Yeah, re-reading my post it's very handwave-y. However, a point you made about more good than bad / more bad than good stuck out to me. I wonder if a survey question "On a scale from 0 to 10, where 0 is every thing that happens to you is a bad thing, and 10 is every thing that happens to you is a good thing, what number would you give to your life?" would provide scores correlated with life satisfaction surveys. (Ideally we would simply track people and, every time a thing happened to them, ask them whether this thing was good or bad. Then we could collate the data and get a more accurate picture than self-reporting, but the gain doesn't outweigh the sheer impracticality so I'll be content with self-reported values).
I feel like if it correlated weakly, you would be right. And now that I think about the experiment, I'm fairly convinced it would come out correlated.
From Mike Darwn's Chronopause, an essay titled "Would You Like Another Plate of This?", discussing people's attitudes to life:
Conclusion, graphs, and references in article. As usual, I recommend reading Chronopause.com as Darwin has many good articles; to quickly link a few: