I didn't realise quite how long you needed to spend near someone for the NHS app to trigger. This is not an official table, but pretty sure it's right. It's also dependent on the number of days since the person got symptoms. It can be a combination of both distances too.

For example, if you were on a bus within 4 approx. metres of someone whose symptoms started 2 days ago then you'd need to be near them for 39 minutes in order to later get an alert when they report their positive test result through their phone.

So it's more "sat at the table near them in a restaurant" than "stood in behind them in queue at supermarket". Is this common knowledge or do some people think it's tracking every single person they walk past in the street?

Interestingly, NHS has decided to override the Google/Apple approach and write their own risk score system. The phone tells the app it thinks there might be an exposure but then the NHS code makes its own assessment. I think this is why people are getting spurious notifications.

Even more techie now, but we can actually see the "days since onset" in the data. Here's a breakdown of the 6526 keys announced yesterday:

It's difficult to infer "number of people" from this data. Maybe if we said the average user has had the app installed for 5 days then this represents about 1300 people inputting a positive yesterday? More of an illustration than an estimate!