Neither does the percentage of tested individuals who are positive:

This is p8 of the daily coronavirus dashboard for today, and each day's is available at mass.gov when ready. The "percentage of all molecular tests that are positive" number is much less useful: we also don't want to think things are good just because MIT is retesting their students at an incredible rate.

When our house relaxed our strict isolation in early July things were looking good: rates were down, testing was up. The per-individual positive test percentage was ~1.5-2%:

Hospitalizations are starting to trend upwards as well, though are still far below their peak:

We talked yesterday as a house, and decided on new house rules:

• Our default is the same as it has been since March: you can only interact with other people outside the household if you're outside, masked, and at least six feet apart. I recommend buying snowpants for comfortable outdoor cold-weather socializing.

• We're still treating a small number of people who have had coronavirus and tested positive for antibodies on a reliable test as fully safe. There is still some chance that they could get it again, with immunity feeding overtime, but we think the additional risk from them is very low.

• Previously each of us could pick one person each week to spend time with us unrestricted: no masks, kissing etc ok. We're not doing this anymore.

• Now, each of us can pick one person each week to spend time with inside, as long as we wear a high-quality mask keep faces at least 2ft apart. It's now possible to get KN-95 (Chinese-spec N95) masks at about $2/mask (ex).

• We're still not going into stores to shop etc.

These rules don't bring our risk down to zero, but we think they're a reasonable trade off between danger of getting or spreading the coronavirus and the benefits of spending time with people.

We will of course revisit as things continue to change.