I haven't followed this saga super closely, but I've read a little. Here is my uneducated prognostication of whats happening.
I think the safety lock is getting disengaged by jostling as the report has shown. I think the sears are releasing because of poor machining, combined with guys drawing their pistols and taking the trigger to the wall (as training says you should), but not firing. Possibly, this is reducing sear engagment on poorly-machined sears, which later causes a sear to randomly release. What I don't get is why the secondary sear notch doesn't catch the striker in those situations.
Most likely because they did not bother analyzing the dynamic behavior of the the fire control parts. The report identified sear bounce through high speed photography. That should have been fixed long before releasing the product to market. But it was not and most likely they also did not check whether deformed first engagement surfaces can kick the sear away from the striker to the point that the second notch has nothing to grab. Or, after slipping off the first notch, the sear contacts the striker, bounces off and again the second notch misses the striker.
There are many items in the fire control group that scream "amateur" (the way they tried and failed to capture the striker safety spring) or "cheapskate" (3rd world MIM garbage) or a mix of both. On top of this, you have multiple undocumented parts revisions, parts without numbers to assess their compatibility, etc.
SIG, once synonymous with sound engineering and uncompromising quality, now looks more like the Boeing of the firearms industry.
