Is it lopsided in the number of connections reported to the 443 VSs or lopsided in the answers the GTM gives out for that wideip? Those are very different issues.
There may very well be a corresponding difference in the connection statistics going to these 443 VSs versus the port 80 VSs.
Perhaps log some connection info on the sets of VSs. You might see a few long duration connections on the 443 that happen to get assigned to one VS might cause a lot more short duration connections to be assigned to the other VS. This could affect the pattern of IPs being returned by the GTM for your wideip as well as the number of connections being reported by the LTM for the VS.