I'm way late to the discussion, and I'll apologize in advance for not quoting all the folks who have inspired the following thread. I just want to provide some observations as someone who was on the NFPA committee at the time.
TL;DR: Code making is public. If you don't like the outcome, get involved. The drag race rules were derived from empirical evidence.
NFPA does codes through a consensus process. Periodically the code is reviewed and revised in a four stage, public process. Literally everyone can contribute suggestions, these are adopted or not by the committee, published for public comment, reviewed and modified (or not) again, and then published for final adjudication before being adopted. The Pyrotechnics committee includes mostly Fireworks industry experts, manufacturers, users, and regulators, with a smattering of rocketry folks. The two groups typically respect each other's subcommittee recommendations if they are not controversial. The rocketry members are reps from NAR, TRA, two Special Experts, four manufacturers, and some law enforcement folks with crossover interests. All the commenting and reviewing is open, internet-hosted, and participation by interested parties is welcome. If you care, get involved.
IIRC the drag race rules adopted in the process culminating in the 2013 Code revisions to both 1122 and 1127 (and both organizations' safety code revisions thereafter) were consensus rules adopted with constructive discussion and little dissent.
The motivation included two separate sources. For Low Power, there were the increasingly huge Boy Scout mass launch record attempts, which evolved into filling the infield of a football stadium with 1000s of rockets and launching them with crowds in the bleachers cheering them, and subsequent posting of Youtube videos of rockets, some under chute and sometimes not, raining down on whoever was upwind or downwind.
For HPR, it was Youtube videos of HPR drag races with various unseen anomalous recoveries in which potential harmful events were easily observed and the outcome not easily avoided other than by the luck of where people were standing or cars were parked.
Fireworks folks have to do math before a public display to show that the likelihood of duds hitting people is very small; we had no such requirement earlier since we had mostly been tracking individual objects in daylight and the feeling had been that observant people could get out of the way. Now there was hard evidence that even observant people were being put at risk by these kinds of events. So risk reduction was needed (and as always when something that is mostly not harmful but sometimes could be gets brought to the attention of a code committee) and the mitigation had to be principled.
There is ample human perception and performance and psychophysical evidence that humans can visually track one moving object at a time, that it takes time to switch tracks and acquire targets (sometimes called Observe, Orient, Decide, Act or OODA loop), and that the apparent speed of the target and cross-retina motion is related to the ease with which it is acquired and tracked.
There is also good empirical evidence (and confirming math) that any one point in a landing area is less likely to be impacted as the target area increases (it's related to the square of the distance to the center so doubling the distance increases safety by a factor of 4 if the size/number of "targets" is constant).
So, increasing distance:
--helps by reducing apparent motion
--helps by increasing available reaction time
--helps by increasing the proportion of area where an unintended landing would nevertheless be not harmful (by a lot, given that if other safety code provisions about overflying spectators, etc are followed, the parts coming down representing a danger to spectators are truly anomalous events--if the impact zone is in the crowd because the field is set up wrong, then the situation is even worse).
In lots of ways, I regret that the hobby is becoming as regulated as it is. One way to prevent that, as has often been cited in this discussion, is to self-regulate--help make sure that our enthusiasm is bounded sufficiently by safe practices so that when the inevitable YouTube video gets posted that scares some member of the general public, or someone heaven forfend gets hurt, we can always tell our insurers and our regulators that that activity is not part of what we do and therefore not in need of intervention.