[Spontaneous recovery can take place at any age, *though the longer a person has stuttered the less likely will the problem disappear of its own accord*.](https://www.mnsu.edu/comdis/kuster/jdwilliamslibrary/williamscourse.html) Emphasis mine. [76% of subjects in a study cited by Ehud Yairi recovered spontaneously within four years.](https://www.stutteringhelp.org/research-early-childhood-stuttering-0) To say that there is no correlation between time-since-onset and spontaneous recovery is dubious according to many if not all sources I've seen/read. I'm 95% sure that my own fluency professor (who I won't name as to lessen the odds of doxxing myself, but you'd likely recognize the name if you're into fluency research at all) taught us that if you're going to spontaneously recover it'll likely be within 15 months of onset, with your odds dropping after that. I *could* be wrong about that figure, though, as I can't seem to find that exact power point.