suppose we design a reasonable experiment here, like taking a random sample of teenagers and mitigating for anonymity concerns. are you seriously claiming that this would show an increase after the ban?

or to put it another way --- the phenomenon you point to is real but completely swamped by the more direct impacts

If we were talking of the USA, or of a European nation like Croatia, I'd bet that smoking rates in youth show a slow but gradual increase in the years following a UK-style ban.

In the UK, however, the way to bet is that the government enforces the ban with maniacal zeal, so smoking rates will probably go down slightly in the near term. (For better or worse -- usually worse -- the UK has more state capacity to enforce a ban than the US or indeed the vast majority of other countries.) But smoking will become, once again, a prestigious activity. All the cool kids will smoke, if only for the social signaling benefits. Over more than a decade, youth smoking rates might surpass what they were before the ban.

Worth noting that the harm done, both economically and personally, by overpolicing smoking would probably be worse than the consequences of the smoking it manages to prevent