Is it possible that zeroes are excluded in intervals and in the summary of a zoomed range? I noticed that these numbers are far too high for parts with a lot of no-pedaling.
Thats correct. The original idea was that you don’t stop pedalling during an interval so a zero is probably a drop out. I might revert that now that Intervals.icu does some automatic fixing of drop outs when analysing the data and also supports manual editing and additional fixes.
Its quite a big change in behaviour though. You can select real avg power which includes zeros: