The fact that they "controlled" for genetics using polygenic scores already is a strong sign of low quality research. Polygenic scores are powerful, but they contain very large amounts of noise compared to the true genetic effects. Controlling for genetics using them is like controlling for income by asking whether the respondents own a Porsche. Also, be aware that Scientific Reports is, if not quite a predatory jour…
Video games, like all things, should not all be treated equal. I could certainly see problem solving skills developing from world building or highly complex games (Civ, PoE, etc.). In fact, most (but not all) highly successful games have depth, which requires time investment and problem solving. The difference in games can be as varied as comparing a marketing pamphlet to Asimov's novels.
I don't dispute your take on the quality of the research though. I would even go further and speculate it would be really really hard to come up with meaningful tests due to game variance. So most anything on the subject is likely fluff.