I've noticed that when downloads get queued (not immediately downloading) the elapsed time apparently begins when the file is added to the queue, not when it actually begins to download. If a file spends ten minutes in the queue, when it actually begins to download the time elapsed will instantly go from zero to ten minutes. It appears the average speed calculation uses the elapsed time. The result of these two things is that average speed for files that have spent a lot of time in the queue is totally inaccurate.
I suppose it does depend on one's perspective on this. An argument can clearly be made that the average speed displayed for the file is correct. Unfortunately, it also isn't very useful. I can't imagine why I'd care that a file that's spent two hours in the queue has an overall effective average speed of 10KB/sec, which amounts to two hours at 0KB/sec and two minutes at 1220KB/sec.
Just to be clear, I'm referring to the number displayed in the "All Download Jobs" window. The average speed in the Job Queue window does appear to be correct for each of the segments there.
Bookmarks