Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Say I have two algorithms that I don't know how they work but I know what they are meant to achieve. I tried algorithm A once and it succeeded. And tried algorithm B 100 times and succeeded 99 times. So A succeeded 100% times and B succeeded 99% times, but my common sense tells me B is still more "reliable" than A. How can I express this mathematically, or, how can I calculate which algorithm should I choose if I have to choose one when all I know about each of them is how many times it has been tested and how many times it succeeded (if it didn't that just means it failed for what I need)?

Please be as simple as possible I'm a newbie at math.

share|improve this question
    
Wouldn't (SuccessfulAttempts - 1) / Tries be a good approximation? –  Juan Dec 13 '10 at 9:15
    
${\rm SuccessfulAttempts}/{\rm Tries}$ is a good approximation provided that ${\rm Tries}$ is sufficiently large. –  Shai Covo Dec 14 '10 at 17:12
    
I do the -1 so that cases with small Tries don't get a very high number, so that, for example, 1 / 1 doesn't get 100% –  Juan Dec 14 '10 at 17:17

1 Answer 1

up vote 2 down vote accepted

Perhaps what you are looking for is "Confidence Interval for Proportion". There are calculators for that online, see e.g. this or this. Here is also a calculator that determines the minimum sample size you would need in order for your sample to meet the desired precision requirements for a study .

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.