I've looked at the specs for the two packages, and if I remember correctly, I think that the Ricardo software did a better job representing the skirt-liner interface than the MDI software did, while the MDI package allows more degrees of freedom in the model. I don't recall reading anything that would lead me to believe that the MDI software can predict wear regions on the skirt, while the Ricardo software definitely can. I look forward to hearing your comments in 2005. Adams engine, from what I've heard, has some pretty sweet features for valvetrain/geartrain dynamics.
How does ADAMS model collisions? I have only limited experience (about 100 to 200 hours) with that software, and I'm not clear on all the details. It's been quite a long time since I used it, too.
When you say that you black-box the pistons, you mean that you give some specs to Mahle or Fed. Mogul (or other) and get a complete piston design back, right? I know (based on some previous work with an engine OEM) that some of those guys use Ricardo's PISDYN in-house, although from what I recall, the features of the software are not fully utilized by them.