Currently the docs for `difflib.SequenceMatcher.quick_ratio()` just says 'Return an upper bound on ratio() relatively quickly', which doesn't give much of an idea about how that upper bound is calculated. `real_quick_ratio` has similarly brief documentation.

I'll raise a PR shortly to add a more verbose description to each of these ratios, so that it is clear when each should be used.

My current suggestions would be:

quick_ratio
    Return an upper bound on ratio() relatively quickly. This is the highest possible ratio() given these letters, regardless of their order.

real_quick_ratio
    Return an upper bound on ratio() very quickly. This is the highest possible ratio() given the lengths of a and b, regardless of their letters. i.e. 2*(min(len(a), len(b))/(len(a) + len(b))

Thanks for the effort, but I'm rejecting this.  The language deliberately defines nothing about how these are calculated.  It defines how `.ratio()` is computed, but that's all.  An implementation is free to do whatever it likes for the "quick" versions, provided only they return upper bounds on `.ratio()`.  Indeed, it's perfectly fine if an implementation merely returns 1.0 for both, regardless of the arguments.

If an implementation is cleverer than that, great, that's fine too - but it would be actively counterproductive to constrain them to be no _more_ clever than the current implementations.