I originally started my chart in an April. So instead of doing a year-end for that year since it would require too much guess-work, I instead did a 'First year' chart, tracking from April to March. From that point on, I just followed the traditional calendar year for updates (also do a 'midyear' after the 26th week just to give a taste of what might end up winning by year's end).
I've been using the basic inverse points system. Each week I make my top 50. The #1 song earns 50 points, the #2 earns 49 points...#50 earns 1 point. Then I just total it for year-end/mid-year charts.
The problem with this system is that while is it the easiest and most convenient, it doesn't account for the individual weeks themselves. Not every week is the same strength. Some weeks I can't listen to music as much, some weeks I'm just bored of every song, other weeks I can listen a lot, or others I'm just really into a lot of songs. My #1 song in the middle of March I might not have enjoyed anywhere near as much as the #5 song in the middle of October. But that song would have gotten more points towards the year-end.
Overall, inverse points is simply too cut and dry to accurately show things for me during year-end.
I'll explain the actual system later, it's fairly complicated and I doubt you'd want to carbon-copy it; what I just explained explains the thought process behind it which is what's important.