Interesting that the misfire is at decel as the injectors are shut off then. Can you post freeze frame data, that might be helpful. Also maybe a cranking/running vacuum waveform. When I get a difficult one I start with a freeze frame and try and duplicate it exactly. Also if you have access to a tech2 you can look at a real time misfire graph and see exactly when it's misfiring. I'm thinking you have a problem like the inline 2.8 and 3.5s that run well but set misfire codes due to partially sealing valves. This can be found using your vacuum transducer. I have an example in the waveform database under a 2006 hummer 3.5. Also think what the computer uses to count misfires, crankshaft acceleration and it may see what you can't feel. I've seen a number of the in line's run fine with valve problems and I've used the graphic to find it.

Michael Byer
Owner/Technician
Mike Byer Auto & Truck Repair, Inc
Asheville, North Carolina, USA