So this interaction from last year came to mind as I start to see some postings of papers that study the effectiveness of measures ("NPIs" and "lockdowns").
One way to look at this is that outbreaks happen in networks. Response effectiveness seems diluted when the response.1/ https://twitter.com/yaneerbaryam/status/1250034133499248641

is at a scale larger than the outbreak, as Dr. Bar-Yam pointed out to me back in April. Maybe easier if we think of islands. Let's say that there is an outbreak in Norfolk Island, population 1748 in the 2016 census - it is part of New Zealand, population 4.9 million (google) 2/

Let's say you tested everyone, measured 175 cases, imposed a hard lockdown and restricted movement in and out of Norfolk I, getting the cases to zero. Outbreak contained, successful outcome. If everyone had become infected (w/movement restrictions) you would have 1573 more.

Technically locking down and restricting movement the rest of New Zealand has no effect (except the collateral effect of, for example, reduced crashes, etc. but let's leave that aside. This drop btw was observed in Arg. in April. But this is an aside.)...

With situations where movement restrictions have holes, or lockdowns exceed scale, etc., the effectiveness of policies seems to me to be quite difficult to measure and even less evaluate.

Match the scale to the intervention. Dr. Joseph Norman outlined the phasing of precaution:
1. Identify systemic threat
2. Clamp down.
3. Detect (test), classify at smallest scale that is functional.
4. Open up scaling up from these units up to global level. Repeat steps... +

This gets you to containment, where you can target your interventions at the smallest level possible (i.e. where the outbreak is.) Based on a discussion among very smart people that I've seen this evening, this issue is going to be around epi twitter/covid twitter for a bit....+

This is my word of caution about how to evaluate what at least a couple of these papers are showing.
One last thing: Here's a link to visualize how contagion happens (Dr. Christakis at Yale.) https://hms.harvard.edu/news/friendship-paradox-helps-predict-spread-flu

Stay sane, and focused.

It's going to get dizzying reading these things coming out.