Since I didn't like the type of retrieve I got from Zeus training it on my own, I went back to my trainer and have been doing it her way with much better results. Zeus would retrieve most objects perfectly fine and quickly, but he wasn't holding the objects tight enough, and it was more of a "hold your breath and hope it works out" kind of thing. I needed to get him more comfortable with holding the dumbbell no matter what (the schutzhund dumbbell is very heavy and large for a dog who is generally uninterested in things being in his mouth.)
But anyways, she has had me go back to the very beginning of having him simply holding the dumbbell in every situation. As in first holding it in a sit while watching you, then the dog stays in sit and you take steps back, step forward, then you rotate to the side of them, and so on. Once you can move anywhere possible and do anything without the dog dropping it, then you start to ask the dog to move and hold it at the same time. This is done by walking on lead and offering the dumbbell at the same time. The dog should continue walking generally with you (not formal heel position) while holding the dumbbell. He should also be ok and immune to your hands doing whatever, you should be able to pet them where ever, mess with the dumbbell or touch their mouths without them dropping it. These are the first parts in having a 'proofed' retrieve.
This is the stuff I skipped when teaching the retrieve, which made Zeus much more sensitive to outside influences whilst doing the actual retrieve. There's no way I could have touched him, moved his head, or walked beside him at first. And I needed to be able to be beside him when I added in the wall. On trial day, the dogs will pull every excuse out of the book, and proofing it this way, helps minimize the possible excuses.
I'll get a video of stuff to do while simply teaching the hold in a stationary position.
The retrieve exercises are 40% of the points in a Sch trial. I am working on that right now. It is one of the things that I have just not spent enough time on and it is certainly not proofed.
We backchain the whole sequence of learning the dumbell retrieve. Actually we don't teach it as a retrieve at all. The problem with throwing a dumbell and then expecting them to run and bring it back as a game is that it is purely a prey exercise, then. In this case typically what you will see is the dog brings it back much slower than he goes out to get it. Also not taking it seriously, hence the chewing and screwing around.
We start off with the front position, and use a wooden dowel (has to be very hard wood) or 1" plastic pvc pipe with a string attached to it. The dog first learns to take it and then maintain the front position with the dowel. a second person yanks on the string attached to the dowel. if the dog is not holding it securely it comes out. He then gets grief in the form of low level e-collar stimulation until he re-grips. The dog learns to not let go. It then progresses to the dog going out and getting the dowel and bringing it back. There is low level stims until the dog brings back the dowel to the front position. We then progress to a break away dumbbell. My training director, Lance Collins invented this. He allowed Schweikert to produce and sell it (they marketed it as the LC dumbbell). Gappay also copied it and from what I understand makes a better one than the Schwiekert. This process can take many sessions to get to the point where you can proof it.
In this process also a second handler is using a long line to invoke the reverse response in the dog, to increase willingness to be in the correct position and to increase speed, precision and power in the exercise.
Several years ago Lance did a 3 day seminar on just the retrieve, so I do not expect anyone to get much from the little I have written.