0:00:00.032,0:00:03.044 Today's question comes from [br]San Francisco, California. 0:00:03.044,0:00:07.079 Blind Five Year Old asks, "How reliable is the site colon, 0:00:07.079,0:00:11.004 query, and why does the total count sometimes change from 0:00:11.004,0:00:12.096 page one to page two?" 0:00:12.096,0:00:14.601 Well, I wouldn't really put it in terms of 0:00:14.601,0:00:16.011 the site colon query! 0:00:16.011,0:00:19.003 Instead, I would pivot and narrow and focus down on the 0:00:19.003,0:00:20.062 results estimates. 0:00:20.062,0:00:21.852 So there's a few things to know. 0:00:21.852,0:00:24.453 Number one, if you haven't noticed, you should notice 0:00:24.453,0:00:26.061 that results estimates are only accurate to about three 0:00:26.061,0:00:27.965 significant digits. 0:00:27.965,0:00:29.051 So they are estimates. 0:00:29.051,0:00:32.631 As we're looking for queries, at some point we say, OK, we 0:00:32.631,0:00:37.009 have enough to constitute the first page, and we see how far 0:00:37.009,0:00:38.097 we've gone through our posting list - the 0:00:38.097,0:00:40.019 amount of data we have-- 0:00:40.019,0:00:42.005 therefore, we can estimate there's about this many 0:00:42.005,0:00:43.356 results left. 0:00:43.537,0:00:46.058 Now when you click from page one to page two, suddenly 0:00:46.058,0:00:47.072 you're going deeper. 0:00:47.548,0:00:49.069 And so you have a better estimate or you have a 0:00:49.069,0:00:51.069 slightly different estimate, because you're starting to 0:00:51.069,0:00:54.055 look not just a little bit of the way through all those 0:00:54.055,0:00:55.077 documents, you're actually starting to 0:00:55.077,0:00:57.002 go deeper and deeper. 0:00:57.002,0:01:00.001 So it's relatively common for our results estimates to 0:01:00.001,0:01:04.066 change going from page one to page two, typically becoming 0:01:04.066,0:01:05.063 more accurate. 0:01:05.063,0:01:08.024 So it's not usually a factor of the site colon query. 0:01:08.024,0:01:10.049 It's usually a factor of us digging deeper through our 0:01:10.049,0:01:13.012 data, and as a result, we have a better or at least a 0:01:13.012,0:01:15.029 different estimate for how many results there are.