The requests do not resemble folders just words with hyphens
asking for www.example.com/folder1/ before asking for www.example.com/folder1/folder2/folder3/ is a good idea.
Your example with slashes won't work. You would block a plain images folder default access:
www.example.com/images/
but you could very well have a friendly link that also includes the images word like:
www.example.com/images/blue-widgets.html
which returns a valid page
These requests are also accessed at other times individually not in particular order nor with the same rate. I can't remember if at some point I gave to google a sitemap with the wrong links ages ago but for a couple of incorrect words I had once, google has a very long history even if the misspelled links were short lived.
I also see although were crawled by words length they were not crawled alphabetically when the words had the same length.
I've seen this behavior for years as well.
Which behavior have you seen exactly? So you have these links:
www.example.com/red-widgets.html
www.example.com/blue-accessories.html
www.example.com/free-stuff.html
And the googlebot accesses
www.example.com/red
www.example.com/free
www.example.com/widgets
www.example.com/red-widgets
Have you seen that?
