Hacker News new | ask | show | jobs
by twooster 1900 days ago
You're right on your first estimation:

P(getting infected per touch) = 1/10000 = .0001

P(not getting infected per touch) = 1 - P(getting infected per touch) = .9999

P(not getting infected after 100 touches) = P(not getting infected per touch) ^ 100 = .9999 ^ 100 =~ .99

P(getting infected after 100 touches) = 1 - P(not getting infected after 100 touches) =~ .01 = 1%

Or, in other words:

1 - ((1 - .0001) ^ 100) =~ .01 = 1% chance of infection

By the same logic, your second estimation is a bit off:

1/5000 = .0002

1 - ((1 - .0002) ^ 100) =~ .02 = 2% chance of infection

That's a far cry off 40%.

To think about it another way, if the chance to get infected per touch is 1/10, and you touch a surface 20 times, is your chance of infection 200%? No, that doesn't make sense. It's 1 - ((1 - 1/10) ^ 20) =~ 87.8%.

That's the problem with statistics -- "everyone can do the math", but it's pretty easy to mix it up and reach the wrong conclusions and spread misinformation.

1 comments

Yes, my 40% was actually with 1/2000. Thanks for checking the numbers :). Agree it's very easy to fumble numbers.