It's not normally shown. I got it directly from support, in response to my request during a short term discussion of a "???" problem. (I offered to comment on the behavior, as a professional programmer- they'd possibly found previous comments along the lines of "this isn't good programming; you SHOULD do this in an upgrade...." to be valuable.) It's proprietary to them, so I won't splatter my graphs across the Internet without approval. But the graphs which support can generate for themselves, after you upload the data, do show this.
I can say that you were smart to "clear off" a bad and recent calibration entry or two which had gotten into the list for that Sensor. ALL of the entries for the same Sensor are being used, but the old ones are given less "leverage" by giving more weight to the newer ones. That's the big software change which makes calibration of the new "plus" work better, causing fewer erroneous "???" periods which can't be fixed, and providing better accuracy. (I TOLD them to do that in a microcode update after seeing the data. Probably, they'd probably already figured figured out the idea before I advised them to do it-- but maybe my independent suggestion of the same idea played a role in moving it along.)
As far as I can tell, ALL of the entries are used since the last "restart", and a tiny bit of your own "personal history" gets used too. But older entries are given less "weight". The original Seven used them all equally, leading to a very bad curve (and less accuracy, more "???" displays) as the Sensor became older and the voltages fell off.
Since it worked a lot more smoothly right after after restart, I think that Rita has the right suggestion: give it an extra 6-12 hours of "warm-up" before inserting the Transmitter and starting the 2-hour countdown. The extra warm-up time avoids a lot of the first day craziness. Every calibration which is entered during "craziness" distorts the curve, and this prevents the need to enter anything. You're still running on the old one.)
I also do this for nearly every new Sensor, shooting them in the night before and starting the warm-up around 9 AM on the following day. (That timing keeps dawn effect and morning bG craziness out of the calibration period, it's all over by the end of the two-hour warmup.) works great!