Quoted from robin:The more I think about it, the more I lean towards option 2. There are other advantages. We will get proper backglass images showing in our collection, reflecting the model we own.
Option 2 is much better. Option 1 will stir up the database too much without adding any real value and probable unnecessary complicate the user interface.
Quoted from robin:The only thing I'm unsure of is having multiple models of a game in the Top 100. The way the new Stern games are going, we could have 4 AC/DC games in the Top 100.
One game, one top 100 entry, I'd say. You don't want the top 100 to be clouded with what is basically 4 times the same machine. I'd say, take the average, that seems fair. And then it's up to Stern (or JJP) to make sure ALL of the versions are worthy of a good spot on the list, right?
On a technical note, we could relatively easily link the various machines in the machines table using a extra column or join table linking the different models by their key value. I think this way we might be able to change the code to calculate the average on all of the models for a machine without complicating things too much.