That would be ‘frame by frame’....so-to-speak?
Not quite, but still bad. Prolly more info than you wanted, but the key to stabilization is finding one high-contrast feature that's visible throughout the clip. Then the software can stabilize around that one feature. Unfortunately, the videographer moves around so much that there are rarely more than a few seconds at a time that can be stabilized. You'd have to break the original out into a bunch of subclips, then stabilize the subclips one at a time, then edit them back together. Make sense?