Man; I really am no good at all the cousinship stuff. It's like doing math but without equations.
From a glance, I think the MRCA (couple or not...but, yeah, if a singleton small segment I personally don't think it can be assumed that both in the couple donated DNA, but those are the WT policy guidelines) is the 3g-grandparent of Missy's mom, so 4g-grandparent to Missy. The Ancestry match is the 3g-grandchild of the MRCA, so her mother is a 2g-grandchild. That would make Missy's mom the 2C2R of the match's mom, and make Missy and the match 3C2R.
I think. My head hurts.
I scrunched the numbers from last year's Shared cM Project update using the standard deviation figures that Blaine provided us for the first time. That scrunching is in a table here. By the CoR, the theoretical sharing between 2C2Rs would be 0.7813%, or about 53.13cM. By the Shared cM Project's scrunched numbers, 95% of the time the value will be 0 to 158cM; 68% of the time, between 29cM and 115cM.
Edited: Missy and Barry got that by email, but I'm deleting the rest of the response. Because I was wrong. Sheesh. Not my day!
I'd tried to walk back the generational dates and thought I'd discovered a discrepancy on the match's side. Never mind. But a singleton 15cM segment is pretty low--if any match is detected at all--for 2C2Rs.
I'm walking away from the keyboard now...