'"But we have data and we can weigh the relative risk. Data rules."
I'd like to look at that data if you can cite the source.'
I doubt that there is, but in addition this beside the point. This is a misuse of the idea of statistical risk analysis since this particular system is shooting for zero incidents. A single incident in this case can be catastrophic. Using reasoning based on averages (near the center of the distributions) does not work when the events that are at issue – in this case catastrophic forest fires – are the rare ones on the tails of the probability distribution. So good example of a statement that sounds really smart but is in fact empty.
There is a lot more complexity in the balance here than is sometimes claimed by people who always argue only from the point of rationality and fairness of the rules/regulations (almost always from how the rules effect them personally). If you think of things on the systems level you need a few other feature of a fire-prevention regulatory system that WORKS that the irate individualist often leaves out.
For example, SIMPLE rules (not that this in practice is the case) will trump a lot of things because they will have to be clear to the least common denominator jackass. I'm not sure the system will be able to handle distinguishing between a jackass-safe alcohol stove and an a jackass-unsafe one (simple cat stove), nor should you expect the rangers to have to constantly make this distinction. In such a case the least common denominator wins, and should win provided the net result makes the system as a WHOLE a bit safer, without making your life much harder.
There is even another important issue left entirely off here, to increase the consciousness of the average Joe on the issue. There is a chance that some people knowingly using an "illegal" alcohol stove will be that much more careful simply because they know the stoves are problematical in some way. I wouldn't be surprised at all if this was a conscious part of the strategy. I may be pushing it here, but this might even explain why the guys at headquarters often say they alcohol stoves are not allowed, and the ranger don't press the issue. If I were a ranger the fact that someone was asking me about certain stove would let me know that people are conscious of the issue, would be extra careful, and so would be safer. Still they might want to have some leeway to deal with obvious jackasses out there.
Lastly, and this one seem to divide people even more into two camps – as evidence I cite the debate we had a while back about if it would be OK to game the service dog system to bring your dog into a NP. If anything this is more "abstract", meaning the system in question that is being optimized is widened even further. The camps are made up roughly of people who say "what is one dog going to do to the park", and people who say its the principle of the thing. Beyond the principle is the question of whether there would be an ACTUAL long term benefit by obeying rules you find silly as an attempt make the systems as a WHOLE work better. To make the rangers job easier. Possibly to influence OTHER people in such a way that the system as a whole works better in the long term. In the dog example, to discourage an increase in abuse rather than writing it off, and therefore long term harm and extra inconvenience to the people who actually need the service dogs. In the case of stoves, to grin and bear the silly regulation that don't apply to you because you are safe, because you understand that anything that gives that one jackass pause might prevent that one forest fire that gets out of hand and does catastrophic damage.
I don't pretend to know where to draw the line. It is a balance between the "trouble" it causes you to obey the "silly" regulations and how (or if) this would have a net beneficial effect. On an even more meta level is if by constantly belittling the people who are trying to do their best running the system, and the whole idea of regulation if you might weaken the system and make it that much more likely for a jackass to extend ignoring the rule to wider and wider areas.
Unfortunately in the case of rare and very catastrophic events there is seldom going to be useful data outside of a computer simulation. But at least everyone on both sides should admit it is a complicated issue if you actually embrace the whole problem and not just the part of the problem you personally care about, and quit QEDing very over-simplified arguments.