While this is a general rule, I don’t think it applies to narrow band sources.
Here is the CIE (2008) physiologically-relevant luminous efficiency function for photopic vision with an actually existing 594nm FWHM 10nm laser Bandpass filter and the resulting product, which peaks at about 591nm.
I don’t know how it could have a maxima at 500nm or 575nm if there is little or no energy in that range.
Again, none of this is to dispute that something might be “off” with the current DRT candidate.
I also suspect some problems in the Yellows, as Pekka has mentioned that region needed some extra attention in the past and was desaturating more than expected with certain functions.
Just don’t want to target the wrong problem, if something else can explain that problem.