Since Elo number is nowadays used as a synonym for rating, we cannot yet assume that FACEIT actually uses Elo's system (though it is a good initial guess).Įvery player on FACEIT has a single integer rating for CS:GO. The name originates from Arpad Elo's rating system which was originally developed for chess but is now wildly used in online games (usually heavily modified). Player ratings in FACEIT are called "Elo ratings".
#Csgo faceit sign up code#
See the plotting source code for how this is exactly handled. This is one clear source of error, but I'll argue that the effects of ignoring a small set of matches are negligible. in the Team Elo difference plots) and we might not have the rating before match stored for all participants meaning that some matches are unsuitable for analysis. It is often meaningful to use the Team Elo values before the match (e.g. Not all collected matches were used in the analysis. Finally, Mathematica was used to produce the graphs and charts. As seeing the Elo history of a player is a FACEIT "Premium" feature, I've decided not to include the database which undoubtedly contains historical Elo data for many players. The historical source code can be found in the commits of this repository, but I've removed the files from the master branch since the files were not up-to-date and the code was frankly hacky and terrible (do not use as an example). The Python source code for data collection ( faceit.py) and analysis ( plot.py) can be found in this repository. Some of the used APIs are public and some a part of the Data API which can only be accessed with a token requested from the developer portal. This data was then stored in an sqlite database for analysis.Įxtra care was taken to be nice and not to overwhelm the APIs with queries. Knowing who participated in which games, participant ratings at the time of the match could be calculated for each match (FACEIT does not provide this information via their Data API). For each player encountered I queried their profile, collecting the mentioned Elo history after each match. I started by sampling ongoing matches and later polling for their result after they had finished.
This data is the basis for our analysis (Note Chapter 6.1).
thesis on rating systems and having used FACEIT quite a bit, I decided to dig a bit deeper into the rating system they use. Similarly to Valve's own matchmaking system inside CS:GO, FACEIT allows players to search and play 5v5 ranked competitive matches.įACEIT tries to provide fair matchups by estimating players' skills with ratings. In this analysis, we will be concentrating on only one of the supported games, Counter-Strike: Global Offensive (CS:GO). IntroductionįACEIT is a freemium matchmaking platform with support for multiple games. An analysis of the rating system in FACEIT (CS:GO) 1.
0 kommentar(er)
