In my reckoning, you have on normal events 6 possible results equally distributed. So the chances to get a specific item is 1/6.
Now it would be interesting, how plarium handles the situation.
If they actually stil calculate only 6 possible results being available, your logic is sound as they change the equal distribution of the results to a weighted distribution which leads to your number of 1/3 and 5 times 2/15. This would fit with doubling the chances. (Unfortunately, I do not have the impression that the charms are that effective)
If on the other hand Plarium is adding another possible result while keeping an even distribution, we now have 7 possible results with 2 of them being the results enhanced by the charm. So we get 2/7 and 5 times 1/7 chance to get a specific type. Even so, Plarium is telling the truth that the charm provides a x2 chance...they are just omitting the addition of *in comparison to other results*,
Your logic would actually make the charms more effective. Do we trust Plarium to give players more control over RNG than needed?
From my admittedly basic understanding in programming, I can imagine that the solution with 7 possible results is much easier to implement. Something like
If *charm type* used add *charm.type* to Array.possible.result
Please do not judge upon the syntax of the line above, it should just give the idea of the simplicity of programming how I would imagine it.
If using your logic, you must implent different percentages for the results to happen and need a table to assign the new percentages for the possible results because you are not using an equal distribution anymore.
Perhaps I am wrong but I would suspect Plarium to take the easy route and just add to the possible results and not actually shifting the weight.