Ok, we know that Google often does strange things with our searches but much of the time it is not obvious that something odd has happened. There are usually some "good enough" answers scattered through the first 20-30 results so that we shrug off the rest as "well, that's Google for you". Occasionally, though, one comes across a search that seems to break Google. One such example was reported on Twitter this morning by Rand Fishkin (@randfish). The search was
this is the best * on the internet
At the top of the first results page Google reported that it had found over a billion results but when @randfish moved to the next page Google showed just "2 of 12 results"! Whatever happened to the other billion or so?
I tried the search myself on my laptop and straightaway got three results but on repeating it that was reduced to two.
I repeated the search having logged out of my Google account, cleared cookies, used Incognito and different browsers. Same results.
I tried a phrase search and the number of hits increased to 17.
Then I removed the quotation marks, got back to my original set of two and ran Verbatim on it. Over a billion hits but, bizarrely, Google claimed to have gone straight page 2!
Note: you normally can't see the number of results after you have run Verbatim because it is obscured by a second menu line. You can toggle between that menu and the number of hits by clicking on the Tools button.
Then I tried a phrase search followed by Verbatim: two results but different from my first set.
I could have gone on trying various advanced search commands but it is very clear that Google is having problems with this particular search. And, no, I have no idea what is going on here.
If Google messes with your search to this extent or comes back with far fewer results than you would expect don't struggle with it; just go to another search engine. As an asterisk is used in this search to stand in for a missing word Yandex.com would be the best option. (See https://yandex.com/support/search/how-to-search/search-operators.html for a list of the main operators).
I think Google is confused because most of the words are stop words. (Not sure about "internet" - always thought it was but not on the lists I just looked at which are mostly old, but probably still valid to an extent).
ReplyDeleteGoogle has coded for known phrases with stop words e.g. "to be or not to be" but this search was not that sort of phrase. So I wonder if what is happening is that Google is confused as it doesn't know what words to actually include and which ones to ignore in its algorithm.
You can test this by missing out words.
I tried ["this is the best" * "on the internet"] and got 15 results.
"this is the " * "on the internet" gave 19 results
"this is the best" * "on the" gave 14 results as did
"this is the best * on the"
"this is the * on the" gave 40 results
["is the * on the" internet] gave 54 results
["is the * on the" planet] gave a number of 588m BUT only 111 were shown (although there was the option to see the rest!)
Planet alone gave 754m results so around 30% more.
[* "the planet"] gave 460m - with several pages of results.
["the best * the planet"] gave 206m with several result pages
BUT ["is the best * the planet"] went down to 89 and ["this is the best * the planet"] gave only 13 results.
So if I'm right, and it's stop words, it's an extra thing to include in search training.
Best solution: learn Dutch. This is the best * on the internet only results in 3 hits on google.nl, but when I repeat the search in Dutch (dit is de beste * op het internet) I get 158.000.000 results. Lucky me/gelukkige ik!
ReplyDelete