Generalizing a bit, probabilities greater than 1 arise when there are multiple ways for something to happen, while negative probabilities arise when there are multiple obstructions to something happening.

]]>Similarly, suppose we have a water tank with a single pipe, but with two valves. Both are closed. Then we could say that the probability of getting water at the outlet is 0. But wouldn’t it be better to say that the probability is -1, to capture the fact that if one valve is opened there is still no water (probability 0), and it takes two valves to be open to get water (probability 1)?

]]>If you want to see all the talks in one place go here; it’s somewhat more avidly curated than the official seminar website. I can tell the people in charge of that website to add links to everything.

]]>I just came across this and was also very happy to hear the video is posted. You might add links to videos on the seminar web site; I had checked for a video of this talk there and was disappointed not to find one. (The slides alone are great, but I couldn’t help feeling that once I got to “the main idea”, it would have been helpful to hear what she said out loud!)

]]>Yes!

]]>After a while we reach

should read

]]>When renormalization was first developed it seemed ad hoc, but now we understand the physical meaning of it, e.g.:

• John Baez, Struggles with the continuum (part 6).

• John Baez, Renormalization made easy.

This fact combined with its remarkable success at predicting results of experiments means few people are trying to get rid of it. Certainly nobody knows how to do so without eliminating its successes. I haven’t heard of attempts to do this by allowing negative probabilities.

]]>