In the absence of other information, it's reasonable to assume linear scaling.
I have news for you - there is other information. This is a graph produced by a company comparing a product sold by that company with an undetermined alternative product. This other information tells you that it is in the interests of the producer of the graph to present their own product in the best light possible, linear scaling be damned.
Throughout this entire thread you have explicitly made a number of assumptions but these have been shown to be false, not least by your own experiments (as in the initial post). I do not understand why you persist in clinging to such false assumptions.
For my part I too have made one assumption: that your own intentions are honourable. I sincerely hope that this one is the exception which turns out to be true.