One quick helpful point might be if you are using WL/2 to design the connection, as the structural engineer what the actual reaction is.  This may really reduce the connection needed, and two bolts might work.

This is just the sort of problem which convinces me that structural engineers should design their own connection details, and take responsibility for them.

If I understand the situation, the design moment on the 2-bolt group to the W8*18 may not only depend on the load on that beam, but also the load on the opposite W8*13, since there is a continuous beam effect that could come into play.

Thus, for example, whether one or both W8 beams are loaded, there will be a tendency to negative (top tension) bending at both connections.

If, however, the design engineer has made the common assumption that these beams are simply supported at the centreline of the W12, then the design moment at the connection would be positive (compression at top).

(Don't ask me how on earth the actual structure is meant to read the designer's mind as to assumed method of action - that is much too deep).

If you are supposed to be treating the W8 as simply supported at their ends, the the answer to "can I reduce the eccentricity because the shear tab is welded to the top flange ?" is an unequivocal NO you can't.  The shear force in the W8 has to be transferred from the bolt locations to the W12 web centreline, and the welds change nothing except conditions in the shear tab itself.

Dougantholz's suggestion that you talk to the structural engineer is a good one, particularly if it results in the structural engineer actually committing to responsibility for what was shown on the design drawings.
  

I agree that you should talk to the engineer, and even offer a different detail that is more conventional. Coping the top flange of the W8's and bringing the bolts closer to the W12 appears to be a better detail.  Sometimes the steel detailers have better ideas about connections, and the costs associated with different detailing methods.  Anyways, the engineer was just probably "trying" to show the most economical detail, and thought that was it.  If you can come up with a detail that doesn't increase fabrication costs, then I'm sure he'd have no problem with it.

ps - It doesn't sound like the eccentricity affects the bolts so much, but rather the gusset pl. (seems a little thin at 1/4")

austim,

I agree with your point about how is the structure supposed to know what the designer is thinking.  This looks like one of those cases where (and I'm just guessing) someone said let's assume the beam (in this case the W8's) are simply supported and thought they had a conservative design. But never really thought about the joint itself. In this case it seems to be some sort of hybrid between a fixed and simple support (I know, I know most joints are really never truly fixed or truly pinned but you know what I mean). I guess my point is I'm okay with assuming a worst case end condition to design a member but I'm not okay with sticking with that assumption when looking at the joint itself. Anyway, so much for philosphy.

Specifically on this joint, if I had to just blow and go I would assume the full eccentricity of 6 3/4" for sizing the bolts.  In other words I would ignore the help you get from welding the plate to the flange of the W12.

Please bear in mind I'm just talking about the bolts.  I think the plate itself should be looked at more closely.  Especially in regards to what austim said about the W8's acting like a continuos beam. This would put compression on the lower edge of the tabs possibly causing them to buckle. Hey, look on the bright side now you would have a simple support. I'm not even going to get into the possible lack of torsional restraint at the ends of the W8's. (Mainly because I wouldn't know what I'm talking about.)