Correct use of "x times lower"

B is half the length of A

B is one-quarter the length of A

When in doubt, rewrite the sentence for clarity. It's always a little tricky to express math in words, that's why it has its own special formats.

If it must be done, remember this rule of thumb: digits tend to emphasize numbers while words tend to diminish their impact.

One could say:

B is one seventeenth to one thirteenth the size of A.

But in modern scientific literature I think it more likely be expressed as something like:

B is 5.9-7.7% the size of A.

Dude it's just one-seventeenth to one-twentieth the size of X.

{Incidentally: you are totally wrong to assert they have to be "in order"; whether for a multiple or a fraction. You can certainly say: "P is 10 to 5 times bigger than Q." No problem. "10 to 5" simply means "the numbers from 10 to 5".}

BTW another common ons is: A is smaller than B by a factor of 13 to 17.

that may be what you have in mind?

B is 13-17 times lower/smaller than A

The problem is then, how would one best rewrite this since "x times" will, at least mathematically when x > 1, always be larger and not smaller?

I disagree with the premise of the question. The example is OK, although a little awkward. In "x times smaller," the word "smaller" inverts the ratio. There are of course other ways of expressing this, but that's a matter of style. Using "by a factor of" doesn't necessarily fix things:

(1) A millimeter is ten times smaller than a centimeter.

(2) A millimeter is smaller than a centimeter by a factor of 10.

(3) A millimeter is smaller than a centimeter by a factor of 0.1.

(4) My car is lighter than your car by a factor of x.

Of 1-3, I think 1 is the best style and is fine mathematically. 2 and 3 show that there's a potential ambiguity. In 4, the ambiguity is a real problem, because we can't tell if x is being defined as a number that's less than 1, or greater than 1.

Edwin Ashworth wrote:

[...] linear scale factors are used to avoid [...] the confusion where in everyday language 'ten times bigger' is used to mean 'x10' whereas 'one time(/s) bigger' (paraphrasing 100% bigger) means 'x2'

I think this is an innumeracy issue, not a language issue. Some people just don't understand how to convert back and forth between ratios and fractional changes, or don't realize that they're different things. For example, if they're told that B is 7% greater than A, and are then asked to find the ratio B/A, they may say 0.07. We're talking about scientific writing, where this kind of innumeracy isn't an issue.

This is one of my pet peeves. In one of the responses above, the writer provides the example "(1) A millimeter is ten times smaller than a centimeter." The writer then goes on to state that this example is "fine mathematically" when in fact, it is incorrect. When you see "is" in a sentence, you replace it with an "=" sign. Thus, the example can be written as 1mm=(10)x(1cm) or 1mm=10cm. This is clearly incorrect. If the mathematical relationship between a millimeter and a centimeter is written this way: 1mm=(0.1)x(1cm) or 1mm=(1-0.9)x(1cm), then the acceptable English language translations are "A millimeter is one-tenth as long as a centimeter" or "A millimeter is 90 percent shorter than a centimeter," respectively.

"A is 13 to 17 times the size of B

this can be written as

A is 13–17 times higher/larger than B"

This is incorrect. Let's do an example:

2 is 2 times the size of 1. but... 2 is 1 times larger than 1. Because we can think of "1 times" to mean 100%: 1 is one times the size of 1. But 2 is one times larger (100% larger) than 1.

SO: "A is 13-17 times the size of B" can be rewritten "A is 12-16 times larger than B"