I took the screenshot then and I was too lazy to do it again before posting :) Also, sometimes the integration that does the prediction becomes more realistic and the difference would have been less strong. But it looks like it didn’t re-evaluate the prediction, and the second half of the day was even worse :)

You are correct of course, generation happens without direct sunlight. In fact, today it was raining and I never saw the sun (until literally 5 minutes before sunset, go figure). It’s actually awesome to get what I get on a rainy day, jokes aside.

I can confidently say that there was no direct sunlight involved today because I have panels on the northwest side, that doesn’t get any direct sunlight these days, and they behaved just like the south. Even pale sun makes the south take off way past whatever the north string can do. Didn’t happen today