That seems like just a fancier way of expressing TWIV's "we can't possibly know anything without lab experiments that establish the exact mechanisms, so we should do nothing" philosophy.

The undeniable fact is that this variant has successfully replaced other already highly prevalent variants. That needs to be explained somehow.

Could it be a founder effect? No.

Could it be random chance? No. Of course a stochastic process could converge like this even with no selective advantage eventually. But the change has been too fast and too consistent in this case.

Could it be a super-spreading event? No. The change has been continuous, and a single event would cause just a step-change in relative prevalance.

Could it be the new variant being more transmissive, or having another similar selective advantage? Yes.

Could it be an unspecified emergent behavior? I guess it could. But how are we supposed to reason about something that vague? We have a simple explanation that's consistent with the known facts, I feel that anyone proposing that it's just emergent behavior should be at least a little bit more specific.

> But which action?

In the UK, my view is the action to take was very clearly to dial back the public's expectations of having a 5 day "free for all" period of relaxation of our restrictions over Christmas.

This 5 day period, as originally planned, was going to lead to massive cross-country travel, plenty of it being completely unnecessary, with no doubt a lot of risk taking in the spirit of "it's Christmas" and an inevitable huge spike of cases and deaths in the weeks that followed.

I mean I love Christmas, but I can cope with not having 5 days of partying for just one year... *

As it happens, the government did take action to dial this back at the very last minute to allow just 1 day of relaxed rules (Christmas Day) rather than the week that was originally on the table. In my view they should have been planning for this right from the start and they unnecessarily screwed up a lot of people's Christmas plans by leaving it until the very last minute to make this change.

Now even if the variant turns out to be a "non issue" (and we still have rising numbers anyway), I think it's better that they took this action to further limit Christmas (as tough as it is for everyone) rather than take the "wait and see" or "hope for the best" approach, both of which seem to have been UK govnt strategies at one time or another during 2020.

* I'm being slightly flippant here. This is clearly a tough time for a lot of people, and many families will be spending Christmas apart and will find this very hard for all sorts of reasons.

Don’t you think this reflects a bias in the opposite direction?

Systems are dynamic and complex, but that doesn’t make them unknowable. If complex systems were impossible to model science wouldn’t be a thing.

Sure, the “new variant” narrative is an easy one to pick up because it’s superficially understandable; but what other complex interactions are resulting in a new strain becoming dominant over an existing one?

Do we just sit on our hands and go, well gee, better do nothing? Guess it’s just too hard to know what to do at this point...

There is a typical government fallacy that to be seen doing something is better than nothing, even when it’s the wrong thing... I get it, that’s bad.

...but this is a case where doing nothing has been a colossal disaster so far; and it’s extremely clear what the result of doing nothing different will be if the new variant hypothesis is correct.

So it’s a risk game. Is the cost * probability of A vs B a better choice?

You’re saying the probability is not known at this point because systems are complicated.

So what? It’s still clear to me that it’s very likely that something is happening, even if exactly what it is, is unclear.

You can still build a risk matrix taking that into account.

Personally I’ve taken it to changing optimization tactics. I’m going to use pseudo-statistical language to illustrate my point so please try to be charitable. In a usual model I’d base my actions on the mean outcome for a given set of choices. For covid, I think it’s smarter to choose to minimize the maximum, worst outcome. Ie assume the absolute worst and react in the harshest way. This new strain may prove to be nothing major, but personally I’ve decided it’s wise to be even more distanced from possible disease vectors. I suspect this may be a good strategy in a universe in which an already rare occurrence has been observed, in a larger sense.

> removing the public's focus on the more evenly distributed measures we are already employing

How is this focus "removed" and why is there any reason to think that evenly distributed measures along with targeted ones would be a problem for the public?

What would be your advice we should do in the face of a system so complex it's impossible to understand or predict?