Did I misunderstand, or was it only a 3dB difference on the yellow meter, for the stock vs. the DonnyFL "Ronin"?
That's in the "mildly perceptible" range.
The yellow meter was completely inaccurate!
Ignore the readings on the yellow meter and observe the readings on the gray meter.
There is approximately a -15db difference between the stock moderator and aftermarket moderators.
How can you tell? His experiment had little control, so there's no way to tell which meter is wrong. Possibly both were wrong.
It looked to me like the yellow reading may have been more consistent, but given the lack of any real controls on the experiment there's no way to have any confidence. This is like trying to predict a trend from two points on a graph.