The current government timetable is to announce any tier changes on Wednesday for implementation on Saturday at 00.01.
But there is evidence that when people fear their tier level will rise, which is a well-grounded expectation in infection-growing London, they bring forward socialising, they discount the shift to increased restrictions.
There is evidence these behavioural changes increase coronavirus infections by 10% to 15%.
It is a basic flaw in a system of tiers being reviewed every couple of weeks.
By the way, there is also evidence that when an area comes out of restrictions, it is like unpopping a cork of a fizzy drink: there is such a sense of relief and release, that again people accelerate their virus-spreading socialising.
All of which is an argument for having fewer tiers, perhaps just two, a moderate-restrictions tier and an extreme-restrictions tier, with an expectation of no change unless local hospitals are in danger of being overwhelmed.
The problem of course is we are unlikely to know whether the current three-tiers system, with fortnightly reviews, is sub-optimal until it has comprehensively failed and we are on the way to third lockdown.