| 摘要 |
Existing player ranking methods, for example, those using simple score-summation ranking, often suffer from problems related to reduced player satisfaction and players becoming dis-incentivised. We address this by determining a first indication of the relative skill of players of a game, updating those indications fully in the light of a game outcome to find a second indication, and then determining a revised indication of the relative skill being between the first and second indications. The indications of relative skill are related to belief distributions associated with belief about skill of players. The revised indication is determined using a level of update which may be adjusted to incentivise players and for other purposes. |
