A new paragraph moves rankings when it genuinely improves the page, better coverage, a closer fit to what the query wants, or information the competing results lack, and it does nothing when the paragraph is filler that adds no detectable value. The refresh effect tracks the value added, not the act of editing. This is the pattern observed again and again: the edit itself is not the lever, the improvement is.
When the addition closes a real gap, the mechanism is straightforward. The page now answers a question it did not answer before, or covers an angle the searcher needed, or brings a detail the other ranking pages skip. The search engine re-evaluates the page on its next crawl, finds it serves the query better than it used to, and the page can rise. The movement is a response to the page becoming a better answer, which happens to coincide with the edit but is caused by the substance, not the timestamp.
When nothing moves, it is because nothing of value was added. A paragraph that restates what the page already said, pads the word count, or sprinkles in keywords without new information gives the search engine no reason to re-rate the page. The crawl may register that the page changed, but a change with no improvement in helpfulness leaves the ranking where it was. This is why “publish fresh content on a schedule to refresh rankings” disappoints so often: it treats editing as a mechanical trigger when only real improvement triggers anything.
So before you add a paragraph expecting a lift, ask what it genuinely improves. If it covers something the page missed, sharpens the answer, or brings detail the rest of the results lack, it has a real chance to move the page. If you are adding words to look active, expect nothing, and spend the effort on a paragraph that actually makes the page a better answer instead.