This page gives a detailed view into the inner workings of the Zero-K rating system. The graph at the top shows the rating over time. Below you can find some more detailed information about current rating, ladder position and rank progress of the selected players.
Ratings consist of an expected skill and variance. The lower the variance, the more certain the system is of the assigned skill. Variance starts at infinity for new players and generally decreases as the number of played games increases. They are calculated using a system called Whole History Rating.
Ladder ratings use a special lower boundary estimate. For active players with a very low variance, the assigned skill is used. For less active players with higher variances the ladder rating will be lower than their true rating. If the variance exceeds a certain value or the player has been inactive for too long, they are removed from the ladder and considered unranked.
Player ranks, visualized as colored chevrons and stars next to their name, will change color depending on their rating. The rating required to reach a new rank is automatically updated so that each rank contains a set percentage of players. Thus, the rank color is a fairly direct indicator of a player's ranking percentile. The rank is calculated for every rating category, but only the best one is displayed.
Zero-K is a free real time strategy (RTS), that aims to be the
best open source multi-platform strategy game available :-)