Are quads needed for a low poly videogame model?

I real always that tris are evil. I have never understood the why, but during my exercises I am always mainteining all in quads in any case.
Now, I am creating this sort of vintage back bag to be suitable for a videogame. This is only an exercise and there are no concrete videogames on the horizon, but I want to understand the “theory”.
So: in the image you can see that there are many “useless loops”, that maybe are only useful for a very little part of the mesh but to maintein the quads-only dynamic I add an entire loop. This, causes shading uglyness and same if I subdivide the surface, that makes the bad shading also more strong.
I could solve it by adding an omogeneous amount of loops even where they aren’t needed, or, using tris.
So, in a videogame asset, what is the concrete advantage to have quads-only topology? I can understand that it is useful for rendering in Blender because quads second the subdivision surface modifier well, while tris may be unpredictable. But what about a model that is meant to stay low-poly?


Are quads needed for a low poly videogame model?

Short answer: No
If your object is not going to have any transformation(bending or something similar), you can get along with triangles without problems.
And you probably won’t use Subdiv in the game, so no problem there also.