Smashed potatoes . . .

[Benjamin Zimmer]

On 1/13/07, Geoffrey Nunberg <nunberg at sims.berkeley.edu> wrote:
>
> Well, not as whopping as all that. If you go to the last page of the
> Google results, you'll find that the phrase actually turns up just
> 829 hits, or 984 if you ask Google to show "similar results" (i.e.,
> additional results from the same url). This is just another example
> of some well-known problems with Google's hit-count estimation
> algorithm.
>
> One way to get a better indication of the true ratio here is to add
> some distractors that lower the hit counts to something more in the
> range of what the algorithm can actually count. "Mashed potatoes
> reply Cleveland" gets 17500 estimated hits; "smashed potatoes reply
> Cleveland" gets 24.
>
> For discussion of this and related issues, see:
>
> http://itre.cis.upenn.edu/~myl/languagelog/archives/001834.html
> http://itre.cis.upenn.edu/~myl/languagelog/archives/001831.html
> http://itre.cis.upenn.edu/~myl/languagelog/archives/001832.html
> http://itre.cis.upenn.edu/~myl/languagelog/archives/000397.html

And also this one:

http://itre.cis.upenn.edu/~myl/languagelog/archives/002658.html

...in which I discuss the "similar results" issue. The 829 vs. 984
results above are not very meaningful, since these results will always
cut off at some number under 1,000 (generally between 800 and 950).

Google's explanation here:

http://www.google.com/support/bin/answer.py?answer=484&topic=359


--Ben Zimmer

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org