I have two problems with close range weapons:
1) Units get spotted from a too long distance in dense terrain, however this cannot be changed in editors and there are reasons for keeping a free visibility zone around each unit.
2) I think weapons are fired only when the center of an enemy unit is inside the (at least) max range coloured circles, which is an important fact for RPGs and SMGs; given their short range often the forward part of a unit would have those weapons in range with an enemy (or a portion of an enemy unit), however my unit wouldn't fire them because the center of the unit is too distant from that enemy.
I have increased slightly the range and accuracy for these weapons editing the Estab, and it increases a lot the fun, in particular RPGs which get a chance to fire at least (always talking about daytime). During night they are just fine.
1) There is a patch in the works to address the spotting issue you mention by taking into account a combination of both atmospheric conditions and time of day in determining effective spotting ranges. That said, there are many instances where a threat can be seen yet not attacked with a viable weapon in war at that time.
As far as the issue with centers of units, keep in mind that the combat modeling and casualty calculations in an operational scale game are performed on a unit by unit basis rather than an individual or single combat vehicle basis -- you can't order a sniper to take out a key commander for example.
When Command Ops 1 was going through its development, there was significant discussion of how direct fire casualties are dispersed into units based on their formation type, fields of fire and ranges. They basically took into account the expanse of the formation's footprint, its disposition and facing, and its cover status to assign casualties for any encounter based on both the ability to "see" enough of a unit to order a firing unit to attack, and the capability of weapons assigned to the firing unit to attack the bulk of what is "seen."
If I understood the discussion correctly, the assumption in performing those calculations is that the ultimate fire impact revolves around a center of gravity for the firing unit formation type / facing, and the effects are calculated against a center of gravity for the target formation type / facing to address both the ability to spot the bulk of a target unit and put fire on that bulk.
It amounts to a compromise for putting a simulation such as CO2 onto a PC platform and obtaining reasonably accurate results in the combat modeling without slowing down the performance to an unbearable crawl.
I worked in support of developing military training simulations until 2009, and the calculating power and electricity to bring a combat simulation just to the vehicle on vehicle level for a battalion-sized training simulation that accounted for individual vehicle casualties during combat required the equivalent of 10 home PC's per vehicle and enough power to run a small city -- a little bit above the minimum requirements to purchase Command Ops 2.