One would also expect this effect to be a function of distance, not of how prettily lined-up the planets are or the shapes they make when viewed from above.

My claim isn't that it's unlikely that any celestial body can affect earthquake likelihood. It's that it's unlikely that earthquake likelihood is affected by pretty geometry of planet arrangements that do not include our own.

For example, on the "About" page of their website, they show this image:

https://ssgeos.org/images/planets/1960/1960-05-21.jpg

They go on to make the claim that earthquakes were more likely on Earth, because if you draw a line from Venus to Neptune, and another line from Mars to Uranus, those lines cross at right angles.