] argue that a one-way ratchet distorts scheduling of new psychoactive substances (NPS) in the United Kingdom, easily cranking up stringency but rarely easing off. Based on their experience within policy circles, they attribute the ratchet to three associations.I am unsure how to test the power of their three associations versus other explanations. I have thought about these issues from the perspective of Type I/Type II errors [2,3]. Policy could err by banning a substance that should be allowed or by allowing one that, in retrospect, we wish had been banned. The latter would create more tangible evidence of error, so policy might naturally shy away from making that type of error.Also, the international conventions underpinning the scheduling apparatus do not recognize any value from drug use that is not medicinal. Therefore, a substance without recognized medical value should-by the treaties' logic-be banned if it poses any risk at all, even if it were a fount of great joy and happiness from 'recreational use'. That could look like a ratchet, but if the real complaint is with the treaties we should temper criticism of bureaucrats' faithful implementation of treaty obligations.I will focus, however, on how big the ratchet is, not why it exists, and distinguish (i) initial scheduling from (ii) subsequent adjustments.Stevens' & Measham's Table 1 lists changes to drug classification in the United Kingdom. They cite eight adjustments, five in the direction of tighter control and three loosening. Five-to-three is not parity, but when flipping eight fair coins there is a 36.3% of getting five or more heads, and a P-value of 0.363 does not reject a null hypothesis of no ratchet.Rather, the asymmetry in Table 1 pertains primarily to whether a substance, or category of substances, has been scheduled at all. Twelve have been added versus just one removed, and one is very unlikely (P = 0.002) to see 12 or 13 heads when tossing a fair coin 13 times.That is not the right test, however. As Stevens & Measham note, there has been a 'profusion' of new substances, and Table 1 omits NPS that were not scheduled. If 12 were scheduled but 120 were not, there would be no evidence of an indiscriminate reflex towards scheduling, so before drawing conclusions we need to understand how many substances were not scheduled, and it would be useful to contrast parallel analyses for other countries that have different regulatory processes and organizational structures.Nor would such a reflex necessarily be bad if it were practical to de-schedule later. Drug use and markets have 'tipping points' and other forms of history dependence [4-6] that make them resemble Pandora's Box. If a new substance arose that could create substantial harm if companies were allowed to distribute it for recreational use, then it would be easier to prevent that market from becoming established than to ban it after there were a pool of dependent users and an industry lobby befriending politicians.The real question concerns how many substances that were placed on the sc...