Firstly I think it’s really helpful to have a basic grasp of vector based collision detection, at least to the point that you know how it can make things easier and also what it is able to achieve. Secondly I think it’s helpful to be aware of quantisation when checking for floating point equality.
The diagram below shows a moving red box being checked for collision with a green box. On the left you can see the full extent of the relative velocity vector - and FYI if you wanted to take it further you could have a rotating square, the maths just “works”. On the right is what vector based collision detection can achieve, the point in time of impact.

Now the question you might want to ask now is, “how do you find the point of collision”? I’m on my way out now, but when I get back I will find a link to one of my previous posts on this forum or another about using vector maths to solve collisions for geometric points, and I’ll upload some of my old collision examples to github 
Now you can see what vector collision detection looks like (at least what it takes in and what it puts out) you might notice something specific about the image on the left that might not have stuck out before. Suppose the boxes passed through each other, if you only checked for collision on specific time slices you are going to miss many collisions when velocities exceed the span of the boxes.
Edit: maybe tomorrow actually, so far I’ve collected links of previous topics, but it’s quite a mouthful right now …