I have done some dyno R&D on CFI units. The biggest problem I saw wasn't at idle or WOT - it was when the throttle plates were at NEARLY WOT. At idle the injectors will make a pretty conical spray pattern more or less uniformly distributed across the throttle plate. GOOD.
At WOT the velocity of the inrushing air pulls the conical pattern of the spray discharge down into something more closely associated to the fuel coming out of a carb booster - less conical and less evenly distributed over the cross-sectional area of the throttle bores. Not great, but still GOOD.
BUTT... at throttle blade angles around 75 or 90 degrees the air velocity is large enough to distort the spray discharge from its ideal conical shape. And because the fuel is entering from what is essentially a point source, the injector pintle, and this point source is located directly above the throttle plate, that's when BAD things start to happen. The fuel impinges on the steeply angled throttle plates and "slides" off to one side of the throttle bore, greatly stratifying the fuel mixture and playing havoc with the fuel distribution downstream. I have seen this condition in V8s that exhibited as much as a five air/fuel ratio spread between cylinders. Imagine an engine with 11:1 A/F ratio in one cylinder and 16:1 A/F in an adjacent cylinder!
If I were trying what you are doing I would try to bias this rich / lean condition so that the orientation of the throttle plates serves to offset the latent fuel biasing encountered with siamesed intake runners.
