Seems like a good calculator. I tested a few things I thought it might screw up, like whether it increased the risk level linearly with the length of an interaction, and it did not screw them up. It seems to come to about the same conclusions that I did before, which is that biweekly Minnstf picnics are a reasonable risk except for particularly vulnerable people. That assumes everyone is vaccinated, and holds with or without masks. My couple-times-a-month indoor hangouts with small groups of friends come with about the same total risk, which is again reasonable.

I'm a little confused by "Describe the scenario", which only lets you say that you'll be with 2 other people, but then later you can set the number of people? I think it is saying that it will start you at 2 and you can adjust it. Or is it saying that you'll be near 2 people, but others will be filtering past?

The biggest weak point is that it only models the risk of getting COVID, not the risk of getting sick, or being hospitalized. There should be an additional multiplier for that, at least optionally.