My Math Forum How to resolve binomial experiment without binomial theorem.

 Probability and Statistics Basic Probability and Statistics Math Forum

 September 16th, 2015, 08:11 AM #1 Senior Member   Joined: Jan 2013 From: Italy Posts: 154 Thanks: 7 How to resolve binomial experiment without binomial theorem. Hi, I have the following exercise: Suppose 20% of the items produced by a factory are defective. Suppose 4 items are chosen at random. Find the probability that: 2 are defective. This is a binomial experiment, but I don't want to apply directly the Binomial theorem. I would want to resolve with basic probability, precentages, and combinations. I have done some reasonings. But I an unable to finish the exercise correctly. Let's see the problem in another way: A factory produces 100 items. 20 items are defective and 80 are not-defective. 4 Items are chosen from the 100 items available, therefore the ways we can do this selection are: $\displaystyle \binom{100}{4}$ Now, we want to calculate the probability that among these 4 selected items, there can be 2 items defective AND 2 items not-defective. I can select the 2 defective items from 20 defective items in $\displaystyle \binom{20}{2}$ ways, AND I can select the 2 not-defective items from 80 not-defective items in $\displaystyle \binom{80}{2}$. Hence, the probability became: $\displaystyle P = \frac{\binom{20}{2}\binom{80}{2}}{\binom{100}{4}} = 0.15311$ Now, my book uses binomial theorem with a given result of 0.1536. I have used another method, obtaining a different one 0.15311. Why I have obtained a different result? What's wrong with my reasoning? Please, can you help me? Many thanks!
 September 16th, 2015, 12:44 PM #2 Global Moderator   Joined: May 2007 Posts: 6,511 Thanks: 584 You started with a finite number (100). The theory is based on an unlimited number. Thanks from beesee
September 17th, 2015, 07:36 AM   #3
Senior Member

Joined: Jan 2013
From: Italy

Posts: 154
Thanks: 7

Quote:
 Originally Posted by mathman You started with a finite number (100). The theory is based on an unlimited number.
Ok. But, does the exercise can be resolved without using Binomial theorem, or not?

September 17th, 2015, 01:26 PM   #4
Global Moderator

Joined: May 2007

Posts: 6,511
Thanks: 584

Quote:
 Originally Posted by beesee Ok. But, does the exercise can be resolved without using Binomial theorem, or not?
You can start from first principles, but all you will end up doing is deriving the binomial coefficients for the particular case.

September 18th, 2015, 09:27 AM   #5
Senior Member

Joined: Jan 2013
From: Italy

Posts: 154
Thanks: 7

Quote:
 Originally Posted by mathman You can start from first principles, but all you will end up doing is deriving the binomial coefficients for the particular case.
please, can you explain me better?

 September 18th, 2015, 01:38 PM #6 Global Moderator   Joined: May 2007 Posts: 6,511 Thanks: 584 A given selection of 4 items with 2 bad and 2 good has a probability of $\displaystyle .8^2.2^2=.0256$ Now count the number of possible combinations with 2 bad and 2 good. 4 items has 4! permutation. However the 2 bad and 2 good are each interchangeable, so that the number of combinations is $\displaystyle \frac{4!}{2!2!}=6$, which is the binomial coefficient for this combination. This leads to .1536. Thanks from beesee

 Tags binomial, experiment, resolve, theorem

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post jbergin Probability and Statistics 2 December 15th, 2014 02:27 PM muhsentdrawjiac Probability and Statistics 1 June 14th, 2012 12:44 PM Keroro Probability and Statistics 6 June 10th, 2012 11:59 AM chocolate Probability and Statistics 1 April 16th, 2012 08:34 AM mikeportnoy Probability and Statistics 2 March 10th, 2009 06:29 AM

 Contact - Home - Forums - Cryptocurrency Forum - Top