Fortunately mine survived its maiden voyage with no lawn darting.
I've been pondering the problem though, and a few observations:
1) The issue seems to be limited to the Big Daddy, and to a lesser extent the Leviathan. However, the Estes 3" is hardly the only sloped shoulder nose cone. If nothing else, every other PSII kit has sloped shoulders. I compared it to my Ventris transition, and that has an even larger sloped shoulder (relative to overall shoulder length), yet deployment seems very reliable (knock on wood). The MDRM has a bigger diameter, but the sloped area is shorter/smaller relative to overall size, and it seems to deploy reliably (knock on more wood).
2) The pressurized volume of the Big Daddy is large for an LPR bird, but it's not that big. In fact, once you get the nose cone fully inserted, there's only a few inches of length between the back of the shoulder and the front CR. I'd imagine there are several 24mm Estes rockets with the same or larger volumes where we don't hear about failure to separate unusually often (the Mean Machine is one extreme example).
3) One thing that is unique is the "stubby" aspect ratio - the pressurized volume is very short compared to its diameter.
4) I also noticed that, due to the length of the shoulder and the width of the tube, once you get past the non-sloped section it's fairly easy to tilt the nose cone back and forth, potentially binding it on the airframe. In the smaller diameter of the Ventris, you can tilt it a bit but it doesn't bind as hard.
So with all that, my theory is as follows:
Ejection is not an instantaneous event. Basically, the gas produced by the burning BP pressurizes the airframe until the force on the base of the nose cone overcomes the friction (and any aero forces, plus the cone's own inertia) holding the cone onto the airframe. At that point the cone starts to move, accelerating until the gas vents out and/or the cone is free. In a perfect world, the cone pops off with enough momentum that it pulls the laundry out behind it.
On the Big Daddy, the short airframe volume and long, sloped shoulder work against you in a couple ways. First, while the short volume helps get the pressure high, it also hurts you because as the cone begins to move, the pressure drops (because the available volume expands) more quickly. Also, once you pass the shoulder slope, you'll vent pressure very quickly because the vent area is large compared to the pressurized volume (on a longer rocket, there's more pressurized volume so the pressure drop won't be as fast). Also, the shoulder slope means that the forces on the base of the cone aren't quite symmetric - it's going to tend to rotate. As I mentioned previously, the Big Daddy cone seems easier to bind with a little rotation.
So basically, I think what's going on is that the cone starts moving quickly, but starts to vent/lose pressure before it can get enough forward momentum to overcome any binding friction on the last part of the shoulder and pull out the recovery system.
My ideas for a fix:
1) The somewhat drastic but probably effective approach of hacking off the whole sloped bit ant replacing it with a bulkplate and eye bolt. Has the added advantage of increasing your reccovery volume. You'd need to make sure you have sufficient shoulder to keep it on straight.
2) If you don't want to do that, make sure the nose cone is pretty tight at the unsloped shoulder, but fairly loose in the sloped part. That way, the pressure builds higher before the cone starts to move, yielding a bit more oomph, and the tail of the shoulder can still slide out freely.
3) Add weight to the nose. While maybe a bit counter intuitive, the friction forces are going to be a lot bigger than the inertia forces from the cone. So adding weight probably won't decrease your ejection speed much, but it will increase the momentum the cone has to pull out itself and the chute.