ct_the_1 wrote:
What i thought of would be to start at 10 and then minus 1 for every checkbox checked.
Use any amount of checkboxes you want (each one means a problem with the game). Then get the rating this way:
rating = #UNCHECKED_boxes / total_boxes * rating_upper_boundary
eg: Rating is a number up to 10, there are 20 checkboxes and 3 were checked. The result is, (20-3) / 20 * 10 = 8.5
If sometime in the future you decide to add more checkboxes... You can do something like this:
- Instead of marking, in the database, which fields are checked, mark the ones that are not, and do the inverse logic. If a field is marked in the database it means that item is ok, and there's no problem in the game (the user did NOT check the checkbox and you marked the field as true)
- When you add new fields update the database and set them to true (only the new fields) for all the records that were already in the DB.
- The rating may increase slightly, depending on how many fields you add, but it will be coherent for all records since it will mean those problems were not reported by any user (which was obviously impossible since there was no checkbox to report them
)
PS: I think you shoud use a weighted mean instead. Clearly not all problems have the same impact... You may also want to consider some exceptions, like "The game doesn't load at all", in which case the rating should automatically be 0. If it's hard to come up with good weights for each item then at least consider the exceptions. If the game has some problem that makes it unplaybale then the rating is automatically 0, otherwise get the arithmetic mean of the remaining problems.
