-
I'd say it's meaningless unless all of that data came from the same climb. As you say, your mean max curve is what tells you what you have produced at every point in time. That's a far better measure of performance gains over time (assuming you have ridden at max effort for the durations you are interested in comparing).
If you are determined to use VAM, find a climb that you can test on repeatedly and just use those data Setswana create a chart (using tags).
Although, I suppose it becomes meaningless when you consider that you don't smash it up every hill... or is that what meanmax accounts for? It does seem to be possible, expression is
meanmax((_elevation - shift(_elevation,1)) / deltatime * 3600)
Gives:
1 Attachment