@ned, I absolutely think that your calibration will be cleaner if you wait until the Dexcom and the real BG have been flat for a while. It is very likely that the Dexcom looks forward when it calibrates, but it cannot predict the actual BG curve. The only situation where the Dexcom may be able to anticipate a non-flat situation is if the rise or fall is constant and steady. Even then, though, I would wait.
The only times when we don’t wait are:
upon sensor start (we want to see a signal, even if we know it’s not perfect)
if we know there is a very large error. In that case, it makes sense to reduce it already, and to do an extra calibration later if you can or want to.
As a caveat, a teenager’s BG is so quarely that we sometimes see periods of 6-8 hours where we are unable to calibrate properly
The Dex will ask you based on a strict time schedule. Definitely wait until it’s flat to calibrate. There’s no need to respond instantly to its requests. Personally, I only end up calibrating about once a day on average, and it works great (and I think would be far superior to twice a day calibrations at times when it’s rapidly changing), but that’s based on experience that that works fine for me (I increase frequency if my sensor seems wackier than usual or something is going on to make me not trust the readings as much).
Just for info, I think we only calibrate once for every 2 days or so (through lack of diligence). Doesn’t seem to affect things for the worse at all. In fact, I think our results are better than they were when we were calibrating more often. This is probably due to not always calibrating when things were exactly flat.
