TheOptimizationIdiot - 1:58 am on May 28, 2013 (gmt 0)
I would think that would depend on the update schedule for a specific portion of the index. I personally think they update different portions of the index at different intervals to save on processing cycles/computing power necessary to update the entire index.
Also, I was editing and got distracted while you were posting, so see my previous post for an example. (Sorry, life got in the way of a post, oops! lol)
Another, possibly better, way of explaining what I'm thinking would be if a portion of the index is updated more frequently (near real time), there could be "days of bouncing" based on historical changes made to a site over 4 or 5 or even 10 days, yet for other portions of the index there may be no "bouncing" due to less frequent rebuilding of that portion of the index.
So, in one case, on-site (or whatever) changes over 4 or 5 or 10 days could have a bouncing effect for that number of days, but in other cases those on-site (or whatever) changes may all be "lumped together" in the next update of that portion of the index. (Hopefully that makes a bit of sense. I know what I'm trying to say, but sometimes I have difficulty in spitting it out lol)