1
$\begingroup$

Dear Stack Mathematicians,

I'm new to the site and I joined to ask a few math questions. I did some research regarding the topic, but I still have doubts.

Let's say I have a $4$-letters alphabet: {a, b, c ,d}, and I want to know the probability of getting a specific 5 digit string, e.g. 'aabcc'. Now, the possible combinations are $4^5$ , i.e. $1,024$. That means, if I pick up $5$ random letters (with replacement), I have 1/1024 chances of getting 'aabcc' (the order does matter). This means $\frac{1}{1024}$ $\approx0.001$ chance.

Now, I'm going to change things, let's say 'a' and 'b' have each $0.4$ chance to be chosen, while 'c' and 'd' have each $0.1$ chance to be chosen. The chance now to get 'aabcc' in that specific order will be: $0.4\cdot 0.4\cdot 0.4\cdot 0.1\cdot 0.1 = 0.00064$. Thus, if I generate $1500$ random $5$-letters words, I would expect (on average) one of them to be 'aabcc'. Is that correct?

I am looking forward to your feedback! :)

$\endgroup$
1
  • $\begingroup$ Yes, that is correct. $\endgroup$ – Théophile Jan 22 '18 at 16:02
0
$\begingroup$

Yes, you're right, but I suggest you to use the words "events" and "permutations" to make your question more concise and clearer, say "the chance now to get the permutation $aabcc$ will be ...", or even $$\Bbb{P}\{aabcc\} = \frac25 \cdot \frac25 \cdot \frac25 \cdot \frac{1}{10} \cdot \frac{1}{10} = \frac{2}{5^5} = \frac{2}{3125}.$$ To express the expected number of occurrences of the event $\{aabcc\}$ under 1500 independent trials, you may write $$\Bbb{E}\left[1500 \cdot 1_{\{aabcc\}}\right] = 1500\, \Bbb{E}\left[1_{\{aabcc\}}\right]=1500 \, \Bbb{P}\{aabcc\} = \frac{3000}{3125} = 0.96.$$ N.B.: The indicator function $1_A$ counts the number of occurrence of the event $A$ in one single trial.

$\endgroup$
0
$\begingroup$

Think about the probability that you never obtain 'aabcc' sequence; it is $$(1-0.00064)^{1500}\approx 0.38$$ so the probability to obtain 'aabcc' at least once is $$1-0.38=0.62$$ which is greater than $0.5$

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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