Algorithms

There's really one main algorithm - the one that works out suggestions for games a user may enjoy. The current algorithm assumes that games are rated on a scale of 1 to 5. The current algorithm is also massively unoptimized. It is likely that additional data structure will be necessary to store/memoize tag scores for each game in the system.

- tag scores start at 0
- for each game the user has rated
    - increase score for each tag by the user's rating of the game (double scores for games marked "completed")
- for each game in the system that shares a tag with the user's tags and that the user has not already played
    - increase score for that game by tag score
- return top N highest scoring games

Problems with this algorithm:

• Speed - it's probably very slow