I've read the book and did it with that method (such as adding and subtracting a 100 kg/Ibs) plotting my new ZFW and with the original ZFW and thats how much baggage/passenger i can carry. But some questions add 50 kgs/Ibs and that doesnt make sense to me at all.
You're missing the point with this stuff. The technique is only a geometric means to an end and has no other significance.
If you run the initial sums and find yourself not where you want to be (CG-wise) you have to rework things to fix that problem. Techniques which spring to mind include
(a) algebraically - too much like hard work so let's give that a miss.
(b) trial and error calculation - fine, if you have set it up in a spreadsheet and then just play with the numbers. A real pain if you do this longhand. Except, perhaps, for some final finessing, let's give that a miss.
(c) set it up with bits of transparent graph paper overlays and run the sums graphically. Works fine but a bit of a pain to set up unless you are going to do a heap of exercises.
.. and you can come up with other ways to the solution but they are too involved for routine use and, for the exams, impracticable or not permitted. So we need a simpler way which is reasonably accurate.
You can plot the CG/load on a CG envelope providing that the envelope is weight x moment (or IU) as when we do the calculations longhand we add weights and add moments (or IUs). It follows that we can do the same, graphically, on a chart presented as weight x moment (or IU). Mathematically, this technique is incorrect if the chart is presented as weight x CG but the error usually is small and we can get away with it if we can accept a small error at the end or run a longhand check calculation.
The technique is based on the fact that, for a constant arm (eg a passenger seat) the plotted line on the weight x moment (IU) chart is a straight line. So, we can add any CONVENIENT weight to see where the straight line goes and we can observe where it crosses whatever CG position we might desire, typically the forward or aft CG envelope limit. It is a simple exercise, then, in the typical problem, to work out what extra weight we might need to add (or what weight we might need to remove if the original load put us outside the limit) to figure out the correct loading.
To minimise plotting and reading value errors, it is best to use large weight increments but, at the end of the day, it really doesn't matter what weight you use so long as you get a useful intersection which will allow you to read off the required information.
So, you pick a suitable weight increment. Whether that is 100 kg, or 50 kg, or 60,3749 kg (that would be a bit silly but you should get the idea) so long as it does the job, it will be fine. If you have done the job on a CG chart presented as weight x CG, it is better then to confirm the result by a quick longhand check calculation.