Nada!

Here is my guess about what is going on.

Screen refresh is happening largely in the video hardware while the script execution is happening on the cpu: so you have some parallel processing going on and it is difficult to time these separately. So, I'm guessing that whichever routine returns first gets to be nearly 0, while the other when gets the rest of the time. This would be necessary if all the debug times have to add up to the actual time between frames -- if you time parallel processes the total time would come out too high.

Yeah -- I think parallel processing explains it but the manual should probably explain this when it describes the debug output.