He's right on. Start at 10/1,000 and move in 2-3/1,000 at a time to see if any significant difference.

I'd not recommend below approx ~3/1,000 unless you work up a load very deliberately engaging the lands. You'll be fighting manufacturing tolerances at that point.

Two asides: Are you using a Hornady/StoneyPoint (or equivalent) gauge to determine seating length for the rifle at these distances? And are you using a bullet ogive comparator (rather than just vanila OAL) to measure seating?
Hornady