I’ve a multiplayer recreation that, just like chess, can have a win/tie/lose end result in a 1v1 setting.
I have been Elo versus Glicko & Glicko-2 and it appears Glicko-2 may be an excellent one to attempt to implement. Nonetheless, I’m confused on the way to construct Glicko-2 in a scalable manner.
Evidently I’ve to load all video games & gamers into reminiscence with a purpose to calculate the brand new rankings of gamers in a match.
From my understanding, the rankings/deviation are supposed to vary the longer a participant hasn’t performed… From my understanding the algorithm ‘auto’ modifications the deviation for gamers who have not performed.
Is there a form of O(1) approach to apply a glicko score to the 2 gamers of a recreation and perhaps each day run some sort of ‘decay’ on everybody’s rankings confidence (for many who have not performed) in a mathematically correct manner?
I am leaning on sticking with Elo however I do know many video games have scaled Glicko-2, so I wish to attempt to discover a manner to try this.
