Hi, Collin. Two things at play, I think, both here and with your recent question about Richard III. First is to note that WikiTree does no DNA matching or comparisons. It can't because we never upload DNA data to WikiTree and it has no access to our our actual test results at AncestryDNA, 23andMe, Family Tree DNA, etc. So when WikiTree shows us people who might be in our genetic tree--whether in the "DNA Connections" panel on a profile or via the Relationship Finder--what it goes by is the information entered here into the genealogical family tree. A mistake in the tree will result in an incorrect proposed DNA connection.
It looks like the profile for Charles Smith (Smith-214542) was edited within the past 24 hours, though all that show on the change log are answers to this G2G question. Nevertheless, for now Charles Smith is showing no genetic connections in the tree; that should change after nightly processing.
Second, your yDNA test with FTDNA is a 37-marker STR panel. This is now considered an entry-level resolution (when I first tested it was 12 markers, and then an upgrade to 25 before the 37-marker test appeared). We've learned, especially with the advent of the Big Y sequencing tests, that a low level of STR testing really can't be an accurate predictor of genealogical relationships. It can be quite useful as negating evidence: if two men have very different haplotypes, they will not be related in the genealogical timeframe. But it's of less use as positive evidence in linking men in a generational fashion.
At 37 markers FTDNA will include anyone as a "match" at a genetic distance of 4 or less. Mind you, that term "genetic distance" has nothing to do with the number of generations; it's simply an indication, using what's called the infinite allele model, of how many mutations are believed to separate the two different sets of STR results. I personally believe the level of GD4 is too lenient, and the infinite allele model itself does not take into account the extremely variable mutation rates exhibited by different STR markers. Some markers mutate at rates that are several factors of magnitude slower or faster than others.
As a result, at 37 markers you could be a GD1 with both your father as well as a patrilineal ancestor with whom you don't connect until the mid-1700s. In my case, we have a fairly mature subproject with almost 40 members, over 60% of whom have taken the Big Y test. From that I've learned I have two GD4 matches at 37 markers with whom I cannot share an ancestor until circa 900-1000 AD, clearly before the genealogical timeframe.
Even at 111 STR markers, where FTDNA allows up to a GD of 10, there can still be relationships beyond the genealogical timeframe. At 111 markers I have 12 reported matches at GD5 and greater, up to GD10, that align not with my immediate family, but with that of the patrilineal branch that split in that period from around 900-1000 AD. We most definitely share a male ancestor; it's just that the full-sequencing data shows us the branches split in the British Isles prior to the common use of consistent surnames.
Unfortunately, a "match" is almost never as straightforward as being a "match." The lower the testing resolution, the more inference plays into it and the more investigation is needed to help determine the true relationship.