I feel that canwesteng's point is the crux of this.
As a result of unintended joint stiffness and joint prestress due to axial load, all base plates are pretty much fixed. Right up until they're not that is. There are two basic paths:
1) The base connection attracts moment until it fails in a ductile manner (rare). The moment in the connection then remains relatively steady at the yield + over-strength value.
2) The base connection attracts moment until it fails in a brittle manner (common). The moment in the connection then drops to zero-ish.
So the goal of design is not to create anything approximating a pin at the column bases. That's improbable given typical detailing practices. Rather, the goal is to ensure that the diaphragm providing lateral support to the tops of the columns doesn't shift very much. As long as diaphragm movement is kept small, the tops of the columns won't move very far, and the column base connections will never attract enough moment to initiate a brittle failure. Consequently, the parameter of interest is the ratio of gravity column lateral stiffness to story bracing lateral stiffness.
Consider two extreme one story building cases where the gravity columns are just a bunch of HSS6x6 pinned to the supported roof beams and conventionally "pin" fastened to the foundations:
1) The lateral system consists of stiff cross bracing at regular intervals on all four walls. In this case, as canwesteng intimated, the lateral stiffness of the story bracing will dwarf that of the gravity columns. The gravity columns will be shielded from attracting moments that would cause brittle base connection failure.
2) The lateral system consists of a handful of the HSS6x6 columns designated as flagpole cantilever moment frames and provided with kick-ass moment resisting base plate connections. In this case, the designated lateral resisting system will not shield the gravity columns from attracting moments and brittle base connection failures may well occur in the gravity columns.
The Take Away
It is counterproductive to use a tight, centrally located anchor bolt pattern in an attempt to approximate a pin connection. Baring an extreme anchor bolt / base plate arrangement designed with the intent of creating flexibility, you're going to wind up about the same amount of base connection rotational stiffness no matter what you do. By using tight, central anchor bolt patter, you just exacerbate the prying/tension loads on your anchor bolts when the inevitable base moment shows up. The goal of design for "pinned" base connections is this: moment capacity < moment demand. Reducing moment demand by stiffening the story bracing helps; reducing the moment capacity of the base connection...hurts.
It's worth noting that effective column buckling length goes another way. It takes very little flexibility in the base connection to turn a fixed based column into a column that is effectively pinned.
The greatest trick that bond stress ever pulled was convincing the world it didn't exist.
