The principles of Soccerratings

Ratings are generally regarded as numbers that express playing abilities, performances, etc. Correctly calculated, they are absolutely objective. Soccerratings therefore mirror the performances of football teams.

We decided to basically build upon a method developed by Prof. Arpad Elo for the purpose of chess. The so called "Elo rating" has proven to be a reliable system to determine the strength of chess players from beginners to grandmasters. It is used by the World Chess Association (FIDE) and superseded all other methods (See: Interesting links).

Basically, the difference between the Elo ratings for two opponents results in a percentage telling which winning expectation either player has got. So, the better your opponent is rated, the more points you will earn for a win or draw; and if you lose, you are not going to lose too much of your rating. Of course, the same goes in return for a player who is weaker than you are. The exact formula is rather complicated; those of you who are interested to know more about it may find it in the list of links.

As chess players we do know the plus and minus of this method for almost 20 years now. The Elo rating is generally a very reliable number. There is a high correlation between the expectation and the actual performance of the players.

Unfortunately, this is not true if you have to face a player whose actual good form does not match his/her current rating. This rating will lag behind his top form. Luckily, for poor players the same goes in reverse; that’s good for you, but the new calculation of both of your ratings will not be 100% accurate.

These deviations are very often seen in football. Teams who take 1st place in the table when the season starts fall back, but finally finish on a spot guaranteeing a ticket to Europe. Others find themselves in the relegation zone after three rounds but finish top at the end of the day.

One year ago we created Soccerratings to make these form changes measurable as early as possible. In order to do so, we made the Elo rating method dynamic. Based on statistics, we do also include the further form development of previous opponents in our unique calculation formula. Because, if the team you beat one round ago wins against the leader afterwards, this subsequently underlines your performance. So, there is a built in correction of this "lag behind" phenomenon that was mentioned aboved.

Later on, we adjusted this dynamic and advanced rating system considering the experiences we made so far. We also created a computer programme which makes it available for a great number of football leagues. By using this tool it can be possible to pick those few games out of 100s which offer the best perspectives to make a profit from.