A research paper on the Microsoft website gives a fascinating insight into the way that gamers will be ranked on the new-style Xbox Live.
The paper outlines some existing ranking systems, noting that the ELO ranking system is used in Chess rankings and football rankings. However, ELO ratings are not designed to work where games have more than two 'teams' or players. Video games, of course, can often have 16 or more 'teams' in a deathmatch scenario.
Consequently, Microsoft have come up with a ranking system they call 'TrueSkill' which uses Bayesian inference to rank players. The inference system uses a bell-curve (come on all you high-school mathmos!) and then calculates the mean and standard deviation within that. The ranking system thereby assumes that the 'skill' of a player in the system can be thought of as the average performance of the player over a number of games, and that the likelihood of deviation from his previous form can be calculated.
Players' ranks can be altered within the game by comparing their performance to others around them in the end of game rankings, and then examining how that relates to their rank.
There is some absolutely brilliant maths in there, and one of the equations is shown in the picture above. All in all, it's fascinating to see just how much work goes into something that most will simply take for granted.
Check out the paper, online, here. Give us your thoughts on the system in the News Forum.
The paper outlines some existing ranking systems, noting that the ELO ranking system is used in Chess rankings and football rankings. However, ELO ratings are not designed to work where games have more than two 'teams' or players. Video games, of course, can often have 16 or more 'teams' in a deathmatch scenario.
Consequently, Microsoft have come up with a ranking system they call 'TrueSkill' which uses Bayesian inference to rank players. The inference system uses a bell-curve (come on all you high-school mathmos!) and then calculates the mean and standard deviation within that. The ranking system thereby assumes that the 'skill' of a player in the system can be thought of as the average performance of the player over a number of games, and that the likelihood of deviation from his previous form can be calculated.
Players' ranks can be altered within the game by comparing their performance to others around them in the end of game rankings, and then examining how that relates to their rank.
There is some absolutely brilliant maths in there, and one of the equations is shown in the picture above. All in all, it's fascinating to see just how much work goes into something that most will simply take for granted.
Check out the paper, online, here. Give us your thoughts on the system in the News Forum.
MSI MPG Velox 100R Chassis Review
October 14 2021 | 15:04
Want to comment? Please log in.