The weight of charges for a given "click" setting varies from measure to measure, and also by operating technique, which is why the system is of limited value for giving other shooters precise load data. IMO this should be done by weight. On the other hand, it is a very handy system for the individual shooter, where the same measure is in use, particularly since Lynwood Harrell started making measures that have six divisions between numbers.
Obviously the weight change per detent varies with the density of powders. On my measure, with the technique that I use, one detent is worth about .09 gr of 133 and about .1 for the LT powders.
I determine this by throwing ten charges into a scale pan at one setting, say 50, weighing it, and then at another setting say 55, and weighing that. Then I subtract one from the other and in this example divide the difference by 300 (five clicks time six divisions times ten charges) to get the value of one detent.
New users to the system may find the reference to clicks a bit confusing, but all they need to remember is a click is a change of one full number. With the my Harrell measure that is set up for sixths of a click, I note increasing charge weights as follows. 50, 50 1/6, 50 1/3, 50 1/2 etc.
Using my measure and 133, thanks to Gene Beggs, whose tunnel testing has identified the charge weight difference between nodes for a typical LV rifle shooting 133 as 1.2 grains, I know that if that is the move that I want to make that I will need to change my measures setting by around two clicks, and that if a two shot test, shot in really good conditions, has paper between the holes, that a one click change is likely to put me pretty close to an accuracy node center.