My understanding as I have learned it, painstakingly I might add on the torque model is this - the map side basically controls the timing at it's core at idle. There is an offset in the map side. The offset is directly defined by how much you changed the VE model and MAF curve (if you lowered the idle area of the VE table in the idle regions say 30% then that's how much you need to increase the map model and then add in the MAF increase for the same corresponding idle cells in those corresponding regions of the torque model). The map side will define the "base torque defined as DELIVERED TORQUE for idle". It then uses that and matches it with the airmass side to get the TIMING. The TORQUE itself is defined by the airmass side. Confusing right since it seems like I just said the opposite??? Oh wait it gets better. Throttle control "usually" is defined by the map side, but it will use the airmass side to help define how much it needs to close. You have to walk a fine line not to close the throttle too much and get the right torque while still getting the timing you want. Curvatures affect things as much as anything else. You can put a curvature into the airmass side, but the map side has to stay flat, so to get everything "right" you have to use the curvatures to balance both out to one another. Also keep in mind even though IT WILL change with torque model changes, fueling needs to be nearly 100% dialed in before making many changes to the torque models.
The jist of all that? You can't just lower the model to get what you're after like you did. This is also why I've really been pushing for a good torque math formula. Braxton almost has one though
Basically you'll need to think about what a cam does and how it affects power for the given loads and then try to change the model as best you can for that on the "maf side - and think about that (maf side)" then work the map side back close then balance them using the airmass side and map alike.
.