Measured (without the IPhing) is taken the first time from peak to peak, the second time from baseline to peak, the third time from an arbitrary point to peak. In the third case a "peak to peak" stretch is not easy to identify.

You seem to always measure "Actual" from baseline to peak.

From what I see, "actual" can be reliable, "measured" (with iPhone) can give any result you want, depending on where you place the beginning of the exposure, which judging from the video seems quite arbitrary.