If you've missed the point of this posting:

1. Click the URL, press ctrl+f, and search for "bananas" in the page's text

2. Read the comments

The experiment's results change on page 3 of the comments, around the time that John Timmer (author of the article) tweeted about the experiment: http://twitter.com/#!/j_timmer/status/63361502316474369

That level of knee jerk reaction is surprising, especially on Ars Technica. I'm guessing 90% was a joke and now it looks like a conservative estimate.

It's a shame that the results will be skewed by people who read about the experiment first. The more I think of it, the more I regret posting this to HN.

I'm deleting this submission. Please do not re-submit. Lets wait until Ars writes a blog post with their own analysis of the data.