No worries! It's easy to see "designing better" and jump to assuming that also could mean "designing perfectly". I have been fighting a losing battle against this misperception for years now.
A few points.
1) All squads have an element of cyclical balance, i.e. the "paper rock scissors" effect. Therefore:
1a) A squad's competitive "score" is not a singular value. If there are N viably competitive squads, then every squad (viable or not) has N different win ratios. A squad's win ratio vs itself is always 0.5, and a squad's win ratio vs any other squad is dependent on their relative power levels and their particular matchup strengths or weaknesses.
1b) If all of a squad's win ratios are equal or less than another squad's win ratios in every matchup, then it is mathematically dominated, and is not competitively viable. I.e. if Squad B is universally better (or at least as good as) Squad A in every single matchup, then Squad A is inferior in the absolute sense.
1c) In the idealized system where all players have perfect information, squads that are mathematically dominated have zero appearance rate competitively. In reality these squads still show up occasionally, since players do not act on perfect information. More on this below.
1d) Squads that are not mathematically dominated are considered to be competitively viable. The steady-state solution for the appearance rate of all squads is the solution of N equations and N unknowns such that the expected win rate of any given squad is 0.5. As long as there is a paper rock scissors effect, the meta will counter itself so no squad has an advantage.
1e) For at least the first few major tournaments after a discovery of a new squad (either by new release or delayed discovery like Parattanni), the system has not reached steady state solution, and some squads will have an expected win rate >0.5. X-wing is generally in this state, hence the competitive focus on "countering the meta game". If the game were to go a full year without releases, then it would certainly reach the equilibrium described in 1d.
1f) Local metas may differ due to a variety of reasons, notable player skill (a squad's win ratios can be a function of player familiarity/skill and therefore by extension geography in the broader sense), or simply preferences with particular squads (not modeled by the idealized system).
Conclusion of 1: you can have some squads that are generally not as good as others, but still see viable competitive play, just in lesser numbers than other squads. Cyclical balance is therefore an extremely powerful design tool for a designer, as there is a built-in margin of error when designing a ship to the game's power curve. The game designer's goal is generally to maximize viable squad diversity, which is indicative of both healthy game design and a fun play experience. Also note that a squad can be mathematically dominated, so its overall expected win rate in a given meta must therefore be <0.5, but if the design tolerances as tight enough, you may be able to play it without necessarily getting completely blown out of the water. In a system with poor balance, "Squad X" (say all X-wings) may have a win rate of 0.1, whereas in a better balanced system it might be 0.3 - 0.4, where "Squad X" is mathematically dominated in both cases.
2) Complexity creep is a design problem in any game system, grows exponentially with time, and inevitably pushes balanced design space into a corner. The designer's goal is to both identify all viable permutations of new design elements, and then quantify them, so a prediction can be made as to its impact on the overall meta-game (point 1d). Neither of these are trivial, but the more effective the design team is at both of these, the longer the game will remain in a healthy state. At risk of going slightly off-topic: as the industry continues to mature and barriers to entry rise, additional investments in both expertise and money (the former being directly related to the latter) will be made for AAA games specifically for better game balance. Identifying permutations is a very difficult problem, but one that I expect automation / AI to play a bigger role in the future, for any companies willing to invest the up front expertise and capital. FFG does not strike me as currently being as such a company, and the industry as a whole may take another 10-20 years to really get to this point.