Smart Recording is just a way to reduce file size by logging data only when it changes. So if your heart rate remains steady at 150 bpm for 5 seconds, it’ll only write one data point instead of 5. It has no effect on GNSS accuracy, distance, or any other statistics. Activities with Smart Recording will look identical to those with every second recording, just with fewer data points and a slightly smaller file size.
I tested the two settings awhile back and found that even with smart recording, GNSS points were still recorded every second. I assume this is because the precision is such that each reading is always going to be slightly different. The only real compression I could see was in the other stats like HR, pace, and elevation.
Even so, I still turn on “every second” recording, because why not?