You are quite right.
This alternative approach fascinated me for years too, and eventually I just had to try it, and it worked better than I ever imagined it would.
I did a lot of back to back testing with two identical sets of panels (my main system split in half) and a Make Sky Blue MPPT controller connected to a pair of Turnigy power meters, both charging the same battery.
Swapping around panels and power meters in various combinations, the results were always within a very few percent either way.
Neither system was significantly superior to the other.
In the end, it came down to the power components, the mosfets and series inductor in the buck converter.
The improvement in power conversion efficiency with larger power components rapidly gets into diminishing returns.
There is no great difference between the two very different algorithms, and exactly as you say, a microcontroller costs no more these days than a direct simple PWM control chip.
I think it comes down to the microcontroller potentially being able to have more features, versus the more basic approach being a much simpler to build and repair controller, especially for those that are software challenged. The commercial units are difficult if not impossible to repair, as no circuits are available.
A further advantage of building your own controller would be a controller for higher voltages.
Higher voltage commercial mppt controllers seem to be pretty thin on the ground.
There is a thread here somewhere about all this, including schematics and some pictures of my controller, and the testing results.
I could not find it earlier.
Ah just found it !
https://diysolarforum.com/threads/simple-home-made-analog-mppt-contoller.47057/