I'd heard that Google had been experimenting with reducing the number of results to a page, but I hadn't actually seen it for myself. Now however I have - and rather wish that I hadn't. I was looking for a particular software game that was produced years ago, and the graphic shows you the search and the results:
Now, there's a fair amount to be said about this. I'm only getting 4 results that match my query. Actually, it's less than that, because the last result from the 4; 'Green battleground' didn't include the word Talon. Actually, it's less than that, because the third result doesn't contain the word 'battleground'. So out of the 4 results, 50% are incorrect. We then have a line break, with "Results for similar searches" listed for me. Why does Google assume that I'm interested in similar searches? Particularly since the searches they've run for me are not accurate. 'Wargames Empire' and 'Empire war games' are totally different, despite a vague semantic similarity.
What's really annoying is that if I go onto page 2 of the results, I carry on with organic results from my first search, and I do get the information that I was after. So, not only does Google know what I wanted, it retrieves what I wanted, then insists on showing me results that it knows are inaccurate! I am therefore forced onto the second page - where I get shown more adverts!
Now, you might well say 'Use the Verbatim tool'. For those not familiar with it, the verbatim tool tells Google to stop trying to think for itself and do as it's told, by finding all the words on the pages it returns. That doesn't actually work though, because the results that I'm getting back don't include all 3 terms - several pages are missing one of the 3 terms! So that's not something that works either.
This is really the absolute worst that I've seen Google do myself, though I'm familiar with other examples of Google totally messing up a search - who can forget Karen Blakeman's Coots and lions post! I really am almost entirely lost for words - just when you think Google can't get any worse - it does! Well, I shall go back to sensible search engines, and leave this one to fester some more.
Google searches have so many variables now. You always have to one browser open in which you're not logged in and are going incognito to grab more relevant search results, not the "intelligent" ones returned by the search engine.
Posted by: Tamara Thompson | March 27, 2013 at 02:26 AM
Thank you! This happened to me and I thought I'd gone mad. I knew it was an obscure search, but I didn't think it was 4-results obscure. I'm feeling happier now I know what's going on. I'm feeling less and less happy with Google, however, every time it tries some new way of being smart. Time to go somewhere else!
Posted by: Merry | March 27, 2013 at 06:43 PM
DuckDuckGo and Blekko are poor substitutes for Google, as (their documentation notwithstanding) they do not correctly implement the OR operator. Try, for example, searching for
"bubble bread" "los angeles" OR glendale
and you'll see results having nothing to do with bubble bread. Despite its quirks, Google gets it right.
Posted by: Richard Johnson | March 27, 2013 at 07:17 PM
The order of precedence with Boolean operators is important here. Your search could be seen as (A and B) OR C, or alternatively A and (B OR C). I looked through the results that Google returned and there were plenty without reference to Glendale.
No, Google doesn't 'get it right' - there are plenty of times when Google returns results that don't contain the term(s) that were asked for. They will automatically exclude some terms that are unusual or for which they think there will be few results. Google also decides to re-write search terms as it feels like - my search for 'teaching search skills' gets rewritten as 'teaching research skills', and I have to do the search a second time to get what I wanted.
Agreed, DDG and Blekko are not perfect - but Google is a very long way from being so itself, as my example above shows.
Posted by: Phil Bradley | March 28, 2013 at 05:09 PM
I figured this must be to cater to the short attention spans of people who are not serious researchers, but then they cut Scholar down to a maximum of 20 per page! Researchers will look at any number, and this just makes a serious search go that much more slowly. Now I'm thinking that it's because those of us who set results to 100/page don't see enough ads, so they'll force us to look at more. I wonder how much longer they'll be able to claim they're "trying to improve the search experience" with a straight face. They jumped the "evil" shark a while back.
Posted by: Caryn Wesner-Early | March 29, 2013 at 08:43 PM
Phil--Whether or not one believes that Google "gets it right" w.r.t. the OR operator depends on one's expectations. Users of Google know that Google fudges on the classical definitions of operators in general. For example, Google users will expect a search for "hounds of the baskervilles" to yield results pointing to The Hound of the Baskervilles, quotes notwithstanding. Google obviously thinks its fudging is more helpful than harmful. Regardless, for Blekko and DDG to return results for the OR search expression I quoted that completely ignore the first term makes those search engines unacceptable to me.
Posted by: Richard Johnson | March 30, 2013 at 12:28 AM
Richard - well, if you prefer Google over other search engines, that's fair enough - at least you gave the others a go, which is more than a lot of people do. All that I'd say is that I'm perfectly happy with a situation where (a+b)OR c sometimes leads to results that are 'c' and not (a+b) because that seems perfectly logical. If you really want A and (b or c) then the search needs to be constructed in a different way.
Posted by: Phil Bradley | March 30, 2013 at 09:33 AM