Grasping the concept of a three-value study wasn't easy or straightforward for me. At first, I found it difficult to look at a reference and decide how to reduce it to just a few values.
What I did find helpful was drawing out a few thumbnail sketches and deciding what I actually wanted to include in the painting. Once I started doing that, figuring out the lights and darks became much easier.
I began to realize that the thumbnail wasn't just a small version of the painting. It was a way of making decisions before I started painting.
I actually found that the three-value study works better for me with landscapes than with urban sketching. With a landscape, I can see how grouping the larger areas into lights, midtones and darks can help simplify the scene. With an urban sketch, I find there are so many different shapes and details that it doesn't seem to work quite as naturally for me.
The end result was actually pretty good, even though I didn't quite achieve the three values I was aiming for. One thing I discovered while working from a photo was that I could bring it into Affinity Photo and take a look at the three values there. I know, cheating, but what the heck! It helped me understand what I was looking at and gave me another way of seeing the lights and darks.
I still use thumbnails occasionally, but I've found that I prefer working from the photo in hand. That way, if I decide something needs to be darker, I can make that decision as I paint rather than being locked into what I worked out in the thumbnail.
I wouldn't say I've mastered three-value studies, but I understand them a lot better than I did when I started. Sometimes an exercise doesn't work exactly the way you expect, but that doesn't mean it wasn't useful. In this case, I think I learned more from trying it and figuring out what worked for me than I would have from simply getting it right.