I don't think I really understand my dilemma either, no prob.
I think downhill skiing is anaerobic, but I'm getting similar spikes during x-country, which is more aerobic.
I think the front end of the exercise data is 'noisy.' I don't know whats going on there, but I suspect I may have been dropping because I was getting little increase in BG after eating quite a lot.
Cereal was 45g, 6 unit Humalog, at 9m, but I blacked out the data because I didn't have sensor readings to go alongside it.
Diet soda.
12-15 grams may increase me about 60 points. So, about 4 or 5 points per gram.
I:C is 2 units per 12-15 grams of carb.
Basal is working O.K. If I'm stable during the day, with 35 units of Lantus, I might take 6 units of correction mid-morning (around 10am) and will tend to run low in the evening. Its remarkably stable overnight - it might vary less than 10 points. Until I wake up and it starts to rise, requiring a mid-morning correction bolus - 6 units on a good day, 12 units on a bad day.
If I lower the Lantus to 34 or 33, I run much higher in the morning, but will stabilize in the eave. If I do that, then I might wake up around 3am and bolus an extra 4 to 6 units to counteract the highs that develop.
I don't know what you consider appropriate for basal behavior, but I am have certainly identified a greater need for basal insulin between waking and 2pm. A pump might have to come to bat for varying basal requirements, especially when I add exercise into the mix.
It gets real bad when I patrol at a mountain with all double and triple black diamond runs. When I started doing that, three or four years ago, I really started to see how volatile the system becomes during exercise. I think I need to be a better diabetic, now, in order to be a better athlete. God help me if the ski patrol has to come to my rescue. I would rather eat a cyanide pill in the woods.