Thanks Ole, that was also my understanding, with the caveats you mentioned. Basically, a "match" at that level of granularity is not proof-positive--but a non-match at that granularity IS a strong refutation. Despite the fact that J-M241 diverged roughly 10,000 years ago, the number of people in this clade is relatively small compared to the total number of clades. On Wikitree, a number of other Schell/Shell males have taken Y-DNA tests. All of them (so far) are definitely NOT M241, so I can eliminate them as close relatives. By contrast, the one male I found who is also predicted to be M241 on 23&Me is indeed my relative, as supported by genealogy and sources. Thus, my conclusion is that autosomal-based haplogroup prediction is a great first step (and it costs nothing) because in research, the strongest result one can obtain is a refutation.