Because it's a proportional thing, SQ - I believe; it's ages since I read it. It's not about having a fixed amount of one quality or another. As you say, 'high' on both would put the result in the middle - ie, they wouldn't actually be high in the given context.
I was responding to the very common complaint about any such system, where people claim they have high levels of both qualities so the system must be wrong.
You're right, it does use 5 different characteristics so it's a sliding scale which allows for a reasonable stab at the the proportional approach. The tests aim to measure a respondent's strength of interest in (drive towards) rule-based structures and their drive towards understanding & interacting with people. As it's all about proportions, the results are given as a quotient.
From the Wikipedia article:
E or S. E-S profiles show reliable sex differences in the general population (more females showing the profile E>S and more males showing the profile S>E).[1]
The E-S theory is a better predictor of who goes into STEM (Science, Technology, Engineering and Mathematics) subjects than is gender.[2]
Neither of those statements claim to say whether the difference is innate or learned.
The first one says more men get a high SQ score, while more women get high EQ.
The second says people with high SQ are more likely to go into science than other people, irrespective of gender.
It would have been better if Baron-Cohen hadn't called it "Male Brain" but his data (and subsequent studies) do show that more men than women have a proportionally high SQ with lower EQ.