Player Rankings

The rankings primarily use the TrueSkill rating system developed by Microsoft Research. Here's how it works:

• Every player starts with a skill estimate (mu) of 25 and uncertainty (sigma) of 8.33.
• TrueSkill uses Bayesian inference to estimate true skill levels.
• The ranking score is calculated as mu - 3*sigma (conservative estimate).
• As players play more matches, uncertainty decreases and ratings become more accurate.
• TrueSkill handles team games and varying match qualities better than ELO.

This creates a sophisticated ranking system that accurately reflects player skill!

We also maintain ELO ratings for comparison. Here's how it works:

• Every player starts with a rating of 1000.
• After each match, points are transferred from the losing team to the winning team.
• The amount of points transferred depends on the expected outcome - upsets result in bigger point swings.
• Teams with higher ratings are expected to win against teams with lower ratings.
• When a higher-rated team beats a lower-rated team, they gain fewer points.
• When a lower-rated team beats a higher-rated team, they gain more points.

ELO provides a simpler, more traditional ranking approach.