams on the women's side being south of 600. The maximum base rating for the women's side is 1600.

How do Elo ratings compare between different generations?

Elo ratings are a system to show how dominate one team is at one given time. We can't directly compare the men's game to the women's as the landscapes are vastly different but even over time, it is tricky to compare the 1999 USWNT to the 2015 squad. Both teams get over 2100 points but it is in reference to their competition. For example, if the rest of the world - for whatever reason - became significantly worse over time, then a 2100 against weaker teams would not be the same as a 2100 when the rest of the world was better. Of course it is tough to say objectively if teams are better or worse over time, but it should be kept in mind when comparing two different periods of time.

How many games do you have in your database?

Currently I have 8677 games that start from April 1, 1920 to July 31, 2016.