The reason to think in terms of ideological Turing test is that "opposite" is almost never correct. Almost nothing can be usefully simplified to a simple one-dimensional aspect where both ends are reasonable and common.

In the mulidimensional space of different personal influences (genetics, upbringing, current social environment, governmental and non-governmental support and constraint networks), there are likely multiple points of belief in the balance of choice vs non-choice. It's just not useful to characterize one cluster as "opposite" of the other.

Personally, I find the three-axis model fairly compelling - it's not that different political leanings come from different points on a dimension, it's that they are focusing on completely different dimensions . Progressives tend to think of oppressor/oppressed, Conservatives about Barbarism/Civilisation, and Libertarians about Coercion/Freedom.

This does get accepted (to some extent - it's still massively oversimple) by both liberal and conservative friends of mine, so passes at least one level of test.