(It's not the computation itself that's the problem. The problem is that it causes a lot of stuff to get loaded from disk that would otherwise not be.)
I just went through and (roughly) calculated my average. I am just over the threshold. Which creates an interesting scenario: A couple of stupid comments from me and I will fall below the threshold.
That creates a really great motivation.
So, if you add this feature back, it might be worth it to run some reports to see if there is a "sweet spot". Not to mention it would be really neat to see the distribution of average comment scores.