Because the point of an attitudinal survey is to gather data about your attitude. "No opinion" directly undermines this goal, and makes the survey pointless. It also provides an easy out for respondents who don't want to put in the cognitive workload to form and express an opinion, but still want to appear "cooperative" by completing the survey -- sort of a manifestation of the Abilene paradox (https://en.wikipedia.org/wiki/Abilene_paradox).
And that's actually the biggest flaw of the Advocates of Small Government's quiz. Not only does it provide a "no opinion" option ("Maybe"), but then it appears to interpret it as a centrist attitude anyway. So, a totally neutral response -- all maybes -- puts the respondent dead center. Or, in the "Libertarian" worldview, half-way to dirty statist. I mean, "Libertarian" evangelism is the whole point of that quiz, so it not surprising that a totally neutral response would be interpreted as proto-statist, it's just bullsh*t is all.
As to "loaded questions," FloorNet went though and answered all the Political Compass questions four times, each time answering exclusively and consistently the same for each question:
All "Strongly Disagree:" https://www.politicalcompass.org/analysis2?ec=0.0&soc=-4.36
All "Disagree:" https://www.politicalcompass.org/analysis2?ec=-0.25&soc=-2.41
All "Agree:" https://www.politicalcompass.org/analysis2?ec=0.38&soc=2.41
All "Strongly Agree:" https://www.politicalcompass.org/analysis2?ec=0.0&soc=4.36
It concludes that the questions on the Political Compass more frequently represent an authoritarian and ever-so-slight right viewpoint. By contrast, if there were an equal number of questions worded so as to represent each quadrant, blanket responses as above should put one dead center in all four instances. Ergo, FloorNet concludes that the specific wording utilized by the Political Compass will cause Left Libertarians and Right Authoritarians to appear more extreme than they actually are.