Ben Langhinrichs

Photograph of Ben Langhinrichs

IBM Champion logo

E-mail address - Ben Langhinrichs







Recent posts

Wed 14 Nov 2018

Short agents



Tue 2 Oct 2018

No cookies for you



Fri 28 Sep 2018

We need more than baby steps


December, 2018
SMTWTFS
      01
02 03 04 05 06 07 08
09 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31

Search the weblog





























Genii Weblog

70 sextillion and counting

Tue 22 Jul 2003, 02:51 PM



by Ben Langhinrichs
On CNN.com today, the headline reads: Astronomical: Study counts 70 sextillion stars

If you read the story, it is not until the fourth paragraph that you get a vague hint that the counting wasn't really counting at all, when the phrase "was calculated by" is used instead of "was counted by".  It is not until the fifth paragraph that the article clarifies that "the number was drawn up based on a survey of one strip of sky, rather than trying to count every individual star".  Furthermore: "Within the strip of sky some 10,000 galaxies were pinpointed and detailed measurements of their brightness taken to calculate how many stars they contained. "

OK, so first, only a survey of one small strip of sky was done, and everything was extrapolated from that.  Second, the brightness of a galaxy was used to calculate the number of stars, with no particular evidence that it wasn't twice as many dimmer stars or half as many stronger ones.  Finally, this is extrapolated not just across the sky, but "multiplied again out to the edge of the visible universe", whatever that means.  I am not a scientist, but the likelihood of serious error has exceeded any possible value by this point.  On top of that, studies show that most people don't read beyond the fourth paragraph of any given news story, so nobody even knew this was a wild guess at best.

"So what?" I hear you mutter under your breath.  With stars, it hardly matters, but this is exactly the way Microsoft counts seats when it claims to be dominant in collaboration/e-mail/pick-some-other-category.  Use a lot of fuzzy math; pile the estimates on estimates;  discount other reasonable interpretations, and then declare a number, with all caveats buried on a website seven clicks from nowhere.  And guess what, people believe them, just the way they believe the astronomers!

Copyright © 2003 Genii Software Ltd.

What has been said:


30.1. Tom Duff
(07/22/2003 04:37 PM)

Yes, but the seat wars are over... Just ask Ed.

And FYI... I find it rather disconcerting that when I post a comment, your picture ends up right next to the comments box. I feel like I'm being watched! :-)


30.2. Ben Langhinrichs
(07/22/2003 05:21 PM)

You and I may listen to Ed, but there are a few others out there who list to Bill Gates as well.

As for the photo, I just want you to remember who you are writing to. lol.


30.3. Steven Alwich
(07/25/2003 05:58 AM)

It's not a "wild guess". I can make wild guesses. They used valid statistical techniques in this case, sampling. The important thing about statistics is that any statistic can be skewed into a favorable or unfavorable light. So when Microsoft makes these claims, they're correct. And when others make counter-claims, they're also correct.