Mark, I was looking over your shoulder at the Symposium at one point, and saw a still shot from, I assume a test jump, where the MARD (not sure who's) released prematurely. It looked pretty nasty. Almost like the reserve bag got ejected out of the container at the same time as the reserve PC.
I wanted to produce a device that I had as much confidence in as the snap shackle that's the first link in the chain. When you think about it, our main deployments rely on the pilot chute being able to pull the pin to initiate the deployment, and then finish it- there's no need for second PC inside the container to take over the deployment once the PC has pulled the pin.
In testing, we found the same to be true with the Infinity MARD- if there's a horseshoe malfunction (for example) where the main is still in the bag, it doesn't create enough drag to open the riser covers, or separate the risers from the harness rings (even if the riser covers are open), let alone pull the reserve pin. So if the jumper cuts away the main, and pulls the reserve (standard EP's), the reserve bridle will disconnect from the Infinity MARD before the risers and RSL can lock the MARD to the bridle. So, from our testing, it appears as though the concept that the MARD needs to be able to "change it's mind" mid deployment isn't accurate.
Maybe there's some fringe entaglement/malfunction scenario where an aerodynamic MARD would be beneficial, but since you're starting with an entanglement already, nothing can really be predicted.